Identification and characterization of a carnitine transporter in Acinetobacter baumannii

Abstract The opportunistic pathogen Acinetobacter baumannii is able to grow on carnitine. The genes encoding the pathway for carnitine degradation to the intermediate malic acid are known but the transporter mediating carnitine uptake remained to be identified. The open reading frame HMPREF0010_01347 (aci01347) of Acinetobacter baumannii is annotated as a gene encoding a potential transporter of the betaine/choline/carnitine transporter (BCCT) family. To study the physiological function of Aci01347, the gene was deleted from A. baumannii ATCC 19606. The mutant was no longer able to grow on carnitine as sole carbon and energy source demonstrating the importance of this transporter for carnitine metabolism. Aci01347 was produced in Escherichia coli MKH13, a strain devoid of any compatible solute transporter, and the recombinant E. coli MKH13 strain was found to take up carnitine in an energy‐dependent fashion. Aci01347 also transported choline, a compound known to be accumulated under osmotic stress. Choline transport was osmolarity‐independent which is consistent with the absence of an extended C‐terminus found in osmo‐activated BCCT. We propose that the Aci01347 is the carnitine transporter mediating the first step in the growth of A. baumannii on carnitine.

choline oxidation to glycine betaine has an energetic benefit for the cell. More important, the product of the reaction sequence, glycine betaine, is a known osmoprotectant. Therefore, the expression of the choline transporter and the choline oxidation pathway genes is regulated by the transcriptional regulator BetI (Sand et al., 2014, Scholz et al., 2016 via the choline and salt concentration of the environment. Thus, induction of choline uptake and oxidation is induced by the presence of choline in the medium and in addition, by high salt (Sand et al., 2014).
Glycine betaine is used by A. baylyi (Sand et al., 2011) and A. baumannii (Zeidler et al., 2018) as compatible solute. It cannot be synthesized de novo but from the precursor choline, as outlined above. In addition, both species can take up glycine betaine from the medium (Sand et al., 2011;Zeidler et al., 2018) and uptake is preferred over synthesis from choline for energetic reasons (Oren, 1999, Sand et al., 2014. In contrast to, for example, Pseudomonas aeruginosa (Lisa, Garrido, & Domenech, 1983), glycine betaine is not further metabolized. When the osmolarity in the ecosystem decreases, glycine betaine is expelled by yet to be identified proteins.
Under anaerobic conditions, Escherichia coli can oxidize carnitine to γ-butyrobetaine that is expelled from the cell by a carnitine:γ-butyrobetaine antiporter, CaiT (Schulze, Köster, Geldmacher, Terwisscha Van Scheltinga, & Kühlbrandt, 2010). In contrast, Acinetobacter calcoaceticus uses carnitine as sole carbon and energy source. The first step is the splitting of the C-N bond in carnitine leading to the release of trimethylamine and malic acid semialdehyde. Subsequent oxidation gives malic acid, a central intermediate of the citric acid cycle that is oxidized to CO 2 (Kleber et al., 1977;Seim, Löster, & Kleber, 1982). Recently, mining for carnitine-degrading enzymes in reference genomes of a Human Microbiome Project (HMP) led to the detection of a putative carnitine utilization gene cluster in Acinetobacter ssp. (Zhu et al., 2014). Mutant studies of A. baumannii ATCC 19606 led to the identification of the carnitine oxygenase genes cntA and cntB, required for the oxidation of carnitine to malic acid (Zhu et al., 2014). The transporter catalyzing carnitine uptake had not been described yet but close to cntA/B is a gene (aci01347) encoding a potential transporter of the betaine/choline/carnitine transporter (BCCT) family. Indeed, this gene was suggested by Zhu et al. to encode a carnitine transporter. Herein, we provide genetic and biochemical evidence that Aci01347 of A. baumannii is a carnitine transporter essential for growth on carnitine.

| Bacterial strains and growth conditions
Escherichia coli MKH13 strains were grown in LB medium (Bertani, 1951) at 37°C with 100 µg/ml ampicillin. A. baumannii strain ATCC 19606 was grown at 37°C in LB medium (Bertani, 1951) (Tschech & Pfennig, 1984) with 20 mM sodium acetate or carnitine as carbon source. The role of choline and carnitine as compatible solutes was addressed by growth experiments in mineral medium with 300 mM NaCl in the presence of 1 mM choline and carnitine, respectively. Kanamycin (50 µg/ml) or gentamicin (100 µg/ml) was added from stock solutions when appropriate. Growth was monitored by measurement of the optical density at 600 nm. The growth experiments were repeated 3 times, and one representative experiment is shown. Growth curves were fitted manually.

| Cloning of aci01347 in E. coli MKH13
The aci01347 gene was cloned with a histidine tag into the vector pBAD/HisA into the MCS (SacI and EcoRI). The histidine tag (his 6 ) was added at the N-terminus. The primers pBAD/HisA_aci01347_ fwd and pBAD/HisA_aci01347_rev (Supporting Information Table   S1) were used to amplify the aci01347 gene from the genome of A. baumannii ATCC 19606. The plasmid was then transferred into E. coli MKH13, which is devoid of all compatible solute transporter.

| [ 14 C]-choline and [ 14 C]-glycine betaine uptake in E. coli MKH13
Escherichia coli MKH13 carrying the plasmid pBAD/HisA_aci01347 and a control strain carrying pBAD/HisA were cultivated at 37°C in LB medium containing ampicillin (100 µg/ml). Expression of the transporter was induced at an OD 600 of 0.6-0.8 with 0.02% arabinose. The cells were harvested 2.5 hr after induction and washed in 0.5 volume KPi-buffer pH 7.5 (25 mM) with 100 mM NaCl. Cells were resuspended in the same buffer containing 30 mM glucose to an OD 600 of 3. Different osmolalities of the KPi-buffer were adjusted with KCl. For compatible solute uptake studies, cell suspensions were diluted 1:1 with the adjusted KPi-buffer, and after 3 min of incubation at 37°C, the assay was started by adding 500 µM [ 14 C]-substrate (1 µCi). Samples (200 µl) were taken at time points indicated. Cells were separated from the medium by filtration using mixed cellulose nitrate filters (pore size 0.45 µm, Sartorius Stedim, Göttingen, Germany) and washed with 20 volumes 0.6 M KPi-buffer. The filters were dried and then dissolved in 4 ml scintillation fluid (Rotizint R eco plus; Carl-Roth GmbH, Karlsruhe, Germany); the radioactivity was determined by a liquid scintillation counter. fwd + aci01347_ctr_rev (Supporting Information Table S1). Integrants were grown overnight in LB containing 10% sucrose, plated on LBagar containing 10% sucrose, and single colonies were analyzed by replica plating onto LB/kanamycin agar with respect to kanamycin sensitivity. The deletion of the aci01347 gene was verified by PCR using the primers aci01347_ctr_fwd + aci01347_ctr_rev (Supporting Information Table S1).

| Complementation of aci01347 mutants
For complementation studies, the vector pVRL1 was used (Lucidi et al., 2018). The aci01347 gene and the promotor region were amplified from genomic DNA using the primer pairs aci01347_compl_fwd + aci01347_compl_rev and aci01347_promo-tor_fwd + aci01347_promotor_rev (Supporting Information Table S1) and inserted in PstI and NotI digested pVRL1. The resulting plasmid pVRL1_aci01347 and the pVLR1 vector control were transformed into

| Deletion of aci01347 abolishes growth on carnitine
To address the role of Aci01347 in A. baumannii ATCC 19606, the gene was deleted from the chromosome via single homologous recombination and segregation of the plasmid using sacB, yielding the markerless deletion mutant Δaci01347 (Stahl, Bergmann, Göttig, Ebersberger, & Averhoff, 2015). The deletion mutant was verified by DNA sequencing. The Δaci0134 mutant was completely defect in growth on carnitine, whereas growth on acetate was not affected ( Figure 2a). To validate the role of aci01347 in carnitine metabolism, the Δaci01347 mutant was complemented with plasmid F I G U R E 1 Genetic organization of the aci01347 locus of Acinetobacter baumannii. The potential regulator of the genes essential for carnitine oxidation is labeled with LysR, TDH is a predicted tartrate dehydrogenase, CntA/B was found to encode an unusual Rieske-type oxygenase (Zhu et al., 2014), MSA-DH is a potential malate semialdehyde dehydrogenase, and Aci01347 is a potential BCCT pVRL1_aci01347 containing the aci01347 gene under the control of the native promotor. The complemented mutant grew with carnitine as sole carbon and energy source, whereas the mutant with the empty vector pVRL1 did not (Figure 2b). This together with the similarities of Aci01347 to BCCT strongly suggests that aci01347 encodes a carnitine transporter. Moreover, the complete loss of the ability to use carnitine as sole carbon and energy source leads to the conclusion that Aci01347 is essential for growth on carnitine.
In previous studies, we analyzed the induction of two distinct choline transporter in A. baylyi and found that they are maximally induced in the presence of choline and high salt (Sand et al., 2014).
High salt also induces the synthesis of glycine betaine transporter.
To address the question whether other potential choline transporter or glycine betaine transporter encoded by A. baumannii are also able to mediate carnitine uptake, the A. baumannii Δaci01347 mutant was grown in mineral medium with carnitine as carbon source in the presence of 1 mM choline and 300 mM NaCl. Still no growth was observed (data not shown) indicating that Aci01347 is the only carnitine transporter in A. baumannii.

| Aci01347 is a secondary active transporter for quaternary ammonium salts
To study the biochemical function of Aci01347, the encoding gene was cloned into the expression vector pBAD/HisA, and the plasmid was transformed into E. coli MKH13. This strain of E. coli is devoid of any transporter for compatible solutes and thus an ideal host to study the function of heterologously produced BCCT (Haardt, Kempf, Faatz, & Bremer, 1995). Transformants were grown in LB medium, washed, and resuspended in KPi-buffer pH 7.5 with 100 mM NaCl and 30 mM glucose. The cell suspensions were used for the transport experiments. The choline uptake was reduced by 4.6%, 2.4%, 9.2%, and 14.5%, respectively. However, only addition of 50-fold excess of carnitine led to a complete inhibition of [ 14 C]-choline transport (Figure 3b).

This complete inhibition of choline uptake by carnitine implies that
Aci01347 mediates the uptake of both, choline and carnitine, at least under these conditions.

| Aci01347 is not activated by osmolarity
The finding that Aci01347-mediated choline uptake raised the question whether Aci01347 is also involved in osmoadaptation. Indeed, the growth of A. baumannii at high salt was stimulated by choline ( Figure 4). This led us to determine a possible osmoactivation of the protein. To this end, E. coli MKH13 pBAD/HisA_aci01347 was washed with KPi-buffer containing 0.2 osmol/kg and resuspended in buffer containing an external osmolality of 0.2 osmol/kg, 0.4 osmol/ kg, 0.6 osmol/kg, 0.8 osmol/kg, or 1 osmol/kg. As can be seen in Figure 5, Aci01347-mediated [ 14 C]-choline uptake activity was not activated by increasing osmolalities but rather decreased over an external osmolality range from 0.2 to 1 osmol/kg. This is consistent with the hypothesis that Aci01347 is not osmotically activated. In line with this hypothesis is the finding that carnitine did not stimulate growth of A. baumannii at high salt ( Figure 4). and virulence factor induction (Chen, Malek, Wargo, Hogan, & Beattie, 2010;Kleber, Schöpp, Sorger, Tauchert, & Aurich, 1967;Lucchesi et al., 1995;Wargo & Hogan, 2009). In Enterobacteriaceae, such as E. coli, Salmonella typhimurium, or Proteus vulgaris, it was found that under anaerobic conditions, when no other electron acceptor is present, carnitine is metabolized to γ-butyrobetaine and carnitine uptake is mediated by a member of the BCCT family which functions as a substrate:product antiporter (Eichler, Bourgis, Buchet, Kleber, & Mandrand-Berthelot, 1994;Jung et al., 2002).

| D ISCUSS I ON
The carnitine transporter in A. baumannii differs from the carnitine transporter in P. aeruginosa and E. coli. Aci01347 is probably a proton:substrate symporter, as other choline transporter (Sand et al., 2014) and the first non carnitine:γ-butyrobetaine transporter described. In silico structural analyses of the glycine betaine transporter BetP of Corynebacterium glutamicum led to the identification of two sodium-binding sites. The first sodium-binding site comprises of T246, T250, and F380, whereas the second sodium-binding site is formed by M150, A147, F464, T467, and S468 (Khafizov et al., 2012). These sodium-binding sites are not present in Aci01347, which corresponds with our suggestion that Aci01347 is probably a proton:substrate symporter.
Aci01347 of different A. baumannii strains contains a conserved glycine-rich segment in transmembrane helix three with the GMGIG motif which is typical for glycine betaine transporter and slightly different in choline-specific transporter (GIGIA/D; Figure 6) but absent in the substrate:product antiporter CaiT of E. coli (Sand et al., 2014, Lamark et al., 1991. The charged residue in this motif is suggested to be crucial for substrate co-ordination for protoncoupled transporter of the BCCT family (Ziegler, Bremer, & Krämer, 2010). A serine is present in position four of the glycine motif in Aci01347 instead of an isoleucine. Whether this serine residue is important for Aci01347-mediated carnitine transport will be subject of future studies.
Three-dimensional structures of the glycine betaine BCCT BetP of C. glutamicum revealed a long hydrophilic C-terminal extension.
Aci01347 exhibits only a short C-terminal hydrophilic stretch of 53 hydrophilic residues which corresponds to the absence of osmoactivation of Aci01347.
We confirmed that A. baumannii uses carnitine as sole carbon and energy source. This together with the abundance of carnitine in the human host (Meadows & Wargo, 2015) suggests that carnitine is important for metabolic adaptation of A. baumannii to the human host.
However, this might only be one physiological benefit of carnitine for A. baumannii. In A. baumannii, carnitine was found to be converted by the Rieske-type oxygenase (CntAB) to trimethylamine (TMA; Zhu et al., 2014). The later is subsequently oxidized in the liver to the proatherogenic species trimethyl-N-oxide (TMAO) which correlates with human cardiovascular health (Koeth et al., 2013). It is tempting to speculate that carnitine uptake and subsequent degradation do not only play a role in metabolic adaptation to the human host but might also facilitate virulence. Whether carnitine metabolism of A. baumannii indeed plays a role in virulence will be subject of future studies.

CO N FLI C T O F I NTE R E S T
The authors declare no conflict of interest.

AUTH O R S CO NTR I B UTI O N
JB, IW, and BA participated in design of the research, analysis of the data, and creating the manuscript. JB performed the experiments.
All authors read and approved the final manuscript.

E TH I C S S TATEM ENT
This research did not involve studies with human or animal subjects, materials or data; therefore, no ethics approval is required.

DATA ACCE SS I B I LIT Y
All data are included in the main manuscript or available as Supporting Information.