Engineering lithoheterotrophy in an obligate chemolithoautotrophic Fe(II) oxidizing bacterium

Neutrophilic Fe(II) oxidizing bacteria like Mariprofundus ferrooxydans are obligate chemolithoautotrophic bacteria that play an important role in the biogeochemical cycling of iron and other elements in multiple environments. These bacteria generally exhibit a singular metabolic mode of growth which prohibits comparative “omics” studies. Furthermore, these bacteria are considered non-amenable to classical genetic methods due to low cell densities, the inability to form colonies on solid medium, and production of copious amounts of insoluble iron oxyhydroxides as their metabolic byproduct. Consequently, the molecular and biochemical understanding of these bacteria remains speculative despite the availability of substantial genomic information. Here we develop the first genetic system in neutrophilic Fe(II) oxidizing bacterium and use it to engineer lithoheterotrophy in M. ferrooxydans, a metabolism that has been speculated but not experimentally validated. This synthetic biology approach could be extended to gain physiological understanding and domesticate other bacteria that grow using a single metabolic mode.

www.nature.com/scientificreports/ single metabolic growth mode prohibits comparative "omics" studies and targeted gene deletions to probe Fe(II) oxidation and carbon flow as mutants defective in these pathways will be unable to grow.
Here we use synthetic biology to study Mariprofundus ferrooxydans PV-1, the founding member of the Zetaproteobacteria 3 which are thought to be the dominant Fe(II) oxidizers in marine environments 3,11 . We develop genetic methods and tools to transform M. ferrooxydans and manipulate its metabolic capacity by expressing foreign genes, yielding an engineered variant capable of using glucose as a carbon source instead of CO 2 .

Results and discussion
We developed a conjugation protocol to successfully transform M. ferrooxydans using the donor strain Escherichia coli WM3064, which is auxotrophic for diaminopimelic acid (DAP) 15 . M. ferrooxydans transformed with pRK2m3 16 continued to grow and produce characteristic twisted iron oxide stalks 3 over successive transfers in the presence of kanamycin (Fig. 1a, b). Wild-type cells incubated with kanamycin was unable to grow (data not shown) and only amorphous iron oxyhydroxides were observed (Fig. 1c), likely produced from abiotic Fe(II) oxidation. 16S rRNA gene sequencing confirmed that the transformed culture was M. ferrooxydans. After ten transfers, E. coli cells were undetectable by microscopy or growth in lysogeny broth (LB) medium augmented with DAP. Maintenance of pRK2m3 in the transformed M. ferrooxydans cells was confirmed by amplifying 330 bp of plasmid specific DNA using total extracted DNA as the template (Fig. 1d) and verified by sequencing. These results demonstrate that M. ferrooxydans was able to replicate pRK2m3 over repeated transfers under kanamycin selection. With a method for transformation and selection established, we were able to express green fluorescent protein (GFP), encoded by gfpmut2 17 and driven by the P neo promoter amplified from upstream the gene encoding kanamycin resistance on pRK2m3. P neo was used because it provided sufficient expression to confer kanamycin resistance in M. ferrooxydans transformed with the pRK2m3 vector. Microscopy confirmed production of GFP in the engineered strain (Fig. 2).
We next sought to leverage our ability to introduce and express foreign genes to augment the metabolism of M. ferrooxydans. The M. ferrooxydans PV-1 genome is predicted to encode genes for glycolysis and the Krebs cycle 12 , but lacks genes encoding glucokinase or an apparent glucose transporter. We hypothesized that by introducing the capability to transport and phosphorylate glucose, M. ferrooxydans could use it as a carbon and energy source. The genes galP and glk from E. coli, encoding a glucose symporter and glucokinase 18 , were cloned into pRK2m3 www.nature.com/scientificreports/ to create pGlu where each gene was individually driven by P neo promoters. M. ferrooxydans cells transformed with pGlu did not yield viable cells when selected heterotrophically using glucose (data not shown). However, growth was observed when M. ferrooxydans cells transformed with pGlu were selected lithoheterotrophically under Fe(II) oxidizing conditions with glucose as the sole carbon source without the addition of carbon dioxide (Fig. 3). M. ferrooxydans transformed with an empty pRK2m3 vector was unable to grow with glucose as the sole carbon source (Fig. 3). The presence of pGlu was confirmed by amplifying and sequencing galP, glk and the kanamycin resistance cassette using total DNA extracted from transformed glucose-grown cells as template (data not shown). Purity of the transformed cells was confirmed by 16S rRNA gene sequencing, microscopy analysis and the absence of bacterial growth in LB medium augmented with DAP. The rate of Fe(II) oxidation by the engineered lithoheterotrophic strain was slower during glucose-dependent growth compared to the Fe(II) oxidation rate during carbon dioxide-dependent growth (Fig. 4). Interestingly, the engineered strain oxidized less total Fe(II) when grown with glucose compared to carbon dioxide (Fig. 4), despite achieving similar final cell densities (Fig. 3). The increase in cell yield per unit Fe(II) oxidized during growth on glucose of the engineered strain can be theoretically attributed to additional energy production from glycolysis and/or biomass precursors provided by glucose.
The inability to transport organic carbon or glycolytic lesions have previously been hypothesized as the reasons for obligate autotrophy in some microorganisms 19 . However, when these deficiencies were addressed using pGlu in M. ferrooxydans, heterotrophic growth was not observed. While the reasons for absence of heterotrophic growth in M. ferrooxydans containing pGlu are unknown, we speculate that the cells may either have insufficient flux through glycolysis or that the genes required for glycolysis and anaplerotic reactions are not expressed under the conditions tested. Another possibility could be the inability to convert NADH/NADPH produced by glycolysis into proton motive force (and then ATP). The obligate requirement of Fe(II) as the energy source even while using glucose in the engineered lithoheterotrophic strain provides an important insight into the metabolic functioning of M. ferrooxydans where glycolysis seems to be partitioned from energy metabolism. Such a metabolism could be one of the reasons driving obligate lithotrophy in M. ferrooxydans. We hypothesize that additional components and alteration of metabolic networks will be required to achieve heterotrophic growth in  Our work provides a proof of concept for using synthetic biology to augment metabolism in microbes with limited or unknown metabolic capabilities to enhance their growth capabilities in laboratory conditions. For example, this approach may be applied to a wide range of bacteria that live by only a single metabolic mode. Additional metabolic enhancement of M. ferrooxydans may yield a fully heterotrophic strain that is able to grow without iron oxidation. A heterotrophic strain of M. ferrooxydans would grow more robustly, without producing iron oxides as a metabolic byproduct, and be amenable to characterization of genes involved in iron oxidation by mutation. Enhancing the metabolic capabilities of metabolic specialists can provide a way to better understand their physiology and provide a blueprint for their domestication.

Materials and methods
Bacterial cultivation and DNA extraction. M. ferrooxydans PV-1 was obtained from National Center for Marine Algae and Microbiota culture collection (https ://ncma.bigel ow.org) and was grown on artificial sea water medium (ASW) 3 , buffered to pH 6.5 with 10 mM MES buffer, using Fe(0) or FeCl 2 as electron donor. When Fe(0) was used, M. ferrooxydans was inoculated into petri plates containing liquid ASW medium and incubated in sealed boxes containing one BD Campy Pack (Catalogue # 4080) to produce a headspace of N 2 :CO 2 :O 2 (80:15:5). When FeCl 2 was used, the culture was grown in 1000 mL serum bottles containing 750 mL of ASW medium with a headspace of N 2 :CO 2 (80:20) and sealed with butyl rubber stoppers. Sealed serum bottles containing the medium were autoclaved and 3 mL of filtered ferrous chloride solution (100 mM) was added to obtain final Fe(II) concentration of 400 µM. 10 mL of filtered air was added to introduce oxygen as the electron acceptor. 3 mL of filtered ferrous chloride solution (100 mM), and 10 mL of filtered air were added to the serum bottles at every 24 h. Growth-curve experiments were performed in 25 mL Balch tubes containing 10 mL of the appropriate medium. Balch tubes were sparged with the appropriate gas to remove oxygen and sealed with butyl rubber stoppers. After autoclaving, sealed Balch tubes were added with 100 µL of filtered ferrous chloride solution (100 mM) to obtain final Fe(II) concentration of 1000 µM. 0.5 mL of filtered air was added to introduce oxygen as electron acceptor. For carbon dioxide dependent growth, ASW medium sparged with N 2 :CO 2 (80:20) was used. For glucose dependent growth, ASW medium lacking bicarbonate and augmented with 500 µM glucose was sparged with argon gas. To check for glucose dependent growth in the absence of Fe(II), only filtered air was added to the sealed Balch tubes after autoclaving and ferrous chloride was omitted. DNA was extracted from M. ferrooxydans cultures using a Qiagen DNeasy PowerSoil kit. Escherichia coli WM3064 was grown in LB medium containing 360 µM DAP. 50 µM kanamycin was added to the medium when required.  Cell and Fe(II) quantification. 200 µL of the sample was collected periodically from the Balch tubes using sterile syringes and needles. 100 µL of the sample was added to 900 µL of 0.5 N HCl to be used for Fe(II) quantification using ferrozine assay 21 performed in microtiter plates. Cells were fixed in 0.8% paraformaldehyde for 2 h, stained with 12.5 mM Syto9 and counted using a Petroff-Hausser counting chamber on an epifluorescent microscope.
Plasmid construction. The pRK2m3 plasmid used in this study confers resistance to kanamycin, is approximately 5 KB in size and contains an origin of transfer (oriT) for conjugative transfer 16 . pRK2m3 is derived from pRK2, which has been shown to be present in low copy in E. coli and Pseudomonas aeruginosa 22 .
A 330 bp pRK2m3-specific DNA fragment was amplified using the following primers: CCA TGT CGG CAG AAT GCT TA and TGT AAA ACG ACG GCC AGT . P neo was amplified from pRK2m3 using pneoF (GAT AGA ATT CTT GAG ACG TTG ATC GGC ACG ) and pneoR (TAG ACT CGA GAA CAC CCC TTG TAT TAC TGT TTA TGT AAGC) primers. To construct the plasmid for GFP expression, gfpmut2 17 was amplified from pUA66 17 using gfpF (ACG ACT CGA GAT GAG TAA AGG AGA AGA ACT TTT CAC TGGA) and gfpR (TAG AGA GCT CTT ATT TGT ACA ATT CAT CCA TAC CAT GGGTA) primers and cloned into pRK2m3 with the P neo promoter driving its expression. To construct pGlu, galP and glk were amplified from E. coli K-12 using galPF (ATT TAC TAG TAT GCC TGA CGC TAA ACAGG) /galPR (ATT CGA GCT CTT AAT CGT GAG CGC CTA TTT CG) and glkF (ACG ACT CGA GAT GAC AAA GTA TGC ATT AGT CGGT) /glkR (TAG AGA ATT CTT ACA GAA TGT GAC CTA AGG TCTG) primers respectively. Amplified galP and glk were cloned under the control of separate P neo promoters in pRK2m3. All the plasmids were transformed into chemically competent E. coli WM3064 15 cells, followed by selection on LB plates containing 50 µM kanamycin and 360 µM DAP.