SEARCH

SEARCH BY CITATION

Summary

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References

Structural backbones of iron-scavenging siderophore molecules include polyamines 1,3-diaminopropane and 1,5-diaminopentane (cadaverine). For the cadaverine-based desferroxiamine E siderophore in Streptomyces coelicolor, the corresponding biosynthetic gene cluster contains an ORF encoded by desA that was suspected of producing the cadaverine (decarboxylated lysine) backbone. However, desA encodes an l-2,4-diaminobutyrate decarboxylase (DABA DC) homologue and not any known form of lysine decarboxylase (LDC). The only known function of DABA DC is, together with l-2,4-aminobutyrate aminotransferase (DABA AT), to synthesize 1,3-diaminopropane. We show here that S. coelicolordesA encodes a novel LDC and we hypothesized that DABA DC homologues present in siderophore biosynthetic clusters in the absence of DABA AT ORFs would be novel LDCs. We confirmed this by correctly predicting the LDC activity of a DABA DC homologue from a Yersinia pestis siderophore biosynthetic pathway. The corollary was confirmed for a DABA DC homologue, adjacent to a DABA AT ORF in a siderophore pathway in the cyanobacterium Anabaena variabilis, which was shown to be a bona fide DABA DC. These findings enable prediction of whether a siderophore pathway will utilize 1,3-diaminopropane or cadaverine, and suggest that the majority of bacteria use DABA AT and DABA DC for siderophore, rather than norspermidine/polyamine biosynthesis.


Introduction

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References

The acquisition of iron from the environment is essential in most organisms due to the wide range of iron-requiring proteins involved in primary and specialized metabolism (Miethke and Marahiel, 2007). One of the most successful mechanisms to have evolved in microbes for acquiring iron is the biosynthesis of small iron-scavenging molecules, siderophores, which are exported into the local extracellular milieu where they bind Fe(III) with high avidity. The siderophore molecule and its iron cargo is then shuttled via active transporters back into the cell, where it relinquishes the iron. Ability to synthesize and use siderophores is a known virulence determinant in bacterial and fungal pathogens (Schaible and Kaufmann, 2004). Most siderophores bind iron by using oxygen ligands derived from catecholate, hydroxamate or (α-hydroxy)carboxylate groups. The best understood siderophore biosynthetic pathways are dependent on non-ribosomal peptide synthetase (NRPS) enzymatic assembly lines (Barry and Challis, 2009; Koglin and Walsh, 2009; Strieker et al., 2010) producing siderophores such as enterobactin and mycobactin. Recently, an NRPS-independent siderophore (NIS) biosynthetic pathway has come to prominence (Challis, 2005; Oves-Costales et al., 2009). This pathway produces structurally diverse siderophores including a number of hydroxamate group-containing molecules for which the corresponding biosynthetic gene clusters have been identified (Fig. 1), including aerobactin (de Lorenzo et al., 1986), rhizobactin 1021 (Lynch et al., 2001), putrebactin (Kadi et al., 2008a), bisucaberin (Kadi et al., 2008b) and desferrioxamine E (Barona-Gomez et al., 2004). Similar siderophore structures (Fig. 1) include schizokinen (Mullis et al., 1971), acinetoferrin (Okujo et al., 1994), arthrobactin (Linke et al., 1972) and synechobactins (Ito and Butler, 2005); however, the biosynthetic gene clusters responsible for synthesizing these have not been identified.

figure

Figure 1. Structures of hydroxamate NIS-type siderophores. Lysine or diamine backbones are indicated in red (and specified in brackets under the name of each siderophore), acyl groups in green, hydroxyl groups in blue, citrate moieties in black and succinyl groups in purple.

Download figure to PowerPoint

The defining characteristic of the NIS biosynthetic pathways is the presence of one or more nucleotide triphosphate-dependent synthetases responsible for condensation reactions during siderophore biosynthesis (Challis, 2005; Oves-Costales et al., 2009). The first examples of NIS synthetases discovered were the IucA and IucC proteins involved in aerobactin biosynthesis (de Lorenzo et al., 1986). Siderophores depicted in Fig. 1 can be seen to constitute two groups: one group contains a citrate moiety and has an open structure (aerobactin, arthrobactin, schizokinen, rhizobactin 1021, synechobactins and acinetoferrin); the other group contains succinate and not citrate, and the siderophores have a cyclized, closed structure (putrebactin, bisucaberin and desferrioxamine E). Except for aerobactin, which contains a lysine moiety, the other siderophores contain the diamines 1,3-diaminopropane, 1,4-diaminobutane (putrescine) or 1,5-diaminopentane (cadaverine). The diamine/lysine moiety is N-hydroxylated and N-acylated by an N-monooxygenase and an N-acyltransferase. It is the N-hydroxyl group, and adjacent carbonyl group from the transferred acyl unit, that forms the hydroxamate functionality for iron binding. The cyclized, closed siderophore structures all contain succinate as the transferred acyl group, which is probably why they are able to form iterative, cyclized structures, via the second carboxylate group forming an amide bond with another N-hydroxy, N-succinyl-diamine unit, catalysed by the NIS condensing enzyme. Although N-monooxygenases hydroxylating ornithine and lysine have been characterized (Olucha and Lamb, 2011; Olucha et al., 2011), and exhibit highly specific substrate preference, siderophore-related diamine N-monoxygenases have not been characterized. The only siderophore-related acyltransferase characterized is the aerobactin-related N-hydroxylysine acetyltransferase, which is also active with N-hydroxyornithine and N-hydroxycadaverine (Coy et al., 1986).

Putrescine- or lysine-based NIS biosynthetic gene clusters, e.g. putrebactin with putrescine and aerobactin with lysine, do not contain the genes for putrescine or lysine biosynthesis and instead use the cellular pools of these precursors. It has been speculated (Tunca et al., 2007) that the cadaverine unit of desferrioxamine E in Streptomyces coelicolor A3(2) is provided by the putative PLP-dependent decarboxylase encoded by desA [NP_627012]. However, there is no proof of any lysine decarboxylase (LDC) activity by DesA, and although a genetic deletion of desA abrogated desferrioxamine biosynthesis, the deletion could not be complemented by expression of an extrachromosomal copy of the desA gene but only by the whole desferrioxamine gene cluster, suggesting a polar effect of the mutation and inadvertant downstream disruption (Tunca et al., 2007). The protein encoded by desA does not exhibit any sequence homology to the two known forms of LDC: the aspartate aminotransferase-fold LDC, which is mainly involved in acid resistance (Kanjee et al., 2011); and the bifunctional ornithine/lysine decarboxylase from the alanine racemase-fold, which is exclusively biosynthetic (Lee et al., 2007). In fact, DesA is highly similar to l-2,4-diaminobutyrate decarboxylase (DABA DC), an enzyme involved in biosynthesis of the diamine 1,3-diaminopropane (Ikai and Yamamoto, 1994; Lee et al., 2009). The substrate for DABA DC, l-2,4-diaminobutyrate, is produced by the enzyme l-2,4-diaminobutyrate: 2-ketoglutarate 4-aminotransferase (DABA AT). Open reading frames encoding both DABA AT and DABA DC are found adjacent to one another in the rhizobactin 1021 NIS biosynthetic gene cluster of Sinorhizobium meliloti 2011, found on the pSyma megaplasmid (Lynch et al., 2001) and presumably synthesize the 1,3-diaminopropane unit of rhizobactin 1021. The DABA AT and DABA DC pair are found as a functional gene fusion in Vibrio cholerae (Lee et al., 2009), which produces 1,3-diaminopropane when expressed heterologously in Escherichia coli. Here we show that the DABA DC homologue, DesA of S. coelicolor A3(2) is a functional lysine decarboxylase. We hypothesized that a NIS biosynthetic cluster encoding a DABA DC homologue in the absence of an adjacent DABA AT would encode an LDC, and we biochemically confirmed the identity of a DABA DC-like LDC from a NIS biosynthetic gene cluster lacking a DABA AT homologue, from Yersinia pestis. The corollary of this hypothesis is that a DABA DC-like gene found adjacent to a DABA AT gene would encode a bona fide DABA DC and we biochemically confirmed the identity of a bona fide DABA DC from a NIS biosynthetic gene cluster of the filamentous cyanobacterium Anabaena variabilis, a species known to synthesize the 1,3-diaminopropane-based siderophore schizokinen (Simpson and Nielands, 1976). By analysing complete genomes for NIS biosynthetic gene clusters encoding only DABA DC homologues or DABA DC and DABA AT pairs, and then creating a phylogenetic tree of the DABA DC-like sequences, we were able to distinguish a subclade of DABA DC homologues that were derived only from NIS biosynthetic clusters containing DABA DC in the absence of DABA AT, and therefore likely representing a novel form of LDC. The available X-ray crystal structure of DABA DC is highly similar to l-glutamate decarboxylase and related enzymes, and thus the DABA DC-like LDC of the desferrioxamine E biosynthetic cluster represents an example of convergent evolution for lysine decarboxylation, constituting a third distinct form of LDC. The evolutionary diversification of substrate specificity resulting in the DABA DC and DABA DC-like LDC proteins has generated structural diversity in bacterial siderophore biosynthesis. Recruitment of DABA AT may have influenced the selection of l-2,4-diaminobutyrate over l-lysine as the preferred substrate for the LDC.

Results

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References

Genomics-based prediction of a novel lysine decarboxylase

The desferrioxamine E biosynthetic gene cluster of S. coelicolor A3(2) contains a putative PLP-dependent decarboxylase encoded by desA (Barona-Gomez et al., 2004). DesA protein amino acid sequence [NP_627012] is very similar to biochemically confirmed DABA DC proteins from the γ-proteobacteria Acinetobacter baumannii (37% amino acid identity, 183/494) (Ikai and Yamamoto, 1994), Haemophilus influenzae (38% id., 180/477) (Ikai and Yamamoto, 1998), Enterobacter aerogenes (40% id., 188/474) (Yamamoto et al., 2000) and V. cholerae (38% id., 177/472) (Lee et al., 2009). The bona fide DABA DC-encoding ORFs are each found immediately adjacent to ORFs encoding DABA AT and in the case of Vibrionales species, the DABA AT- and DABA DC-encoding ORFs are fused (Lee et al., 2009). In contrast, there is no DABA AT-encoding ORF present in the S. coelicolor A3(2) desferrixoamine biosynthetic gene cluster (Barona-Gomez et al., 2004). Desferrioxamine E contains a cadaverine unit (Fig. 1), which must be derived by lysine decarboxylation, therefore the desA gene product, although exhibiting high similarity to DABA DC, may have evolved to decarboxylate l-lysine for desferrioxamine biosynthesis. As a comparison, we predicted that a DABA DC homologue [YP_323346] encoded by A. variabilis ATCC 29413, which is found adjacent to a putative DABA AT ORF in a NIS-like biosynthetic gene cluster would be a bona fide DABA DC. The A. variablis putative DABA DC is the only DABA DC homologue in the genome and it is located in the only NIS biosynthetic gene cluster. The 1,3-diaminopropane-based NIS-type siderophore schizokinen (Fig. 1) is known to be produced by Anabaena species (Simpson and Nielands, 1976) and therefore it is a reasonable assumption that the A. variabilis DABA DC homologue is involved in producing the 1,3-diaminopropane component of schizokinen. DesA protein from S. coelicolor A3(2) and the putative DABA DC from A. variabilis exhibit 40% amino acid identity. The similarity of the reaction substrates and products for the two enzymes is shown in Fig. 2, with the glutamate to γ-aminobutyrate reaction of glutamate decarboxylase shown for comparison. The N-terminally T7-tagged S. coelicolor A3(2) DesA and A. variabilis DABA DC recombinant proteins were assayed in vitro with 5 mM of the substrates l-2,4-diaminobutyrate, l-lysine and l-ornithine, and the expected reaction products 1,3-diaminopropane, 1,5-diaminopentane (cadaverine) and 1,4-diaminobutane (putrescine) were detected by HPLC (Fig. 3). Activity of the A. variabilis putative DABA DC was greatest with l-2,4-diaminobutyrate; l-ornithine and l-lysine were decarboxylated less efficiently. The DesA protein displayed a marked preference for l-lysine but was also able to decarboxylate l-2,4-diaminobutyrate and l-ornithine at a lower rate.

figure

Figure 2. Reaction substrates and products of l-lysine, l-2,4-diaminobutyrate and l-glutamate decarboxylases.

Download figure to PowerPoint

figure

Figure 3. Reaction specificity of Anabaena variabilis and Streptomyces coelicolor DABA DC homologues. Products of reactions catalysed by DABA DC homologues from A. variabilis, [YP_323346] (A) and S. coelicolor, DesA (B), with l-ornithine (red), l-lysine (black) and l-2,4-diaminobutyrate (blue) as substrates, detected by HPLC. To ensure peaks were within the scale of the fluorescence detector, different amounts of each reaction were loaded onto the HPLC column and therefore the peak areas are not directly comparable. DAP, 1,3-diaminopropane; PUT, 1,4-diaminobutane (putrescine); CAD, 1,5-diaminopentane (cadaverine); FI, fluorescence intensity.

Download figure to PowerPoint

Kinetic characterization of S. coelicolor DesA and A. variabilis DABA DC

Usually with polyamine biosynthetic decarboxylases, such as ornithine and lysine decarboxylases, activity can be assayed using 14C-labelled amino acid substrate followed by trapping of the released 14C-labelled CO2. However, due to lack of commercially available 14C-labelled l-2,4-diaminobutyrate, the CO2 release assay was not possible with DABA DC. Kinetic constants of the S. coelicolor DesA and A. variabilis putative DABA DC were therefore determined by measuring the production of cadaverine and 1,3-diaminopropane using HPLC detection and quantification. For recombinant DesA, similar k cat values with l-lysine and l-2,4-diaminobutyrate were obtained; however, the K m for l-lysine was more than 20-fold lower than with l-2,4-diaminobutyrate (Table 1). In contrast, with the A. variabilis DABA DC, the K m for l-2,4-diaminobutyrate and l-lysine was similar but the k cat with l-2,4-diaminobutyrate was more than 40-fold higher than with l-lysine (Table 1). Although the S. coelicolor DesA and the A. variabilis DABA DC exhibit a relatively high level of sequence identity, the catalytic efficiency of DesA for l-lysine is 17-fold higher than with l-2,4-diaminobutyrate, whereas the catalytic efficiency of the DABA DC is 24-fold higher with l-2,4-diaminobutyrate than with l-lysine (Table 1). These results confirm that the S. coelicolor A3(2) desA gene encodes a novel LDC and the A. variabilis ATCC 29413 YP_323346 protein is a bona fide DABA DC.

Table 1. Steady-state kinetic parameters for Streptomyces coelicolor DesA and Anabaena variabilis DABA DC
EnzymeSubstrate k cat (s−1) K m (mM) k cat/ K m (M−1 s−1)
  1. Product formation (1,5-diaminopentane and 1,3-diaminopropane) was quantified by HPLC. Enzymatic assays were performed at 30°C. DABA, 2,4-diaminobutyrate.

S. coelicolor DesALysine0.040 ± 0.0030.058 ± 0.02690 ± 200
DABA0.052 ± 0.0031.3 ± 0.340 ± 10
A. variabilis DABA DCDABA0.340 ± 0.0300.54 ± 0.1630 ± 100
Lysine0.008 ± 0.0010.30 ± 0.126 ± 10

Polyamine contents of S. coelicolor and A. variabilis cells

In the case of the putrescine-based NIS molecules alcaligin (Brickman and Armstrong, 1996) and putrebactin (Kadi et al., 2008a) the biosynthetic gene cluster does not contain a putrescine biosynthetic module and therefore the putrescine unit must be recruited from cellular putrescine pools that contribute to normal polyamine biosynthesis. The presence of a cadaverine-synthesizing LDC in the S. coelicolor desferrioxamine biosynthetic gene cluster and a DABA AT/DABA DC 1,3-diaminopropane biosynthetic module in the A. variabilis NIS-type biosynthetic cluster suggests that neither species possesses cellular pools of either cadaverine or 1,3-diaminopropane, respectively, in iron-replete medium. If they did, the cellular pools would provide the cadaverine or 1,3-diaminopropane required for siderophore biosynthesis induced by iron scarcity, and dedicated LDC and DABA AT/DABA DC ORFs would not be required in the biosynthetic gene clusters. To test this hypothesis, S. coelicolor A3(2) and A. variabilis ATCC 29413 cells were grown in liquid, iron-replete growth media. The only polyamine detected in A. variabilis cells was the triamine sym-homospermidine (Fig. 4). A large peak of 1,3-diaminopropane and smaller peaks of putrescine and the triamine spermidine but no cadaverine were detected in stationary-phase S. coelicolor cells. This explains therefore why the S. coelicolor desferrioxamine biosynthetic gene cluster must carry a cadaverine biosynthetic module, and it also explains the need for a 1,3-diaminopropane biosynthetic module in the putative schizokinen biosynthetic gene cluster of A. variabilis. The 1,3-diaminopropane peak in the S. coelicolor extracts was unexpected. However, there is another DABA DC homologue in the S. coelicolor genome (NP_626278) adjacent to an N-monooxygenase suggesting a partial NIS biosynthetic cluster. There is also a DABA AT homologue in the genome (NP_626132), which is part of the ectoine biosynthetic cluster. Ectoine is a compatible solute containing a 2,4-diaminobutyrate moiety (Bursy et al., 2008) and the presence of these two physically unlinked genes encoding DABA AT and a putative DABA DC could explain the synthesis of 1, 3-diaminopropane in S. coelicolor.

figure

Figure 4. Polyamine contents in cells of A. variabilis ATCC 29413 and S. coelicolor A3(2).

A. Polyamines in 7-day-old culture of A. variabilis cells. HSPD, sym-homospermidine; R, fluorescence tag; IS, internal standard (diaminoheptane).

B. Polyamines in stationary-phase S. coelicolor cells. DAP, diaminopropane; PUT, putrescine; SPD, spermidine; R, fluorescence tag; IS, internal standard (diaminononane).

Download figure to PowerPoint

A novel lysine decarboxylase orthologue in Y. pestis

Based on the above evidence, the genomic presence of a DABA DC homologue in the absence of an adjacent DABA AT partner would appear to predict a novel LDC. Such a potential LDC gene is found in the γ-proteobacterium Y. pestis, the causative agent of pneumonic, septicaemic and bubonic plagues, within a NIS-type biosynthetic gene cluster similar to the S. coelicolor desferrioxamine operon. We decided to focus on the potential LDC from this species because desferrioxamine production has not been reported in this important pathogen. We expressed the Y. pestis putative LDC in Saccharomyces cerevisiae to assess whether cadaverine, a non-native polyamine in baker's yeast, is accumulated. In Fig. 5, it can be seen that in the parental S. cerevisiae strain 2602 only spermidine is accumulated. Accumulation of non-native cadaverine could potentially be masked in the parental strain due to aminopropylation by spermidine synthase to form N1-aminopropylcadaverine, which is not separated from spermidine by the analytical method used. To avoid aminopropylation of cadaverine, the Y. pestis putative LDC was expressed in a gene deletion strain of 2602 (Y362) in which S-adenosylmethionine decarboxylase and spermidine synthase had been deleted. Expression of the Y. pestis putative LDC ORF produced a distinct peak of cadaverine, confirming that it is another example of the novel LDC. No peak of cadaverine was detected after expression of the A. variabilis DABA DC in the yeast cells.

figure

Figure 5. Polyamines in S. cerevisiae strains expressing DABA DC homologues. HPLC chromatograms of polyamines detected by fluorescence. Black, wild type [parental (2602)] transformed with p427TEF empty vector; red, Y362 (ΔSPE1ΔSPE2) transformed with p427TEF empty vector; blue, Y362 expressing A. variabilis putative DABA DC [YP_323346]; green, Y362 expressing Y. pestis putative LDC [NP_405115]; orange, Y362 expressing S. coelicolor DesA [NP_627012]. Cad, cadaverine; Spd, spermidine; R, fluorescent labelling dye; IS, internal standard (diaminoheptane).

Download figure to PowerPoint

Genomic organization of 1,3-diaminopropane- and cadaverine-based NIS biosynthetic gene clusters

Identity of the genomic biosynthetic gene cluster for the 1,3-diaminopropane-based siderophore rhizobactin 1021 has been genetically confirmed in S. meliloti (Lynch et al., 2001). Biosynthetic gene clusters for the cadaverine-based siderophores desferrioxamine E (Barona-Gomez et al., 2004) and bisucabin (Kadi et al., 2008b) have been confirmed biochemically. We analysed bacterial and archaeal genomes for the presence of NIS-type biosynthetic gene clusters containing either a DABA DC homologue in the absence of an adjacent DABA AT (i.e. potentially a novel LDC) or a DABA AT/DABA DC adjacent pair. The confirmed biosynthetic gene clusters for cadaverine-based siderophores bisucaberin from Alteromonas haloplanktis and desferrixoamine E from S. coelicolor, and the 1,3-diaminopropane-based siderophore rhizobactin 1021 from R. meliloti are shown in Fig. 6, along with NIS biosynthetic gene clusters from other species. All clusters containing a juxtaposed DABA AT/DABA DC pair possess two synthetase ORFs, with only a low level of identity (20–25%) between the two synthetase proteins within the same biosynthetic gene cluster. The cyanobacteria A. variabilis and Synechococcus sp. PCC 7002 have an acyltransferase fused to the C-terminus of the second, downstream NIS synthetase ORF. In gene clusters containing a DesA-like LDC, many species have the acyltransferase fused to the N-terminus of the single NIS synthetase ORF. With both classes of NIS biosynthetic gene clusters (potentially cadaverine and 1,3-diaminopropane producing), a Major Facilitor Superfamily (MFS) transporter is often present in the gene cluster and probably represents the siderophore export transporter. In some of the DesA-like (cadaverine-type) gene clusters, an ORF of unknown function DUF1624, which appears to be related to acyltransferases, is present at different positions within the cluster in different species. All clusters contain an N-monooxygenase (N-hydroxylase), at least one acyltransferase and at least one NIS synthetase ORF.

figure

Figure 6. NIS biosynthetic gene clusters. N-MOX, N-monooxygenase; AcT, acyltransferase; iucA/C, NIS synthetase; MFS, Major Facilitator Superfamily (transporter); FeR, ferric iron reductase; DUF, domain of unknown function; DABA AT, l-2,4-diaminobutyrate: 2-ketoglutarate 4-aminotransferase; DABA DC, l-2,4-diaminobutyrate decarboxylase; LDC, DABA DC-like lysine decarboxylase. Protein accession numbers for the first and last ORFs in each cluster are presented below the cluster. Size of each predicted protein in amino acids shown below each ORF.

Download figure to PowerPoint

The similar configurations of ORFs within the NIS biosynthetic gene clusters, and conserved amino acid sequences of the constitutent ORFs in phylogenetically distant phyla suggest considerable horizontal gene transfer. NIS gene clusters very similar to the S. meliloti rhizobactin 1021 NIS gene cluster are found in a number of halobacterial species from the euryarchaeota and suggest the biosynthesis of a 1,3-diaminopropane-based siderophore such as rhizobactin 1021 or schizokinen in the some of the Halobacteria, including Haloferax volcanii, Haloterrigena turkmenica, Halobacterium mukohataei and Halobacterium sp. NRC-1. However, the NIS biosynthetic clusters in Halogeometricum boriquense and Haloalkalicoccus jeotagli have been disrupted by transposase ORFs. Currently, no hydroxamate-type siderophore has been described from archaeal sources.

When all DABA DC homologues are aligned and a Neighbor Joining phylogenetic tree is derived from the alignment, the potential novel LDC and bona fide DABA DC proteins separate into two distinct clades (Fig. 7). One clade contains only DABA DC homologues in the absence of an adjacent DABA AT but within a NIS biosynthetic gene cluster, whereas the other clade contains only DABA DC homologues with an adjacent DABA AT, although some of the DABA AT/DABA DC pairs are not located in a NIS biosynthetic gene cluster. An interesting example of a genome containing both putative novel LDC and DABA DC orthologues is Aliivibrio salmonicida LFI1238. The DABA DC orthologue of this species is fused to DABA AT and is found in an operon with carboxynorspermidine dehydrogenase and carboxynorspermidine decarboxylase, and the putative LDC is found in a NIS biosynthetic cluster. The LDC [YP_002261684] and DABA DC [YP_002263508] share 38% amino acid identity (180/469). Thus the novel LDCs can be distinguished from the bona fide DABA DC sequences both by sequence and by presence or absence of an adjacent DABA AT. It appears that all DABA DC homologues found without an adjacent DABA AT, i.e. desA homologues (novel LDCs), are found in NIS biosynthetic gene clusters. Species of Acinetobacter have a DABA AT/DABA DC pair separated from the rest of the NIS biosynthetic genes. In A. baumannii, only one NIS biosynthetic cluster is present, containing two NIS synthetase ORFs (A1S_1647 and A1S_1652), which are likely to be responsible for synthesis of the 1,3-diaminopropane-based siderophore acinetoferrin (Fig. 1). Other γ-Proteobacteria such as Actinobacillus, Moraxella and Haemopilus species have a DABA AT/DABA DC gene pair but do not possess a NIS synthetase.

figure

Figure 7. Neighbor Joining (NJ) tree of DABA DC homologues. Proteins in the red block are DABA DC homologues present in NIS biosynthetic gene clusters in the absence of DABA AT partners. Proteins in the blue blocks are DABA DC homologues adjacent to DABA AT ORFs in NIS biosynthetic gene clusters. Purple blocks are DABA DC homologues adjacent to DABA AT ORFs that are not part of NIS biosynthetic gene clusters. The NJ tree was constructed as previously described (Lee et al., 2007). Numbers on the tree represent percentage bootstrap support from 1000 replications.

Download figure to PowerPoint

A comparison of all sequenced genomes from Firmicutes species indicates that DABA AT- and DABA DC-encoding ORFs are not found in the same genomes as ORFs encoding carboxynorspermidine dehydrogenase and carboxynorspermidine decarboxylase, except for Bacillus halodurans C-125. Even in B. halodurans C-125, the DABA AT/DABA DC ORFs are found within a NIS biosynthetic gene cluster, suggesting that the 1,3-diaminopropane produced is incorporated in a siderophore in this species. This indicates that for most species outside of the Vibrionales, the DABA AT/DABA DC module is used to synthesize 1,3-diaminopropane that is not incorporated into norspermidine. It also means that the carboxy(nor)spermidine dehydrogenase/carboxy(nor)spermidine decarboxylase module is used to synthesize spermidine rather than norspermidine in species outside of the Vibrionales.

Comparative structural analysis of the novel lysine decarboxylase

A CLANS (Cluster Analysis of Sequences) (Frickey and Lupas, 2004) analysis of the novel LDC sequences indicates that besides the bona fide DABA DCs, which are the closest related sequence, other sequences related to the DesA LDC are pyridoxal-5′-phosphate-dependent human glutamate decarboxylase-related sequences, cysteine sulphinic acid decarboxylase- and DOPA decarboxylase-related sequences (Fig. 8). An X-ray crystal structure is available for the DABA DC domain (PDB: 2qma) of the Vibrio parahaemolyticus DABA AT-DABA DC fusion protein, although there is no associated publication. Due to the high sequence identity between the DABA DC and novel LDC, it is very probable that the LDC would have a very similar structure to DABA DC. A comparison of the V. parahaemolyticus DABA DC structure with closely related structures (Fig. 9) identified by VAST (Vector Alignment Search Tool) analysis (Gibrat et al., 1996) reveals clearly the similarity between the V. parahaemolyticus DABA DC and human glutamate decarboxylase and cysteine sulphinic acid decarboxylase structures. The substrates and products of the DABA DC and glutamate decarboxylases are shown in Fig. 2. Other very similar structures include the pig DOPA decarboxylase, archaeal tyrosine decarboxylase and plant and E. coli glutamate decarboxylases. Although DABA DC has no sequence similarity to the acid inducible LDC CadA, both the DABA DC (PDB: 2qma) and CadA LDC (PDB: 3q16) belong to the aspartate aminotransferase protein fold superfamily (Fig. 9). The DABA DC, and therefore the DesA-like LDC on the one hand, and the CadA acid inducible LDC on the other, likely have a distant common structural ancestor but they have evolved substrate specificity for lysine independently and therefore represent convergent evolution for lysine decarboxylation. A pyridoxal-5′-phosphate-dependent lysine decarboxylase has also evolved from an entirely unrelated protein fold, the lysine/ornithine decarboxylase (PDB: 2plj) from the alanine racemase family (Fig. 9).

figure

Figure 8. CLANS visualization in two-dimensional space of the pairwise sequence similarities of DABA DC-related sequences. The PDB (Protein Data Bank) identities of the structures shown in Fig. 9 are indicated. Dots represent each sequence, lines connecting the dots are based on pairwise blast similarity (cut-off E-value 0.001). Connections are coloured in grey scale with black being the closest match. Sequences are clustered in 2D according to the blast similarity.

Download figure to PowerPoint

figure

Figure 9. Protein structures similar to l-2,4-diaminobutyrate decarboxylase. Structure 2plj is the biosynthetic lysine/ornithine decarboxylase of Vibrio vulnificus and part of the PLP-dependent alanine racemase family and is unrelated to the other decarboxylase structures. Other decarboxylase structures are members of the PLP-dependent aspartate aminotransferase family, including the glutamate decarboxylase and acid-inducible ornithine decarboxylase major domain subfamilies. The N-terminal wing domains (residues 1–131 for LDC, 1–109 for ODC), consisting of a response regulator REC receiver domain (Burrell et al., 2010) of the Escherichia coli lysine (3q16) and Lactobacillus sp. 30a ornithine (1ord) decarboxylases have been removed to facilitate the comparison. Glutamate decarboxylase family structures: 2qma, 2,4-diaminobutyrate decarboxylase (closely related at the amino acid sequence level to the novel LDC) from Vibrio parahaemolyticus; 2okk, glutamate decarboxylase from human; 2jis, cysteine sulphinic acid decarboxylase from human; 1js3, DOPA decarboxylase from pig; 1pmm, glutamate decarboxylase from E. coli; 3hbx, glutamate decarboxylase from Arabidopsis thaliana; 3f9t, tyrosine decarboxylase from Methanocaldococcus jannaschii. The LDC (3q16) and ODC (1ord) major domain structures were compared with the glutamate decarboxylase family structures using DaliLite (Holm et al., 2008). All pairwise Dali Z-scores were normalized to the higher of the two self-scores, and the normalized scores were converted to distances using –ln(X). Trees were generated using these pairwise distances with the fitch program of the PHYLIP package (Felsenstein, 1997), with global optimization.

Download figure to PowerPoint

Discussion

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References

The DABA DC homologue encoded by the desA gene of S. coelicolor has been shown here to preferentially decarboxylate l-lysine rather than l-2,4-diaminobutyrate, and closely homologous genes appear to be confined to bacterial NIS biosynthetic gene clusters. Evolution of this novel form of LDC (DesA) therefore appears to be intimately linked to the evolution of hydroxamate-type NIS biosynthetic pathways. The two other known forms of LDC, the aspartate aminotransferase-fold decameric form (CadA) responsible for acid resistance in E. coli (Kanjee et al., 2011), and the dimeric alanine racemase-fold bifunctional ornithine/lysine decarboxylase (Lee et al., 2007), which have evolved from core polyamine biosynthetic enzymes (Lee et al., 2007; Burrell et al., 2010), have no sequence similarity to the DesA LDC. Together, they represent three examples of what has been described as non-homologous isofunctional enzymes (Omelchenko et al., 2010), i.e. proteins with different folds catalysing the same reaction. The DesA and CadA proteins although distant, have in fact evolved from a common ancestral protein as both are part of the aspartate aminotransferase fold superfamily. The three distinct forms of LDC are all pyridoxal-5′-phosphate-dependent enzymes. In the case of the polyamine biosynthetic enzyme arginine decarboxylase, there are four protein folds encoding arginine decarboxylase enzymes, two are pyridoxal-5′-phosphate dependent and two are pyruvoyl cofactor dependent (Burrell et al., 2010).

The DesA LDC, and the bona fide DABA DC proteins appear to have evolved more recently from the l-glutamate decarboxylase (GadB) family, which is an acid resistance-associated enzyme in bacteria (Richard and Foster, 2003). Substantial sequence identity between the novel DesA-like LDCs and the bona fide DABA DC proteins, compared with a much lower level of sequence of identity of both to glutamate decarboxylases suggests that they did not evolve independently from glutamate decarboxylase but from a more recent ancestor. It is unclear whether the LDC evolved from DABA DC or vice versa. However, it is relevant that DABA AT homologues are found independently of DABA DC in many bacterial pathways where l-2,4-diaminobutyrate, the product of DABA AT, is a biosynthetic intermediate. For example, DABA AT is required for the synthesis of the compatible solute ectoine (Reshetnikov et al., 2011a,b), in the non-ribosomal peptide synthetase-dependent pathway for synthesis of the siderophore pyoverdine (Vandenende et al., 2004), and in the NIS biosynthetic pathway for the hydroxycarboxylate siderophore achromobactin (Berti and Thomas, 2009; Schmelz et al., 2009; 2011; Owen and Ackerley, 2011). This is in marked contrast to bona fide DABA DC orthologues which are not found independently of either DABA AT or hydroxamate-type NIS biosynthetic clusters. The DABA AT/DABA DC pair is found independently of NIS clusters in a number of bacterial species, especially within the Vibrionales, where the two ORFs are fused and adjacent to ORFs encoding carboxynorspermidine dehydrogenase and carboxynorspermidine decarboxylase. Together these enzymes synthesize norspermidine from 1,3-diaminopropane (Lee et al., 2009). In V. cholerae, norspermidine is used in the synthesis of the non-ribosomal peptide synthetase-dependent siderophore vibrobactin (Keating et al., 2000).

In the absence of DABA AT, there will be no substrate for DABA DC and so the NIS biosynthetic gene clusters containing a DABA DC homologue but lacking DABA AT are likely to encode an LDC. We correctly predicted that Y. pestis contained a DesA-like LDC because no DABA AT homologue was present in the Y. pestis NIS biosynthetic cluster. The more parsimonious explanation for the evolution of the DABA DC and the DABA DC-like LDC is that the bona fide DABA DC evolved from the LDC because DABA DC in the absence of DABA AT has no substrate. Inherent substrate promiscuity of the DesA LDC towards l-2,4-diaminobutyrate may have been selected for resulting in an enzyme that evolved into a specific DABA DC.

The fish pathogen A. salmonicida possesses a gene pair encoding DABA AT and DABA DC that is likely involved in norspermidine biosynthesis. It also has a DABA DC homologue in a NIS biosynthetic cluster that is likely to be a DesA-like LDC. There is only 38% amino acid identity between the two DABA DC-like homologues, suggesting that at least one of the copies has been acquired relatively recently by horizontal gene transfer, rather than by gene duplication of an ancestral DABA DC homologue. In the human pathogen Acinetobacter baumanii, the DABA DC/DABA AT genes, which have been biochemically characterized (Yamamoto et al., 1991; Ikai and Yamamoto, 1997), are physically juxtaposed but distant from the sole NIS biosynthetic gene cluster. It is very probable that the DABA AT/DABA DC gene pair are responsible for synthesizing the 1,3-diaminopropane components of the A. baumannii amphiphilic siderophore acinetoferrin (Okujo et al., 1994) (Fig. 1). Physical separation of the gene pair from the rest of the NIS biosynthetic cluster may explain why 1,3-diaminopropane is accumulated to a relatively high level in A. baumannii, a species which seems to synthesize no polyamine other than 1,3-diaminopropane (Auling et al., 1991; Hamana and Matsuzaki, 1992). Transposon insertion mutants of the DABA AT- and DABA DC-encoding genes of A. baumannii no longer exhibit surface-associated motility and do not make 1,3-diaminopropane (Skiebe et al., 2012). When A. baumannii is grown is iron-depleted media, its only NIS gene cluster, containing the NIS synthetases A1S_1647 and A1S_1652, which are likely responsible for acinetoferrin biosynthesis, is transcriptionally upregulated more than 25-fold, coinciding with a loss of surface-associated motility (Eijkelkamp et al., 2011). Thus 1,3-diaminopropane and subsequent acinetoferrin biosynthesis may be necessary for surface-associated motility, an important virulence determinant in A. baumannii (Skiebe et al., 2012).

The recruitment of genes encoding DABA AT and DABA DC and the DesA-like LDC to NIS biosynthesis and corresponding biosynthetic gene clusters was a key factor in generating bacterial siderophore structural diversity. Putrescine (1,4-diaminobutane) is widely distributed in bacteria but putrescine-based siderophores are less common than 1,3-diaminopropane- or cadaverine-based siderophores. In E. coli, glutamate and aspartate, which are involved in 1,3-diaminopropane biosynthesis, and lysine (for cadaverine biosynthesis) are orders of magnitude more abundant than ornithine used in putrescine biosynthesis (Bennett et al., 2009). It may be that substrate availability for diamine biosynthesis was a driving force in diversifying siderophore structures. Why isn't putrescine incorporated into the 1,3-diaminopropane- or cadaverine-based siderophores, considering that it is ubiquitous? It is likely that the gate-keeping function conferring selectivity on diamine incorporation into the siderophore is performed by the N-monooxygenase enzyme that N-hydroxylates the diamine, since lysine and ornithine N-hydroxylases involved in biosynthesis of lysine and ornithine-based siderophores exhibit a highly specific selection of lysine over ornithine or vice versa (Olucha and Lamb, 2011; Olucha et al., 2011).

The identification of a novel class of DABA DC-like LDC enzymes involved in NIS-type siderophore biosynthesis allows us to predict from a given genome sequence the likely NIS structural backbone (cadaverine, 1,3-diaminopropane or neither of these) that will be synthesized in the corresponding strain. We previously noted that outside of the Vibrionales the distribution of bacterial genomes encoding DABA AT and DABA DC did not overlap much with those encoding carboxy(nor)spermidine dehydrogenase and carboxy(nor)spermidine decarboxylase (Lee et al., 2009), suggesting that in many species 1,3-diaminopropane is not being used to synthesize norspermidine. Our current results indicate that in numerous and diverse bacteria 1,3-diaminopropane is synthesized to provide the structural backbone of NIS-type siderophores.

Experimental procedures

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References

Bacterial strains and growth media

Cyanobacterium A. variabilis ATCC 29413 was obtained from the American Type Culture Collection. Cultures of A. variabilis ATCC 29413 were grown in ATCC liquid medium #616 (BG-11) at 25°C under lights (16h light/8h dark) with constant shaking. Actinobacterium S. coelicolor A3(2) was kindly supplied by Dr Mark Buttner, John Innes Centre, Norwich, UK, and was grown on NMMP medium (Hodgson, 1982) containing 1.5% (w/v) agarose (agar contains polyamines) at 30°C until sporulation occurred. Spores were resuspended in 20% (v/v) glycerol and filtered through sterile cotton wool to remove mycelial fragments. For growth in liquid culture, spores were first pre-germinated by incubating in NMMP lacking casamino acids, then harvested by centrifugation, followed by resuspension in the same medium and then sonication for 5 s. Germinated spores were used to inoculate 30 ml of NMMP medium lacking casamino acids, in a 250 ml Erlenmayer flask containing a 30 cm length of stainless steel spring to act as a baffle, and cultures were grown at 30°C with shaking at 200 r.p.m.

Yeast strains and methods

The yeast strains used were the parental haploid strain 2602 (MATα ura3-52 his6 leu2) and a derived haploid S-adenosylmethionine decarboxylase (SPE2)/ornithine decarboxylase (SPE1) double knockout strain Y362 (MATα ura3-52 his6 leu2 SPE1 SPE2) previously described (Balasundaram et al., 1994). Bacterial genes were expressed from a p427TEF yeast expression plasmid (Dualsystems Biotech), which were introduced into yeast using a modified lithium acetate method (Elbe, 1992). Cultures for polyamine analysis were grown in polyamine-free liquid SD medium from an initial OD600 of 0.01 at 30°C with shaking at 200 r.p.m.

Protein expression and purification

Genes encoding a putative DABA DC from A. variabilis [GenBank protein Accession No. YP_323346] and putative LDC enzymes from S. coelicolor [NP_627012] and Y. pestis [NP_405115] were synthesized by Genscript (Piscataway, NJ), with codons optimized for expression in E. coli. The genes were subcloned into the BamHI and XhoI sites of either pET45b (A. variabilis DABA DC, Y. pestis DABA DC/LDC) or pET21a (S. coelicolor DABA DC/LDC), allowing expression of the proteins with N-terminal His- or T7-tags.

For expression of putative decarboxylases, E. coli BL21 was transformed with protein expression vectors and grown to an OD600 of 0.3 in LB medium containing 100 μg ml−1 ampicillin at 37°C. Protein expression was induced with 0.4 mM IPTG and cultures were incubated for a further 3 h at 37°C. For purification of his-tagged proteins, cells were resuspended in 20 mM sodium phosphate (pH 8.0) containing 500 mM NaCl, 20 mM imidazole and 0.02% (v/v) Brij35 before being broken by sonication. The cell lysate was cleared by centrifugation and applied to a HiTrap™ Chelating HP (GE Healthcare) that had been charged with Ni2+ and equilibrated with the above buffer. The column was washed with the above buffer and bound protein was eluted using a 0.02–1.00 M imidazole gradient. The T7-tagged S. coelicolor LDC was purified using the T7·Tag Affinity Purification Kit (Novagen) according to the manufacturer's instructions. Purified proteins were concentrated using Amicon Ultra-4 Centrifugal Filter Units, buffer exchanged with 20 mM Tris·HCl (pH 7.5) containing 20 % (v/v) glycerol and 2 mM dithiothreitol and stored at −80°C.

Enzyme assays

Decarboxylase activity was quantified by HPLC. Unless otherwise stated, assays were buffered in 50 mM HEPES (pH 7.5) containing 50 mM NaCl, 2 mM dithiothreitol and 0.1 mM pyridoxal-5′-phosphate and contained 0.08–10 mM substrate. Reactions were incubated at 30°C for 10 min before being stopped by the addition of 5% (v/v) trichloroacetic acid (TCA). For HPLC analysis, polyamines were labelled using the AccQ-Fluor™ reagent kit (Waters Coporation, Milford, MA) according to the manufacturer's instructions. Labelling reactions contained 5 μl of stopped enzyme assay and 1.25 μM diaminoheptane or diaminononane as an internal standard and were heated to 55°C for 10 min. Derivatized polyamine samples were analysed by HPLC using a reverse-phase C18 column (Phenomenex, Luna 5 μ) on a Dionex Summit HPLC System. Polyamine separation was performed using 10 μl of derivatized sample. The system was operated at 33°C and equilibrated with Eluent A (70 mM acetic acid, 25 mM triethlyamine, pH 4.82) at 1.2 ml min−1. Elution was performed using the following linear gradients of Eluent B (80% acetonitrile): 22% for 5 min, 39% for 12 min with 6% methanol, 33% for 30 s with 14% methanol, 10% for 6.5 min with 70% methanol and finally 100% for 21 min. Polyamines were monitored by fluorescence (Dionex RF 2000 detector) with a 248 nm excitation filter and a 398 nm emission filter, and identified by comparison of retention times with known standards which were derivatized and analysed in parallel with the enzyme assays.

Analysis of cellular polyamines

For HPLC analysis of bacterial cellular polyamines, cells were washed in 20 mM Tris·HCl (pH 7.5) and resuspended in 100 mM MOPS containing 50 mM NaCl, 20 mM MgCl2 before being broken by three 15 s sonication pulses. Protein was precipitated by the addition of 10% (v/v) TCA followed by 5 min incubation on ice and the lysate was cleared of cell debris by centrifugation. Polyamines were analysed as described above. For extraction of polyamines from S. cerevisiae, cells were harvested by centrifugation, washed twice in 1 ml of phosphate-buffered saline, and polyamines extracted with 3 μl of 5% TCA per mg of cells fresh weight. Extractions were performed overnight at 4°C, then extracts were centrifuged (12 000 g, 5 min, 4°C) and 5 μl of supernatant flourescently labelled using the AccQ-Fluor™ reagent kit (Waters).

Acknowledgements

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References

This work was supported by a Core Strategic Grant from the Biotechnological and Biological Sciences Research Council to the Institute of Food Research, and by UT Southwestern Medical Center. We would like to thank Anthony E. Pegg for helpful comments on the manuscript.

References

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References
  • Auling, G., Pilz, F., Busse, H.J., Karrasch, S., Streichan, M., and Schon, G. (1991) Analysis of the polyphosphate-accumulating microflora in phosphorus-eliminating, anaerobic-aerobic activated sludge systems by using diaminopropane as a biomarker for rapid estimation of Acinetobacter spp. Appl Environ Microbiol 57: 35853592.
  • Balasundaram, D., Xie, Q.W., Tabor, C.W., and Tabor, H. (1994) The presence of an active S-adenosylmethionine decarboxylase gene increases the growth defect observed in Saccharomyces cerevisiae mutants unable to synthesize putrescine, spermidine, and spermine. J Bacteriol 176: 64076409.
  • Barona-Gomez, F., Wong, U., Giannakopulos, A.E., Derrick, P.J., and Challis, G.L. (2004) Identification of a cluster of genes that directs desferrioxamine biosynthesis in Streptomyces coelicolor M145. J Am Chem Soc 126: 1628216283.
  • Barry, S.M., and Challis, G.L. (2009) Recent advances in siderophore biosynthesis. Curr Opin Chem Biol 13: 205215.
  • Bennett, B.D., Kimball, E.H., Gao, M., Osterhout, R., Van Dien, S.J., and Rabinowitz, J.D. (2009) Absolute metabolite concentrations and implied enzyme active site occupancy in Escherichia coli . Nat Chem Biol 5: 593599.
  • Berti, A.D., and Thomas, M.G. (2009) Analysis of achromobactin biosynthesis by Pseudomonas syringae pv. syringae B728a. J Bacteriol 191: 45944604.
  • Brickman, T.J., and Armstrong, S.K. (1996) The ornithine decarboxylase gene odc is required for alcaligin siderophore biosynthesis in Bordetella spp.: putrescine is a precursor of alcaligin. J Bacteriol 178: 5460.
  • Burrell, M., Hanfrey, C.C., Murray, E.J., Stanley-Wall, N.R., and Michael, A.J. (2010) Evolution and multiplicity of arginine decarboxylases in polyamine biosynthesis and essential role in Bacillus subtilis biofilm formation. J Biol Chem 285: 3922439238.
  • Bursy, J., Kuhlmann, A.U., Pittelkow, M., Hartmann, H., Jebbar, M., Pierik, A.J., and Bremer, E. (2008) Synthesis and uptake of the compatible solutes ectoine and 5-hydroxyectoine by Streptomyces coelicolor A3(2) in response to salt and heat stresses. Appl Environ Microbiol 74: 72867296.
  • Challis, G.L. (2005) A widely distributed bacterial pathway for siderophore biosynthesis independent of nonribosomal peptide synthetases. Chembiochem 6: 601611.
  • Coy, M., Paw, B.H., Bindereif, A., and Neilands, J.B. (1986) Isolation and properties of N epsilon-hydroxylysine:acetyl coenzyme A N epsilon-transacetylase from Escherichia coli pABN11. Biochemistry 25: 24852489.
  • Eijkelkamp, B.A., Hassan, K.A., Paulsen, I.T., and Brown, M.H. (2011) Investigation of the human pathogen Acinetobacter baumannii under iron limiting conditions. BMC Genomics 12: 126.
  • Elbe, R. (1992) A simple and efficient procedure for transformation of yeasts. Biotechniques 13: 1820.
  • Felsenstein, J. (1997) An alternating least squares approach to inferring phylogenies from pairwise distances. Syst Biol 46: 101111.
  • Frickey, T., and Lupas, A. (2004) CLANS: a Java application for visualizing protein families based on pairwise similarity. Bioinformatics 20: 37023704.
  • Gibrat, J.F., Madej, T., and Bryant, S.H. (1996) Surprising similarities in structure comparison. Curr Opin Struct Biol 6: 377385.
  • Hamana, K., and Matsuzaki, S. (1992) Diaminopropane occurs ubiquitously in Acinetobacter as the major polyamine. J Gen Microbiol 38: 191194.
  • Hodgson, D. (1982) Glucose repression of carbon source uptake in Strpetomyces coelicolor A3(2) and its perturbation in mutants resistant to 2-deoxyglucose. J Gen Microbiol 128: 24172430.
  • Holm, L., Kaariainen, S., Rosenstrom, P., and Schenkel, A. (2008) Searching protein structure databases with DaliLite v.3. Bioinformatics 24: 27802781.
  • Ikai, H., and Yamamoto, S. (1994) Cloning and expression in Escherichia coli of the gene encoding a novel L-2,4-diaminobutyrate decarboxylase of Acinetobacter baumannii . FEMS Microbiol Lett 124: 225228.
  • Ikai, H., and Yamamoto, S. (1997) Identification and analysis of a gene encoding L-2,4-diaminobutyrate:2-ketoglutarate 4-aminotransferase involved in the 1,3-diaminopropane production pathway in Acinetobacter baumannii . J Bacteriol 179: 51185125.
  • Ikai, H., and Yamamoto, S. (1998) Two genes involved in the 1,3-diaminopropane production pathway in Haemophilus influenzae . Biol Pharm Bull 21: 170173.
  • Ito, Y., and Butler, A. (2005) Structure of synechobactins, new siderophores of the marine cyanobacterium Synechococcus sp. PCC 7002. Limnol Oceanogr 50: 19181923.
  • Kadi, N., Arbache, S., Song, L., Oves-Costales, D., and Challis, G.L. (2008a) Identification of a gene cluster that directs putrebactin biosynthesis in Shewanella species: PubC catalyzes cyclodimerization of N-hydroxy-N-succinylputrescine. J Am Chem Soc 130: 1045810459.
  • Kadi, N., Song, L., and Challis, G.L. (2008b) Bisucaberin biosynthesis: an adenylating domain of the BibC multi-enzyme catalyzes cyclodimerization of N-hydroxy-N-succinylcadaverine. Chem Commun (Camb) 41: 51195121.
  • Kanjee, U., Gutsche, I., Alexopoulos, E., Zhao, B., Bakkouri, M.E., Thibault, G., et al. (2011) Linkage between the bacterial acid stress and stringent responses: the structure of the inducible lysine decarboxylase. EMBO J 30: 931944.
  • Keating, T.A., Marshall, C.G., and Walsh, C.T. (2000) Vibriobactin biosynthesis in Vibrio cholerae: VibH is an amide synthase homologous to nonribosomal peptide synthetase condensation domains. Biochemistry 39: 1551315521.
  • Koglin, A., and Walsh, C.T. (2009) Structural insights into nonribosomal peptide enzymatic assembly lines. Nat Prod Rep 26: 9871000.
  • Lee, J., Michael, A.J., Martynowski, D., Goldsmith, E.J., and Phillips, M.A. (2007) Phylogenetic diversity and the structural basis of substrate specificity in the beta/alpha-barrel fold basic amino acid decarboxylases. J Biol Chem 282: 2711527125.
  • Lee, J., Sperandio, V., Frantz, D.E., Longgood, J., Camilli, A., Phillips, M.A., and Michael, A.J. (2009) An alternative polyamine biosynthetic pathway is widespread in bacteria and essential for biofilm formation in Vibrio cholerae . J Biol Chem 284: 98999907.
  • Linke, W.-D., Crueger, A., and Diekmann, H. (1972) Stoffwechselproduckte von Mikroorganismen. 106. Mitteilung*. Zur Konstitution des Terregens-Faktors. Arch Microbiol 85: 4450.
  • de Lorenzo, V., Bindereif, A., Paw, B.H., and Neilands, J.B. (1986) Aerobactin biosynthesis and transport genes of plasmid ColV-K30 in Escherichia coli K-12. J Bacteriol 165: 570578.
  • Lynch, D., O'Brien, J., Welch, T., Clarke, P., Cuiv, P.O., Crosa, J.H., and O'Connell, M. (2001) Genetic organization of the region encoding regulation, biosynthesis, and transport of rhizobactin 1021, a siderophore produced by Sinorhizobium meliloti . J Bacteriol 183: 25762585.
  • Miethke, M., and Marahiel, M.A. (2007) Siderophore-based iron acquisition and pathogen control. Microbiol Mol Biol Rev 71: 413451.
  • Mullis, K.B., Pollack, J.R., and Neilands, J.B. (1971) Structure of schizokinen, an iron-transport compound from Bacillus megaterium . Biochemistry 10: 48944898.
  • Okujo, N., Sakakibara, Y., Yoshida, T., and Yamamoto, S. (1994) Structure of acinetoferrin, a new citrate-based dihydroxamate siderophore from Acinetobacter haemolyticus. Biometals 7: 170176.
  • Olucha, J., and Lamb, A.L. (2011) Mechanistic and structural studies of the N-hydroxylating flavoprotein monooxygenases. Bioorg Chem 39: 171177.
  • Olucha, J., Meneely, K.M., Chilton, A.S., and Lamb, A.L. (2011) Two structures of an N-hydroxylating flavoprotein monooxygenase: ornithine hydroxylase from Pseudomonas aeruginosa . J Biol Chem 286: 3178931798.
  • Omelchenko, M.V., Galperin, M.Y., Wolf, Y.I., and Koonin, E.V. (2010) Non-homologous isofunctional enzymes: a systematic analysis of alternative solutions in enzyme evolution. Biol Direct 5: 31.
  • Oves-Costales, D., Kadi, N., and Challis, G.L. (2009) The long-overlooked enzymology of a nonribosomal peptide synthetase-independent pathway for virulence-conferring siderophore biosynthesis. Chem Commun (Camb) 43: 65306541.
  • Owen, J.G., and Ackerley, D.F. (2011) Characterization of pyoverdine and achromobactin in Pseudomonas syringae pv. phaseolicola 1448a. BMC Microbiol 11: 218.
  • Reshetnikov, A.S., Khmelenina, V.N., Mustakhimov, I.I., Kalyuzhnaya, M., Lidstrom, M., and Trotsenko, Y.A. (2011a) Diversity and phylogeny of the ectoine biosynthesis genes in aerobic, moderately halophilic methylotrophic bacteria. Extremophiles 15: 653663.
  • Reshetnikov, A.S., Khmelenina, V.N., Mustakhimov, I.I., and Trotsenko, Y.A. (2011b) Genes and enzymes of ectoine biosynthesis in halotolerant methanotrophs. Methods Enzymol 495: 1530.
  • Richard, H.T., and Foster, J.W. (2003) Acid resistance in Escherichia coli . Adv Appl Microbiol 52: 167186.
  • Schaible, U.E., and Kaufmann, S.H. (2004) Iron and microbial infection. Nat Rev Microbiol 2: 946953.
  • Schmelz, S., Kadi, N., McMahon, S.A., Song, L., Oves-Costales, D., Oke, M., et al. (2009) AcsD catalyzes enantioselective citrate desymmetrization in siderophore biosynthesis. Nat Chem Biol 5: 174182.
  • Schmelz, S., Botting, C.H., Song, L., Kadi, N.F., Challis, G.L., and Naismith, J.H. (2011) Structural basis for acyl acceptor specificity in the achromobactin biosynthetic enzyme AcsD. J Mol Biol 412: 495504.
  • Simpson, F.B., and Nielands, J.B. (1976) Siderophores in cyanophyceae: isolation and characterization of schizokinen from Anabaena sp. J Phycol 12: 4448.
  • Skiebe, E., de Berardinis, V., Morczinek, P., Kerrinnes, T., Faber, F., Lepka, D., et al. (2012) Surface-associated motility, a common trait of clinical isolates of Acinetobacter baumannii, depends on 1,3-diaminopropane. Int J Med Microbiol 302: 117128.
  • Strieker, M., Tanovic, A., and Marahiel, M.A. (2010) Nonribosomal peptide synthetases: structures and dynamics. Curr Opin Struct Biol 20: 234240.
  • Tunca, S., Barreiro, C., Sola-Landa, A., Coque, J.J., and Martin, J.F. (2007) Transcriptional regulation of the desferrioxamine gene cluster of Streptomyces coelicolor is mediated by binding of DmdR1 to an iron box in the promoter of the desA gene. FEBS J 274: 11101122.
  • Vandenende, C.S., Vlasschaert, M., and Seah, S.Y. (2004) Functional characterization of an aminotransferase required for pyoverdine siderophore biosynthesis in Pseudomonas aeruginosa PAO1. J Bacteriol 186: 55965602.
  • Yamamoto, S., Tsuzaki, Y., Kishi, R., and Nakao, H. (1991) Occurrence of L-2,4-diaminobutyrate decarboxylase activity in Acinetobacter . Chem Pharm Bull (Tokyo) 39: 24512453.
  • Yamamoto, S., Mutoh, N., Tsuzuki, D., Ikai, H., Nakao, H., Shinoda, S., et al. (2000) Cloning and characterization of the ddc homolog encoding L-2,4-diaminobutyrate decarboxylase in Enterobacter aerogenes . Biol Pharm Bull 23: 649653.