nihms400410.pdf

Photomixotrophic growth of

Rhodobacter capsulatus

SB1003 on

ferrous iron

Sebastian H. Kopf

and

Dianne K. Newman

1,2,3

Dianne K. Newman: dkn@caltech.edu

Division of Geological and Planetary Sciences, Pasadena, CA 91125

Division of Biological Sciences, California Institute of Technology, Pasadena, CA 91125

Howard Hughes Medical Institute, Pasadena, CA 91125

Abstract

This study investigates the role iron oxidation plays in the purple nonsulfur bacterium

Rhodobacter capsulatus

SB1003. This organism is unable to grow photoautotrophically on

unchelated ferrous iron [Fe(II)] despite its ability to oxidize chelated Fe(II). This apparent paradox

was partly resolved by the discovery that SB1003 can grow photoheterotrophical-ly on the

photochemical breakdown products of certain ferric iron - ligand complexes, yet whether it could

concomitantly benefit from the oxidation of Fe(II) to fix CO

was unknown. Here, we examine

carbon fixation by stable isotope labeling of the inorganic carbon pool in cultures growing

phototrophically on acetate with and without Fe(II). We show that

R. capsulatus

SB1003, an

organism formally thought incapable of phototrophic growth on Fe(II), can actually harness the

reducing power of this substrate and grow photomixtotrophically, deriving carbon both from

organic sources and fixation of inorganic carbon. This suggests the possibility of a wider

occurrence of photoferrotrophy than previously assumed.

1. Introduction

Microbial processes throughout Earth's history have had a profound impact on the

biogeochemical cycling of iron (

Kappler and Straub, 2005

;

Ehrlich and Newman, 2008

While much attention has been paid to iron's ability to serve as an electron donor or electron

acceptor in catabolic processes, beyond a crude accounting for electrons in the metabolisms

of a few model organisms, we have little appreciation for how cells make use of iron's redox

chemistry. For example, we would expect multiple elements within a cell to be affected by

an imbalance in iron homeostasis, which in turn would be expected to change how a cell

might regulate its export or uptake of substrates containing these elements. Similarly, we

would expect intracellular redox homeostasis to be influenced by iron in myriad ways. How

these and other more subtle effects manifest themselves is poorly understood, yet they may

be important drivers of the overall iron biogeochemical cycle.

Correspondence to: Dianne K. Newman,

dkn@caltech.edu

Published as:

Geobiology

. 2012 May ; 10(3): 216–222.

HHMI Author Manuscript

In recognition of this knowledge gap, we chose to explore how ferrous iron [Fe(II)] is used

by the anoxygenic phototroph,

Rhodobacter capsulatus

strain SB1003. We chose this

organism as a model system because it exhibited a curious phenotype during phototrophic

growth. Unlike other

Rhodobacter

species (

Ehrenreich and Widdel, 1994

R. capsulatus

SB1003 does not oxidize iron (and grow photolithotrophically) in medium containing Fe(II)

chloride as the sole source of reducing power (

Croal et al., 2007

). However, in the presence

of chelating agents such as citrate and NTA, Fe(II) oxidation is enabled and SB1003 can

grow photoheterotrophically on supplementary carbon sources (

Croal et al., 2007

;

Poulain

and Newman, 2009

), or, in the case of photoactive ferric [Fe(III)]-ligand complexes such as

Fe(III)-citrate, on the photochemical breakdown products of the ligand (

Caiazza et al.,

2007

). This shows that Fe(II) oxidation can benefit the organism indirectly but does not

resolve whether Fe(II) oxidation can benefit

R. capsulatus

directly. (

Poulain and Newman,

2009

) first explored the ambiguous role of Fe(II) oxidation in

R. capsulatus

SB1003 and

proposed Fe(II) oxidation as a potential detoxification mechanism. Preliminary data on gene

expression furthermore revealed that several Calvin cycle genes are upregulated in the

presence of Fe(II) (Poulain and Newman, unpublished data), suggesting a potential link to

Fe(II) oxidation. Herein, we expand on these studies and show that

R. capsulatus

SB1003

can grow photomixotrophically using Fe(II) as an electron donor for carbon fixation.

2. Results and Discussion

2.1 Fe(II) oxidation promotes growth

To elucidate whether Fe(II) oxidation itself confers any growth benefit to

R. capsulatus

, we

assessed phototrophic growth on Fe(II) complexed by nitrilotriacetate (NTA) in anoxic

freshwater medium, completely devoid of additional electron donors (trace organics or

hydrogen, see Materials and Methods), with bicarbonate as the sole carbon source (Figure

1). The lack of growth in the control experiment (NTA only) shows that

R. capsulatus

cannot use NTA as a carbon source. When provided with Fe(II)-NTA in the light, the

organism grew rapidly for 2 days and appeared to grow slowly for the remainder of the

experiment. Optical density likely reflects growth as iron oxides did not precipitate (all

Fe(III) complexed by NTA), although small variations in optical density could also be due to

morphological changes as the cells aged. Fe(II) was completely oxidized after 2 days (data

not shown) and remained oxidized for the remainder of the experiment. No growth or Fe(II)

oxidation occurred in the dark (Figure 1). When provided with Fe(III)-NTA instead, the

organism had no direct source of reducing power and showed only very little growth, similar

to later stage (2+ days) on Fe(II)-NTA (Figure 1). Small quantities of reduced Fe(II)

(360±80 μM) accumulated in the medium within 2 days. Since the Fe(III)-NTA complex can

be photochemically active at the experimental pH and light regime (

Andrianirinaharivelo et

al., 1993

), the observed accumulation of Fe(II) is likely due to photoreduction of Fe(III) in

the Fe(III)-NTA complex. This process is expected to photodegrade the ligand, yielding

NTA breakdown products that could serve as a carbon source for photoheterotrophic growth

ordissimilatory Fe(III) reduction (

Dobbin et al., 1996

). The apparent slow growth in the

presence of Fe(III)-NTA, both when supplied initially or provided by oxidation of Fe(II),

suggests that the organism can benefit from photoreduction of Fe(III), photolytic breakdown

of NTA, or both. Because Fe(II)-NTA does not photolyze, and photolysis of unbound NTA

Kopf and Newman

Page 2

Geobiology

. Author manuscript; available in PMC 2015 September 29.

HHMI Author Manuscript

is exceedingly slow (

Larson and Stabler, 1978

), the rapid initial growth in the presence of

Fe(II)-NTA cannot be explained by growth on photochemical breakdown products and

suggests that the oxidation of Fe(II) provides a growth benefit to

R. capsulatus

2.2 Fe(II) oxidation allows for carbon fixation

To test whether Fe(II) serves as an electron donor for carbon fixation, we conducted isotope

labeling experiments with

C labeled bicarbonate. If CO

is fixed during growth on Fe(II)-

NTA, inorganic

C should be strongly incorporated into cell carbon. However, purple non-

sulfur bacteria like

R. capsulatus

also use CO

as a sink for excess reducing equivalents to

achieve redox homeostasis during photoheterotrophic growth (

Tabita, 2004

), potentially

obscuring this signal. This is further complicated by any potential contribution to growth

from photolytic breakdown products of NTA, which would introduce unlabeled carbon into

the cell. To allow a quantitative interpretation of the labeling experiments, we thus explored

three different growth conditions: phototrophic growth on [A] acetate alone, [B] Fe(II)-NTA

alone, and [C] acetate and Fe(II)-NTA together. Table 1 documents optical density as well

as acetate and Fe(II) concentrations at the onset and conclusion of each experiment, in

addition to the

C content of the harvested cells (see Materials and Methods for

experimental details). Because all organic carbon sources were unlabeled (acetate, NTA)

and the entire inorganic pool was labeled (bicarbonate), these isotopic data represent the net

assimilation of organic vs. inorganic carbon into biomass for each growth condition. The

low variability between biological replicates provides confidence in the reproducibility and

comparability of the experimental conditions.

The isotopic data from growth condition [A] (acetate only) illustrates the role of carbon

fixation for redox homeostasis during photoheterotrophic growth. During photoheterotrophic

growth, purple non-sulfur bacteria like

R. capsulatus

generate energy from cyclic

phosphorylation while building cell carbon directly from organic carbon sources. The use of

organic substrates for biosynthesis, however, can lead to a buildup of excess reducing

power, requiring these phototrophs to find an electron sink to maintain redox homeostasis.

In organisms limited fornitrogen, nitrogenase can serve this function by sinking excess

electrons into N

and H

, producing ammonium and H

(

Hillmer and Gest, 1977

;

Mckinlay

and Harwood, 2011

). Additionally, certain alternative electron acceptors such as

dimethylsulfoxide can provide the necessary electron sink (

Richardson et al., 1988

). More

commonly though, redox homeostasis under photoheterotrophic growth of

R. capsulatus

achieved by using CO

as a sink for excess reducing equivalents through the Calvin-Benson-

Bassham pathway (the Calvin cycle) (

Tichi and Tabita, 2000

2001

;

Bauer et al., 2003

;

Tabita, 2004

). During the experimental conditions employed in this study,

photoheterotrophic growth of

R. capsulatus

in the presence of excess ammonium and the

absence of alternative electron acceptors, CO

fixation via the Calvin cycle provides the

only available sink for excess reducing power. This introduces cell carbon derived from the

inorganic carbon pool into the cell. Our data indicate that close to 17% (Table 1) of cellular

carbon is derived from the inorganic carbon poolduring photoheterotrophic growth of

capsulatus

SB1003 on acetate. This result is in agreement with a detailed metabolic flux

analysis of another purple phototroph,

Rhodopseudomonas palustris

, growing

photoheterotrophically on acetate (

Mckinlay and Harwood, 2010

R. palustris

metabolizes

Kopf and Newman

Page 3

Geobiology

. Author manuscript; available in PMC 2015 September 29.

HHMI Author Manuscript

22% of the provided acetate to CO

via central metabolic pathways and reincorporates 68%

of the released CO

into cell carbon via the Calvin cycle, ultimately deriving approximately

16% of its cellular carbon from CO

(

Mckinlay and Harwood, 2010

). Several pathways of

acetate assimilation that can explain the observed exchange of carbon with the inorganic

pool have been discovered in the purple phototrophs (

Blasco et al., 1989

;

Willison, 1998

;

Filatova et al., 2005

;

Meister et al., 2005

). However, why these organisms build up excess

reducing power (that is disposed of via the Calvin cycle or other redox sinks) during growth

on substrates that are more oxidized than cell carbon is still poorly understood (

Mckinlay

and Harwood, 2010

). Under the experimental conditions employed in this study, the

energetic cost of carbon fixation via the Calvin cycle is unlikely to significantly affect the

energy available to

R. capsulatus

. ATP generation via cyclic photophosphorylation provides

energy independently of the growth-limiting sources of reducing power (organic carbon and

Fe(II)).

The large incorporation of labeled inorganic carbon in cultures grown phototrophically on

Fe(II)-NTA alone (close to 63%, growth condition [B], see Table 1) confirms that Fe(II)

serves as an electron donor for carbon fixation. However, the dilution of the inorganic signal

(99%

C) indicates that

R. capsulatus

must be capable of assimilating some organic carbon

from the chelator NTA, the only unlabeled pool of carbon available in the medium. Because

the organism is unable to metabolize NTA directly (Figure 1), the isotopic data suggest that

it can benefit from photolysis of the ligand. Previous studies (

Caiazza et al., 2007

) have

shown that under similar conditions, growth of

R. capsulatus

SB1003 on Fe(II)-citrate

occurs as a result of Fe(II) oxidation and Fe(III)-citrate photochemistry, which produces

acetoacetic acid as a consequence of ligand breakdown, a carbon source accessible to the

organism. A similar model is conceivable for Fe(II)-NTA with the well studied

photochemically active Fe(III)-NTA complex breaking down to iminodiacetic acid (IDA),

formaldehyde (HCHO), CO

and hydroxyl radicals (

Trott et al., 1972

;

Stolzberg and Hume,

1975

;

Andrianirinaharivelo et al., 1993

;

Bunescu et al., 2008

). IDA can further disintegrate

to formaldehyde and glycine, although this second photolytic step proceeds at slower rates

(

Stolzberg and Hume, 1975

). Based on the presence of genes annotated as hydroxymethyl-

transferase (

glyA

, RCC00438), serine-glyoxylate aminotransferase (RCC03109) and

hydroxypyruvate reductase (

ttuD

, RCC02615) in the sequenced genome of

R. capsulatus

SB1003, formaldehyde assimilation by the serine pathway shouldbe possible in SB1003,

making for maldehydea potential carbon source(

Chistoserdova et al., 2003

). This pathway is

a common functional module in methylotrophs and well understood at the biochemical level.

An additional pathway of formaldehyde metabolism using the glutathione-dependent

formaldehyde dehydrogenase (

adhC

, RCC00869) present in the genome of SB1003 is also

possible as pointed out previously (

Caiazza et al., 2007

). This pathway leads to the net

generation of reducing equivalents by oxidizing formaldehyde and could be used for fixing

inorganic carbon but would not lead to direct assimilation of the organic carbon (which is

oxidized to CO

R. capsulatus

is unable to grow photoheterotrophically on IDA (no

growth observed on 5mM IDA over the course of 3 days, data not shown) but can grow on

glycine as the sole carbon source (growth on 5mM glycine up to an optical density of 0.3 at

675nm, data not shown), rendering glycine an additional potential source for carbon. Lastly,

the radicals formed during Fe(III)-NTA photolysis can potentially interact with NTA or IDA

Kopf and Newman

Page 4

Geobiology

. Author manuscript; available in PMC 2015 September 29.

HHMI Author Manuscript

to provide additional accessible, yet unidentified carbon sources. Our isotopic data provides

evidence that

R. capsulatus

can assimilate some of these NTA breakdown products,

however, which photolytic product of NTA degradation is metabolized by the organism and

how, remains to be shown.

The intermediate incorporation of labeled inorganic carbon in cultures grown

phototrophically on acetate and Fe(II)-NTA (28%, growth condition [C], see Table 1) is

consistent with a combination of the effects observed during growth on acetate alone and

growth on Fe(II)-NTA alone (conditions [A] and [B]).

These results indicate that

R. capsulatus

SB1003 grows photomixotrophically by fixing CO

with Fe(II) as the electron donor (photoautotrophic metabolism) while simultaneously

assimilating organic carbon sources (photoheterotrophic metabolism). Why the organism

can benefit from the oxidation of Fe(II)-NTA but fails to oxidize unchelated Fe

unclearand merits further research. It could reflect a requirement for ligand-bound Fe(II) to

be recognized for efficient uptake into the cell, and/or result from a toxic effect of the free

metal ion, as suggested by (

Poulain and Newman, 2009

2.3 Mass balance model

The isotopic data shows that

R. capsulatus

can incorporate a mixture of carbon sources

during phototrophic growth and provides a basis for quantitative evaluation of their

respective contributions. In the presence of complexed Fe(II), the organism seems capable

of exploiting simple, naturally wide-spread organic acids (such as acetate), photochemically

mobilized refractory carbon (such as ligand breakdown products) as well as the reducing

power of the Fe(II) itself (summarized schematically in Figure 2). Isotopic mass balance

yields the relative contributions of these carbon sources:

Where

[Inoc], [Ac], [Fe(II)]

and

[NTA]

are the concentrations of the different carbon pools

(initial inoculum, acetate assimilation, carbon fixation through Fe(II) oxidation, and

acquisition of carbon from Fe(III)-NTA breakdown) contributing to total cell carbon

[TCC]

denotes the net efficiency of acetate assimilation,

C/Fe

the net ratio of molecules of

fixed into cell carbon per atoms of Fe(II) oxidized.

Inoc-C

, %

Ac-C

, %

Fe-C

and

NTA-C

indicate the isotopic composition of cell carbon derived from these different

sources, respectively.

TCC

is the isotopic composition of total cell carbon, as measured

in the isotope labeling experiments in this study (Table 1).

The net efficiency of carbon assimilation from acetate during photoheterotrophic growth of

R. capsulatus

was 1.75±0.14 mol cell C / mol acetate (

=88% carbon conservation

efficiency, average from 5 biological replicates ± SD), in agreement with similar

measurements for the anoxygenic phototroph

R. palustris

(

Mckinlay and Harwood, 2010

The net ratio of CO

fixation to Fe(II) oxidation was estimated to be

C/Fe

=0.23 since 4.3

electrons from Fe(II) are required to reduce inorganic carbon to the carbon redox state of of

Kopf and Newman

Page 5

Geobiology

. Author manuscript; available in PMC 2015 September 29.

HHMI Author Manuscript

R. Capsulatus

biomass (-0.3, see Supporting Information for details). The contribution from

the inoculum to final biomass was estimated to be 1% of the biomass generated from growth

on acetate (the 1% inoculum). Assuming that acetate metabolism proceeds by very similar

pathways irrespective of the presence or absence of Fe(II)-NTA metabolism and vice-versa,

the mass balance equations for the different experimental conditions [A, B and C] provide

the estimates for the unknown parameters (

NTA-C

and

[NTA]

) reported in Table 2.

NTA breakdown products are metabolized with an exchange of

∼

18.6% of assimilated

carbon with the inorganic pool (either by dear boxylation/recarboxylation reactions or

oxidation and refixation via the Calvin cycle), which is similar to our measured value for

acetate assimilation (16.9%). The total amount of cell carbon derived from the assimilation

of NTA breakdown products (0.70 mM C), corresponds to 0.12mM NTA (a 6C compound) -

or 1.2% of the total NTA pool - if all carbon in the photolytic breakdown products can be

metabolized by the organism. The precise rate of photolytic breakdown is difficult to

estimate because Fe(III)-NTA photolysis is strongly pH and wavelength dependent, but

would be expected to be slow at circumneutral (or higher) pH with little irradiation in the

UV (shorter than

∼

365nm) (

Andrianirinaharivelo et al., 1993

The estimate for

[NTA]

also enables calculation of an electron mass balance based on the

total electron content of the generated

R. capsulatus

biomass and the total electron content

of the consumed substrates (acetate, Fe(II) and NTA breakdown products). For this

calculation, we assume that the oxidation state of carbon in the assimilated NTA breakdown

product corresponds to the oxidation state of carbon in formaldehyde; making this

assumption, we can estimate the total electron recovery for each experimental condition

(Table 2).

2.4 Obligate mixotrophy

Interestingly, the model results and Figure 1 suggest that

R. capsulatus

cannot fully benefit

from the assimilation of NTA breakdown products in the absence of Fe(II). In the presence

of Fe(II), NTA breakdown products contribute as much as 41% to cell carbon. If provided

with Fe(III)-NTA, however, this contribution is not visible in the growth curve. Because all

organic carbon sources available to

R. capsulatus

in our experiments are slightly more

oxidized than bulk biomass (see Supporting Information for details), the organism requires

some reducing power for net biosynthetic reduction of the substrate. In the case of

photoheterotrophic growth on acetate alone, the oxidation of some acetate can provide the

necessary reducing power to assimilate the remaining acetate, ultimately contributing to the

observed suboptimal efficiency (88%) of acetate assimilation. In the case of NTA

breakdown products, however, some of the reducing power available from the oxidation of

Fe(II) might be required to fully benefit from these carbon sources (hypothesized pathway,

Figure 2). If this were the case, the observed phototrophic growth on Fe(II)-NTA would be

truly obligate photomixotrophy. This possibility cannot be fully resolved here but provides a

testable hypothesis for further research. Detailed metabolic flux experiments could help

elucidate the interdependence of these pathways and explore the role Fe(II) oxidation might

play in the assimilation of refractory organic carbon sources. Analogously, the ambiguous

function of Fe(II) oxidation in anaerobic chemotrophic Fe(II) oxidizers (

Kappler et al.,

Kopf and Newman

Page 6

Geobiology

. Author manuscript; available in PMC 2015 September 29.

HHMI Author Manuscript

2005

) might similarly be explained by considering Fe(II) oxidation as an auxiliary

mechanism for redox balancing during the assimilation of organic carbon.

3. Conclusion

The existence of mixotrophic growth itself is not surprising and has previously been

suggested to occur both in phototrophic (

Widdel et al., 1993

) and chemotrophic (

Hallbeck

and Pedersen, 1991

)Fe(II) metabolisms. However, its significance is rarely appreciated

despite its likely importance in nature. Our results show that an organism not previously

considered capable of growing by Fe(II) oxidation, in fact originally thought

incapable

oxidizing Fe(II) altogether, can use Fe(II) for growth under certain conditions. The ability to

grow photomixotrophically on Fe(II) might be more widespread than previously assumed,

even for cultured organisms that have simply not been exposed to conditions that allow this

mode of growth to be observed in the laboratory. It will be interesting to learn whether

mixotrophic growth accounts for a significant proportion of cellular Fe(II) oxidation activity

by different types of Fe(II) oxidizing organisms. Future research into the different enzymatic

pathways of Fe(II) oxidation and a more detailed understanding of their regulation could

permit a more accurate assessment of how widespread and environmentally significant

microbial Fe(II) oxidation is today and has been throughout Earth's history.

4. Materials and Methods

4.1 Experimental conditions

Rhodobacter capsulatus

SB1003 was grown phototrophically in anoxic, minimal-salts

freshwater medium, prepared as previously described (

Ehrenreich and Widdel, 1994

). The

medium was buffered at pH 7.0 with 22mM sodium bicarbonate. All experiments were

prepared in an oxygen- and hydrogen-free, anaerobic chamber under an atmosphere of pure

. All reagents and glassware were stored in the chamber at least three days prior to use to

remove traces of oxygen. In addition to standard heat sterilization procedures used for all

equipment and medium preparation, glassware was precombusted in a muffle furnace at

550°C to remove all remaining traces of organic materials potentially adhered to the glass.

Cells were grown an aerobically at 30°C under constant illumination from two 60W

incandescent light sources at 30cm distance, providing a total irradiance of ca. 40 W/m

(45.5% visible light, 54.4% IR,

∼

0.1% UV). Growth was followed by optical density at

675nm (OD675). This wavelength was used to decrease distortion by Fe(III)-NTA, which

absorbs strongly at 600nm. OD675 underestimates optical density as compared to a

measurement at 600nm.

4.2 Phototrophic growth in the presence of iron

Phototrophic growth was assessed in bicarbonate buffered freshwater medium amended with

4mM ferrous Fe(II) complexed by 10mM nitrilotriacetate (NTA), or 5mM ferric Fe(III)

complexed by 10mM NTA. Medium amended with either 5mM NTA only or with Fe(II)-

NTA incubated in the dark was tested as controls. No additional carbon or electron sources

were provided. NTA was chosen as the complexing agent to avoid fast photolytic

breakdown of the Fe(III)-ligand complex. Previous work with citrate revealed a high

degradation of the more photoreactive Fe(III)-citrate complex (

∼

0.6 mM / day,

Caiazza et

Kopf and Newman

Page 7

Geobiology

. Author manuscript; available in PMC 2015 September 29.

HHMI Author Manuscript

al., 2007

)), which allowedrapid accumulation of acetoacetate, a substrate that can be readily

metabolized by

R. capsulatus

and obscures the effect of Fe(II) oxidation. The chosen

concentrations of NTA were previously found not to interfere with growth on other

substrates (data not shown) while ensuring virtually all Fe(II) and Fe(III) remain complexed,

preventing precipitation of iron hydroxides (

Poulain and Newman, 2009

). Experiments were

conducted in biological triplicates. Fe(II) concentrations were measured at the start and end

of the experiment. Fe(II) was quantified throughout this study using the FerroZine assay

(

Stookey, 1970

4.3 Isotope labeling

For isotope labeling, freshwater medium was buffered with 22mM labeled sodium

bicarbonate (NaH

, CAS# 87081-58-01) purchased from Cambridge Isotope

Laboratories, Inc. (catalogue # CLM-441, purification grade: >99%

C). Freshwater

medium was amended with 3mM acetate only [A],

∼

4mM Fe(II)-10mM NTA only [B], or

both [C]. Cultures were harvested upon reaching early stationary phase (20 hours after

inoculation for [A], 29 hours for [B, C]), washed thrice in deionized water and lyophilized

overnight. Acetate and Fe(II) concentrations were determined before inoculation and at the

time of harvest. Acetate was measured using a Dionex ICS-3000 ion chromatography

system with a 4×240mm AS-11 IonPac column and NaOH elution gradient (0.5 to 5.0 mM

NaOH in 3.5 min followed by 5.0 to 37mM NaOH in 12 min at a flow rate of 2 ml/min).

Isotopic composition of bulk cell carbon was determined by EA-IRMS at the UC Davis

Stable Isotope Facility (Davis, CA). Carbon isotopic compositions from labeled experiments

are reported in terms of atom percent %

C = 100 [

C / (

C +

C)] throughout this study.

Isotopic composition of acetate, NTA as well as the in oculum culture used in this study

were measured and confirmed to be approximately natural abundance (

≈

1.1 %

C >

40‰, data not shown). The labeled bicarbonate was assumed to comply with the

manufacturer's specifications (

≈

99 %

C). Experiments were conducted in biological

triplicates [A] or quadruplicates [B, C]. Abiotic controls indicated no significant oxidation

of Fe(II) over the course of the experiment.

4.4Carbon assimilation efficiency

The carbon assimilation efficiency of

R. capsulatus

during phototrophic growth on acetate

was determined from the consumption of acetate and concomitant increase in cell dry weight

of cultures grown to late exponential phase on acetate as the only carbon source. Cultures

were pelleted by centrifugation, filtered onto pre-weighed Spin-X centrifuge tube filters and

dried to constant weight in a 60°C drying oven. Total cell carbon content was derived from

dry weights based on the elemental composition of

R. capsulatus

(CH

1.83

0.183

0.5

Dorffler et al., 1998

). Acetate consumption was measured by ion chromatography.

Supplementary Material

Refer to Web version on PubMed Central for supplementary material.

Kopf and Newman

Page 8

Geobiology

. Author manuscript; available in PMC 2015 September 29.

HHMI Author Manuscript

Acknowledgments

We thank Alexandre Poulain for teaching S.K. how to work with

R. capsulatus

and Alexa Price-Whelan and other

members of the Newman Lab for stimulating discussions. Three anonymous reviewers are gratefully acknowledged

for constructive criticism. This work was supported by grants to D.K.N. from the National Science Foundation

(grant MCB-0616323) and the Howard Hughes Medical Institute. D.K.N. is an Investigator of the Howard Hughes

Medical Institute.

References

Andrianirinaharivelo SL, Pilichowski JF, Bolte M. Nitrilotriacetic acid transformation photo-induced

by complexation with iron(III) in aqueous solution. Transition Metal Chemistry. 1993; 18:37–41.

Bauer C, Elsen S, Swem L, Swem D, Masuda S. Redox and light regulation of gene expression in

photosynthetic prokaryotes. Philos T Roy Soc B. 2003; 358:147–153.

Blasco R, Cardenas J, Castillo F. Acetate metabolism in purple non-sulfur bacteria. FEMS

Microbiology Letters. 1989; 58:129–132.

Bunescu A, Besse-Hoggan P, Sancelme M, Mailhot G, Delort AM. Fate of the nitrilotriacetic Acid-

Fe(III) complex during photodegradation and biodegradation by

Rhodococcus rhodochrous

Applied and Environmental Microbiology. 2008; 74:6320–6326. [PubMed: 18757580]

Caiazza NC, Lies DP, Newman DK. Phototrophic Fe(II) Oxidation Promotes Organic Carbon

Acquisition by

Rhodobacter capsulatus

SB1003. Applied and Environmental Microbiology. 2007;

73:6150–6158. [PubMed: 17693559]

Chistoserdova L, Chen S, Lapidus A, Lidstrom M. Methylotrophy in

Methylobacterium extorquens

AM1 from a genomic point of view. Journal of Bacteriology. 2003; 185:2980–2987. [PubMed:

12730156]

Croal LR, Jiao Y, Newman DK. The fox operon from

Rhodobacter

strain SW2 promotes phototrophic

fe(II) oxidation in

Rhodobacter capsulatus

SB1003. Journal of Bacteriology. 2007; 189:1774–1782.

[PubMed: 17189371]

Dobbin P, Warren L, Cook N, McEwan A, Powell A, Richardson D. Dissimilatory iron(III) reduction

Rhodobacter capsulatus

. Microbiol-Uk. 1996; 142:765–774.

Dorffler M, Steindorf A, Oelze J. Determination of C/N ratios required for de-repression of

nitrogenase in

Rhodobacter capsulatus

. Z Naturforsch C. 1998; 53:961–967.

Ehrenreich A, Widdel F. Anaerobic Oxidation of Ferrous Iron by Purple Bacteria, a New-Type of

Phototrophic Metabolism. Applied and Environmental Microbiology. 1994; 60:4517–4526.

[PubMed: 7811087]

Ehrlich HL, Newman DK. Geomicrobiology. 2008:279.606.

Filatova L, Berg I, Krasil'nikova E, Tsygankov A, Laurinavichene T, Ivanovsky R. A study of the

mechanism of acetate assimilation in purple nonsulfur bacteria lacking the glyoxylate shunt:

Acetate assimilation in

Rhodobacter sphaeroides

. Microbiology+. 2005; 74:265–269.

Hallbeck L, Pedersen K. Autotrophic and mixotrophic growth of

Gallionella ferruginea

. Journal of

General MIcrobiology. 1991; 137:2657–2661.

Hillmer P, Gest H. H2 metabolism in the photosynthetic bacterium

Rhodopseudomonas capsulata

: H2

production by growing cultures. Journal of Bacteriology. 1977; 129:724–731. [PubMed: 838685]

Kappler A, Straub K. Geomicrobiological cycling of iron. Rev Mineral Geochem. 2005; 59:85–108.

Kappler A, Schink B, Newman DK. Fe(III) mineral formation and cell encrustation by the nitrate-

dependent Fe(II)-oxidizer strain BoFeN1. Geobiology. 2005; 3:235–245.

Larson RA, Stabler PP. Sensitized photooxidation of nitrilotriacetic and iminodiacetic acids. Journal of

Environmental Science and Health Part A: Environmental Science and Engineering. 1978;

13:545–552.

Mckinlay JB, Harwood CS. Inaugural Article: Carbon dioxide fixation as a central redox cofactor

recycling mechanism in bacteria. Proceedings of the National Academy of Sciences. 2010;

107:11669–11675.

Kopf and Newman

Page 9

Geobiology

. Author manuscript; available in PMC 2015 September 29.

HHMI Author Manuscript

Mckinlay JB, Harwood CS. Calvin Cycle Flux, Pathway Constraints, and Substrate Oxidation State

Together Determine the H2 Biofuel Yield in Photoheterotrophic Bacteria. mBio. 2011;

2:e00323-00310–e00323-00310. [PubMed: 21427286]

Meister M, Saum S, Alber B, Fuchs G. L-malyl-coenzyme A/beta-methylmalyl-coenzyme A lyase is

involved in acetate assimilation of the isocitrate lyase-negative bacterium

Rhodobacter capsulatus

Journal of Bacteriology. 2005; 187:1415–1425. [PubMed: 15687206]

Poulain AJ, Newman DK. Rhodobacter capsulatus Catalyzes Light-Dependent Fe(II) Oxidation under

Anaerobic Conditions as a Potential Detoxification Mechanism. Applied and Environmental

Microbiology. 2009; 75:6639–6646. [PubMed: 19717624]

Richardson DJ, King GF, Kelly DJ, McEwan AG, Ferguson SJ, Jackson JB. The role of auxiliary

oxidants in maintaining redox balance during phototrophic growth of

Rhodobacter capsulatus

propionate or butyrate. Arch Microbiol. 1988; 150:131–137.

Stolzberg RJ, Hume DN. Rapid formation of iminodiacetate from photochemical degradation of iron

(III) nitrilotriacetate solutions. Environ Sci Technol. 1975; 9:654–656.

Stookey LL. Ferrozine - a new spectrophotometric reagent for iron. Analytical Chemistry. 1970;

42:779–781.

Tabita F. Research on carbon dioxide fixation in photosynthetic microorganisms (1971-present).

Photosynth Res. 2004; 80:315–332. [PubMed: 16328829]

Tichi M, Tabita F. Maintenance and control of redox poise in Rhodobacter capsulatus strains deficient

in the Calvin-Benson-Bassham pathway. Arch Microbiol. 2000; 174:322–333. [PubMed:

11131022]

Tichi M, Tabita F. Interactive control of Rhodobacter capsulatus redox-balancing systems during

phototrophic metabolism. Journal of Bacteriology. 2001; 183:6344–6354. [PubMed: 11591679]

Trott T, Henwood RW, Langford CH. Sunlight photochemistry of ferric nitrilotriacetate complexes.

Environ Sci Technol. 1972; 6:367–368.

Widdel F, Schnell S, Heising S, Ehrenreich A, Assmus B, Schink B. Ferrous iron oxidation by

anoxygenic phototrophic bacteria. Nature. 1993; 362:834–836.

Willison JC. Pyruvate and Acetate Metabolism in the Photosynthetic Bacterium

Rhodobacter

capsulatus

. Journal of General MIcrobiology. 1998; 134:2429–2439.

Kopf and Newman

Page 10

Geobiology

. Author manuscript; available in PMC 2015 September 29.

HHMI Author Manuscript

Figure 1. Fe(II) oxidation promotes growth

Phototrophic growth was assessed in freshwater medium amended with NTA only, ferrous

Fe(II)-NTA (light and dark), or ferric Fe(III)-NTA. Symbols represent the averages of

biological triplicates. Error bars indicate standard deviation and may be smaller than symbol

size. OD (675nm) is optical density at 675nm.

Kopf and Newman

Page 11

Geobiology

. Author manuscript; available in PMC 2015 September 29.

HHMI Author Manuscript

Figure 2. Mixotrophic growth of

R. capsulatus

Schematic overview of the various pathways that can contribute to photomixotrohpic growth

in the presence of Fe(II)-NTA. Question marks (?) indicate hypothesized pathways

(discussed in text) that require further investigation. Abbreviations: Calvin-Benson-Bassham

pathway (CBB), photosynthetic reaction center (RC), NTA breakdown products (NTA

BDP), photochemically/photosynthetically active radiation (

Kopf and Newman

Page 12

Geobiology

. Author manuscript; available in PMC 2015 September 29.

HHMI Author Manuscript

Kopf and Newman

Page 13

Table 1

Fe(II) oxidation allows carbon fixation

Bulk isotopic composition of

R. capsulatus

after phototrophic growth in freshwater medium containing 22mM H

and amended with acetate only

[A], Fe(II)-NTA only [B], or both [C]. Optical density (OD at 675nm), acetate and ferrous Fe(II) concentrations were measured at inoculation (t=0) and at

the time of harvest (20 hours after inoculation for [A], 29 hours for [B, C]). Values represent averages of biological triplicates [A] and quadruplicates [B,

C] respectively. Reported error is one standard deviation.

Sample

OD (675nm)

Acetate [mM]

Fe(II) [mM]

Bulk

C [%]

Inoculation

Harvest

Start

End

Start

End

Acetate only [A]

0.007±0.001

0.353±0.001

3.02±0.03

<0.1

none

16.8±0.1

Fe(II)-NTA only [B]

0.007±0.001

0.145±0.003

none

4.31±0.08

0.21±0.03

62.8±0.5

Acetate & Fe(II)-NTA [C]

0.008±0.003

0.451±0.026

3.02±0.03

<0.1

4.32±0.13

0.33±0.11

28.1±0.3

Acetate at time of harvest could no longer be detected and is reported to be below the lower limit of determination.

Geobiology

. Author manuscript; available in PMC 2015 September 29.

HHMI Author Manuscript

Kopf and Newman

Page 14

Table 2

Carbon and electron mass balance

Carbon mass balance includes total amounts and isotopic composition of cell carbon derived from the different carbon sources. Electron recovery is based

on total e- donor consumption and cellular e- content. Derived quantities are reported with errors derived by error propagation. Reported error is one

standard deviation.

Carbon sources

Derived cell carbon

Acetate only [A]

Fe-NTA only [B]

Acetate & Fe-NTA [C]

[mM C]

Relative contributions [%C-source]

Inoculum

0.05±0.01

1.1

1.0±0.2%

3.1±0.4%

0.8±0.1%

Acetate assimilation

5.3±0.4

16.9

99.0±0.2%

—

76±2%

Carbon fixation by Fe oxidation

0.95±0.07

99.0

—

56±6%

14±1%

Assimilation of NTA breakdown products

0.7±0.2

18.6

—

41±6%

10±1%

Total biomass C [mM]

5.3±0.4

1.7±0.2

7.0±0.5

Total biomass e

[mM]

22.9±2.4

7.3±0.9

29.9±2.9

Total e-donor e

[mM]

24.4±0.3

7.1±0.8

31.2±0.8

e-recovery [%]

94±10%

103±18%

96±10%

based on e

content of

R. capsulatus

biomass derived in Supplemental Information

assuming substrate from NTA breakdown to be primarily in the oxidation state of formaldehyde (see Table S1)

ratio of total e

recovered in biomass / e

available from e-donors

Geobiology

. Author manuscript; available in PMC 2015 September 29.