Vanadium Requirements and Uptake Kinetics in the Dinitrogen-Fixing Bacterium Azotobacter vinelandii

Abstract

Vanadium is a cofactor in the alternative V-nitrogenase that is expressed by some N₂-fixing bacteria when Mo is not available. We investigated the V requirements, the kinetics of V uptake, and the production of catechol compounds across a range of concentrations of vanadium in diazotrophic cultures of the soil bacterium Azotobacter vinelandii. In strain CA11.70, a mutant that expresses only the V-nitrogenase, V concentrations in the medium between 10⁻⁸ and 10⁻⁶ M sustain maximum growth rates; they are limiting below this range and toxic above. A. vinelandii excretes in its growth medium micromolar concentrations of the catechol siderophores azotochelin and protochelin, which bind the vanadate oxoanion. The production of catechols increases when V concentrations become toxic. Short-term uptake experiments with the radioactive isotope ⁴⁹V show that bacteria take up the V-catechol complexes through a regulated transport system(s), which shuts down at high V concentrations. The modulation of the excretion of catechols and of the uptake of the V-catechol complexes allows A. vinelandii to precisely manage its V homeostasis over a range of V concentrations, from limiting to toxic.

The transition metal vanadium (V) has relatively few known biological functions. It is found in marine algae's haloperoxidase enzymes (5). It also is accumulated by ascidians (11), but its biological function (if any) in these organisms is still mysterious. Most importantly, V is found at the active center of an alternative form of the enzyme nitrogenase, which fixes atmospheric N₂ gas into bioavailable ammonia and is responsible for the natural input of new nitrogen into the earth's ecosystems. The molybdenum (Mo)-nitrogenase, which is the most common and efficient form of the enzyme, has a Mo cofactor at its active site, but when Mo is not available, some bacteria can express an alternative V-nitrogenase, which uses a V cofactor in place of Mo (3, 4). Some organisms also have an Fe-only nitrogenase, which requires only Fe at its active center and is used when neither Mo nor V is available (14, 20).

V is toxic at high concentrations. Unlike most transition metals (but like Mo), the most stable form of V in oxic environments is a negatively charged ion, the oxoanion vanadate (H₂VO₄⁻/H₂VO₄²⁻). Vanadate is a structural and electronic analogue of phosphate, and it competes against phosphate for uptake in freshwater algae (24). It is also a potent inhibitor of phosphorylases (11). In addition, vanadium can be toxic by interfering with iron uptake. At millimolar concentrations, it binds to the strong iron chelators (siderophores) needed for iron uptake in Pseudomonas aeruginosa and possibly catalyzes the production of the toxic superoxide anion O₂⁻ (1). At lower concentrations, it interferes with the activity of ferric reductase, a key enzyme involved in the uptake of Fe-siderophores in the gram-negative soil diazotroph Azotobacter vinelandii. It also induces the accumulation of the tris(catechol) siderophore protochelin in the growth medium (10).

For N₂-fixing bacteria that can express the V-nitrogenase, the management of V uptake is critical, as V may be needed for growth or may be toxic depending on its concentration.

There are few studies of V uptake in microorganisms, but recently a high-affinity uptake system for vanadate was found in the N₂-fixing cyanobacterium Anabaena variabilis (27). This system is an ABC-type transporter similar to those used for Mo uptake in various prokaryotes, including A. vinelandii (25). In another study, we showed that, in A. vinelandii, catechol siderophores are involved in V uptake (J.-P. Bellenger, T. Wichard, A. B. Kustka, and A. M. L. Kraepiel, unpublished data). A. vinelandii excretes in the growth medium various types of catechol ligands, the two monocatechols 2,3-dihydroxybenzoic acid and aminochelin, the bis(catechol) azotochelin, the tris(catechol) protochelin, and azotobactin, which has a hydroxamate group and an α-hydroxycarboxylic acid group in addition to a catechol group. These catechols are siderophores used for iron acquisition. But even when Fe is not limiting, protochelin and azotochelin accumulate at micromolar concentrations in the growth medium and bind vanadate in strong 1:1 complexes (log K_app^V-Az = 8.8 at pH 6.6, where K_app is the apparent complexation constant and V-Az represents the V-azotochelin complex) (2; Bellenger et al., unpublished). Metals bound to organic ligands usually are not available for uptake, as shown for the protochelin and azotochelin complexes of tungsten (W) in A. vinelandii (36). But A. vinelandii cells grown at low [V] can take up the V-azotochelin and V-protochelin complexes, indicating the existence of a specific uptake system for these complexes (Bellenger et al., unpublished). This study investigates the role of the catechol siderophores in the management of V supply in diazotrophic cultures of A. vinelandii under conditions spanning V limitation to V toxicity.

MATERIALS AND METHODS

Reagents.

All chemicals used were of analytical grade from Sigma or Fisher Bioblock and were used without further purification. Azotochelin was synthesized according to published procedures (6, 22). Protochelin was extracted and purified from an A. vinelandii culture medium as previously described (9; Bellenger et al., unpublished). The radioactive isotope ⁴⁹V was purchased from the Los Alamos National Laboratory.

Bacterial strains and growth conditions.

Strain OP (wild type) and strain CA11.70 (a double-deletion mutant [ΔnifHDK ΔanfHD70::kan] expressing only the V-nitrogenase) (18) of A. vinelandii were used in this work. A. vinelandii was grown aerobically at 22°C in a liquid medium containing the following: KH₂PO₄ (5 × 10⁻³ M), K₂HPO₄ (2.3 × 10⁻³ M), CaCl₂·2H₂O (6.8 × 10⁻⁴ M), MgSO₄·7H₂O (4.05 × 10⁻⁴ M), CuCl₂·2H₂O (10⁻⁸ M), MnCl₂·4H₂O (2.25 × 10⁻⁷ M), CoCl₂·6H₂O (2.43 × 10⁻⁸ M), ZnSO₄·7H₂O (5.3 × 10⁻⁸M), FeCl₃·6H₂O (5 × 10⁻⁶ M), EDTA (10⁻⁴ M), glucose (5.5 × 10⁻² M), and d-mannitol (5.48 × 10⁻² M), pH 6.7. The vanadium and molybdenum concentrations were adjusted by additions of NaVO₃ and Na₂MoO₄, respectively. The cultures were continuously agitated at 180 rpm. Bacterial growth was monitored by measuring the optical density at 620 nm (OD₆₂₀) in a 1-cm polypropylene cell on an HP8453E spectrophotometer. The conversion from OD₆₂₀ to cell density ([1.16 ± 0.16] × 10⁸ cells ml⁻¹ OD₆₂₀⁻¹) was obtained by counting 4′,6′-diamidino-2-phenylindole (DAPI)-stained culture aliquots using epifluorescence microscopy (26).

Acetylene reduction assays.

Nitrogen fixation rates were estimated using the acetylene reduction assay (31). Rates were linear during the assay period of 30 min.

Catechol production.

Forty- to 50-ml samples of three independently grown bacterial cultures were centrifuged (16,000 × g, 20 min, 10°C). The supernatant was spiked with the internal standard (20 μl of 3,4-dihydroxy benzaldehyde [200 μg ml⁻¹ in MeOH]). Each sample then was concentrated on an OASIS-HLB cartridge (Waters, Milford, MA), washed with 4 ml water, and then eluted with 4 ml methanol. After concentration under an argon stream (down to 0.4 ml), 50 μl of the concentrated methanolic eluate was analyzed by reversed-phase high-pressure liquid chromatography at pH 2.2 for the determination of catechol compounds. Catechols were separated on a Discovery C₁₈ column (150 by 4.6 mm; 5-μm particle size) equipped with a C₁₈ guard column (20 by 4 mm; 5-μm particle size; Supelco) using a solvent system of water-acetonitrile containing 0.07% trifluoroacetic acid at a flow rate of 1 ml min⁻¹. A gradient from 0% acetonitrile to 25% acetonitrile (7 min) held for 7 min and continued to 50% acetonitrile (5 min) was used. The high-pressure liquid chromatography was developed on a JASCO system equipped with two nonmetallic pumps (PU-1580i) and a diode array detector (MD1515). The maximum absorbance and absorbance at 310 nm (characteristic of catechols) were monitored. Protochelin and azotochelin were identified by UV-visible spectrum spectroscopy and coinjections of standards, as previously described (36).

The limit of quantification for the standards was 0.24 nmol for azotochelin and 0.80 nmol for protochelin on columns corresponding to a concentration of 1.9 × 10⁻⁹ and 6.4 × 10⁻⁹ M in the growth medium, respectively. The variation between technical replicates (subsamples from the same biological replicate) averaged 5% for protochelin and 6% for azotochelin. The recovery of catechol siderophores was determined by analyzing triplicates of the supernatant of the culture medium spiked with 1.8 × 10⁻⁶ M azotochelin and 3.6 × 10⁻⁶ M protochelin. The recovery rate was 84% ± 4% for protochelin and 86% ± 1% for azotochelin after taking into account the amount of these catechols measured in the unspiked medium.

Metal quotas.

A. vinelandii was grown as described above. When the OD₆₂₀ reached 0.8, 8 ml of the culture was collected and filtered on an acid-cleaned polycarbonate filter (25 mm in diameter, 0.6 μm in pore size) placed in an acid-cleaned polypropylene holder (Fisher). The cells were rinsed two times (3 min each) with 1 ml of an oxalate-EDTA-KCl solution (33). Each filter then was placed in a Teflon tube, to which 800 μl of 50% HNO₃ (optima grade) was added. The tubes were sealed tightly and placed in an oven at 90°C overnight. After digestion, the tubes were filled with Milli-Q water to achieve a final volume of 8 ml and then centrifuged at 12,000 × g for 10 min at 20°C. The supernatants were collected, spiked with two internal standards (2 ppb In and 2 ppb Sc), and analyzed for their concentrations of P, V, W, Mo, and Fe with an inductively coupled plasma-mass spectrometer (Element2; Thermo Finnigan) at medium resolution.

Short-term uptake of vanadium.

A. vinelandii was grown at vanadium concentrations ranging from 5 × 10⁻⁹ to 10⁻⁵ M. When the OD₆₂₀ was close to 0.9, 25 ml of the culture was aseptically centrifuged (16,000 × g, 10 min), washed with 10 ml of clean medium, and centrifuged again. The cells were resuspended into 50 ml of fresh culture medium spiked with the radioactive tracer ⁴⁹V and containing various concentrations of vanadium, with or without a ligand (azotochelin, protochelin, or desferrioxamine B [DFB]) added. At regular intervals, aliquots of the bacterial suspension were collected and filtered on a 0.6-μm polycarbonate filter, washed three times with 30 ml of a concentrated solution of cold vanadium ([V] = 10⁻⁵ M), and then washed with 30 ml of an oxalate-EDTA-KCl solution (33) for 2 min. Finally, the cells were washed twice with 30 ml of the cold vanadium solution. The filter was collected and placed in a 20-ml vial filled with 16 ml of scintillation solution (ScintiSafe plus 50%). ⁴⁹V activity (X-ray emission at 5 keV) was determined with a Beckman LS6500 multipurpose scintillation counter.

RESULTS

Vanadium requirements of A. vinelandii.

To study the V requirements of A. vinelandii, for a number of our experiments we used the double mutant strain CA11.70, which expresses only the V-nitrogenase. Under diazotrophic conditions, the growth rate of this mutant is dependent on the [V] (Fig. 1). At a low [V] (<10⁻⁷ M), there is a first phase of rapid growth, followed by a second phase of lower growth (Fig. 1). Using cultures spiked with the radioactive isotope ⁴⁹V, we showed that the second phase, which occurs at higher cell densities as the concentration of V increases, begins when the bacteria have taken up essentially all of the V (>98%) initially present. The lower growth rates in the second phase are associated with lower dinitrogen fixation rates (data not shown), as bacteria use their decreasing cellular stock of V to fix dinitrogen. Because the growth curves of the low-V cultures are not simple exponentials, growth rates were calculated based on the first phase. The range of V concentrations that sustain maximum growth is relatively narrow, as V is limiting below 10⁻⁸ M and toxic above 10⁻⁶ M (Fig. 1 and 2A).

FIG. 1.

Growth curves of A. vinelandii (mutant strain CA11.70) at various V concentrations. (A) V limitation at low concentrations. (B) V toxicity at high concentrations.

FIG. 2.

Growth rates (μ) and intracellular vanadium quotas (Q, in moles of V per cell) of A. vinelandii (mutant strain CA11.70) as a function of [V]. V:P is the intracellular vanadium/phosphorus molar ratio. (A) Growth rates calculated using a log-linear fit of the early part of the growth curves, with or without azotochelin (Az; 10⁻⁴ M) added to the growth medium. The error bars correspond to the standard deviations from two to five independent experiments. (B) Cellular V quotas measured at a cell density of (9 ± 0.6) × 10⁷ cells ml⁻¹. The error bars correspond to the standard deviations from three independent experiments. The line corresponds to the calculated cellular V quotas if all of the vanadium initially present in the medium had been taken up by the bacteria.

We measured the intracellular V quota at a cell density of (9 ± 0.6) × 10⁷ cells ml⁻¹, which corresponds approximately to the end of the first growth phase for low-V cultures. At low V concentrations, intracellular V quotas increase with the V concentration until they reach a plateau between [V] = 10⁻⁷ M and [V] = 10⁻⁶ M (Fig. 2B). For [V] < 6 × 10⁻⁸ M, the cellular V quota closely matches the initial V concentration, since the bacteria take up nearly all available V (Fig. 2B). Above [V] = 10⁻⁶ M, the decreasing growth rates correspond to an increase in the intracellular quota, as the cells apparently cannot regulate vanadium uptake. It is remarkable that a small increase in the V quotas (designated Q), from Q = 5.6 × 10⁻¹⁹ mol cell⁻¹ (at [V] = 10⁻⁶ M) to Q = 7.2 × 10⁻¹⁹ mol cell⁻¹ (at [V] = 10⁻⁵ M), results in a 30% decrease in the growth rates.

Vanadium uptake.

To determine how bacteria regulate their V uptake as the V concentration increases from limiting to toxic levels, we first studied the kinetics of vanadate transport. Cells grown at low V concentrations were collected and resuspended in fresh medium spiked with the radioactive isotope ⁴⁹V at various V concentrations. Approximately constant uptake rates were measured for at least 30 min (Fig. 3A). The kinetics of vanadate uptake followed Michaelis-Menten kinetics, with K_m^V ∼ 10⁻⁷ M and V_m^V ∼ 4.6 × 10⁻²⁰ mol cell⁻¹ min⁻¹ (where V_m is the maximum rate).

FIG. 3.

Short-term uptake of vanadium by A. vinelandii. (A) Short-term uptake of unchelated vanadate ([V] = 1.5 × 10⁻⁸ M), V-azotochelin ([V] = 1.5 × 10⁻⁸ M; azotochelin concentration, 10⁻⁶ M), V-protochelin ([V] = 1.5 × 10⁻⁸ M; protochelin concentration, 10⁻⁶ M), and V-DFB ([V] = 4 × 10⁻⁸ M; DFB concentration, 3 × 10⁻⁴ M) by the mutant strain CA11.70. The V concentration in the resuspension medium was identical to the V concentration used for preconditioning the bacteria. (B) Short-term uptake of unchelated vanadate ([V] = 2 × 10⁻⁸ M) and V-azotochelin ([V] = 2 × 10⁻⁸ M; azotochelin concentration, 10⁻⁴ M) by the wild-type (WT) strain. The bacteria were preconditioned at [V] = 2 × 10⁻⁸ M and [Mo] < 3 × 10⁻⁹ M (circles) or [V] = 2 × 10⁻⁸ M and [Mo] = 6 × 10⁻⁸ M (triangles). Az, azotochelin.

Interestingly, the uptake rate (v) of vanadate is regulated by molybdenum: in the wild type, cells grown at low Mo and V concentrations take up unchelated vanadate very rapidly (v = 3.4 × 10⁻²¹ mol cell⁻¹ min⁻¹), whereas in cells grown at high Mo, the uptake rate is below our detection limit (v < 8 × 10⁻²³ mol cell⁻¹ min⁻¹) (Fig. 3B).

Free vanadate, however, is not the main form of vanadium in the growth medium, because of the complexation with the catechols azotochelin and protochelin released into the growth medium, except at the very beginning of the batch cultures (Bellenger et al., unpublished, and see below). The bioavailability of the catechol complexes was confirmed in an experiment in which cells preconditioned at a low [V] were resuspended in fresh medium containing V and an excess of azotochelin ([V] = 1.5 10⁻⁸ M; azotochelin concentration, 10⁻⁶ M) (Fig. 3A). In the resuspension medium, the concentration of unchelated vanadate ([V′]) is considerably reduced because of the binding of V to azotochelin ([V′] = 2.4 × 10⁻¹¹ M; log K_app^V-Az = 8.8 at pH 6.6) (2). If unchelated vanadate were the only form available for uptake, the uptake rate, calculated on the basis of K_m^V and V_m^V (Table 1), would be 1.1 × 10⁻²³ mol cell⁻¹ min⁻¹, which is about 200 times lower than the measured uptake rate (v = 2.6 × 10⁻²¹ mol cell⁻¹ min⁻¹) (Fig. 3A). A similar result is obtained with protochelin, the affinity of which for V must be similar to that of azotochelin, since it binds V with the same functional groups and the same stoichiometry (Bellenger et al., unpublished). In another study, the availability of the V complexes with azotochelin and protochelin for uptake was demonstrated directly (Bellenger et al., unpublished).

TABLE 1.

K_ms of vanadate and molybdate uptake systems

Species	Km (V)	Km (Mo)
Azotobacter vinelandii	10−7a	∼10−9c
Anabaena variabilis	2.8 × 10−9b	3.3 × 10−10b
Escherichia coli	NDe	2.5 × 10−8d

^aDetermined in this work.

^bFrom Pratte and Thiel (27).

^cFrom Mouncey et al. (23).

^dFrom Corcuera et al. (7).

^eND, not determined.

To get a better understanding of the regulation of the uptake of the V-catechol complexes, we measured short-term uptake rates of the V-azotochelin complex as a function of the V concentration in cells preconditioned at a low [V] (Fig. 4A). Cells grown at [V] = 10⁻⁸ M for several generations were resuspended in fresh growth medium with various concentrations of vanadate and an excess of azotochelin. At a low [V], the uptake of V-azotochelin follows simple Michaelis-Menten kinetics, with K_m^V-Az = 2 × 10⁻⁸ M and V_m^V-Az = 5.5 × 10⁻²¹ mol cell⁻¹ min⁻¹ (note the log scale in Fig. 4A). But for [V] > 6.5 × 10⁻⁸ M, the cells down-regulate the uptake of the complex within the first 2 min of the short-term uptake experiment, resulting in decreasing uptake rates with increasing V concentrations. The decrease in uptake rates between V concentrations of 6 × 10⁻⁸ and 7 × 10⁻⁸ M is extremely sharp and becomes more gradual between 7 × 10⁻⁸ and 10⁻⁶ M. Above [V] = 10⁻⁶ M, when V is toxic, the uptake rate of the V-azotochelin complex is below our detection limit.

FIG. 4.

Short-term uptake rates of V-catechol complexes by A. vinelandii (mutant strain CA11.70) as a function of V concentrations. (A) Uptake rates (± standard deviations; n = 2) of the V-azotochelin complex. Bacteria were preconditioned at [V] = 10⁻⁸ M. (B) Uptake rates (± standard deviations; n = 2 to 4) of V-azotochelin (circles) and V-protochelin (triangles). V concentrations in the preconditioning and resuspension media were identical. Az, azotochelin; proto, protochelin.

In another series of experiments, the short-term uptake rate of V-azotochelin was measured in A. vinelandii cells (mutant strain CA11.70) preconditioned at various V concentrations and resuspended at the same [V] as that used for preconditioning (Fig. 4B). The uptake rates for V-azotochelin measured under these conditions are very close to those measured in cells preconditioned at [V] = 10⁻⁸ M and resuspended at various V concentrations. Most importantly, they exhibit the same sharp decrease near [V] = 6 × 10⁻⁸ M (compare Fig. 4A to B), demonstrating that the rapid down-regulation at V concentrations greater than 6 × 10⁻⁸ M is maintained in the long term. In addition to being regulated by V, the uptake system for the V-azotochelin complex (like that of free vanadate) is regulated by Mo in the wild type and is activated only in cells grown at low Mo concentrations (Fig. 3B).

In an experiment similar to the one just described for V-azotochelin, cells preconditioned at various V concentrations were resuspended in fresh medium with the same V concentration as that used for preconditioning the cells and containing protochelin (Fig. 4B). The uptake rates for the V-protochelin complex are comparable to those for V-azotochelin, but they are consistently lower, except possibly at very low V concentrations. They also follow the same general trend and decrease rapidly for V concentrations above 6 × 10⁻⁸ M.

The kinetic data show that A. vinelandii takes up the complexes of V with azotochelin and protochelin, two catechols that it excretes in its growth medium. To determine whether the bacterium also can take up other organic complexes of V, we investigated the uptake of the V complex with desferrioxamine B (DFB). DFB is a tris(hydroxamate) siderophore that binds vanadate in the conditions of the experiment (log K_app = 7.0 at pH 6.7) (15). Cells preconditioned at a low V concentration and resuspended in fresh medium containing the V-DFB complex do not take up vanadium (Fig. 3A), thus showing that the V-DFB complex is not available to the bacteria. Similar results were obtained using aerobactin, a tridentate siderophore with two hydroxamate and one α-hydroxycarboxylic acid groups (data not shown). Neither DFB nor aerobactin is produced by A. vinelandii.

Production of catechols.

Cultures of A. vinelandii excrete in their growth medium large amounts of azotochelin and protochelin, which bind vanadate. Catechol concentrations were greater than the detection limit after 20 h in a typical culture, and they increased steadily with cell densities (Fig. 5C). The catechol concentrations of cultures grown at various V concentrations were compared between the mutant and the wild-type strains. In cultures of the wild type growing at V concentrations below the toxicity threshold ([V] < 10⁻⁶ M), the sum of protochelin and azotochelin concentrations in the medium is roughly constant at the average value of (2.1 ± 0.6) × 10⁻⁶ M. In cultures of the mutant strain, the catechol concentrations decrease with the V concentrations, from concentrations that are similar to those of the wild-type cultures to below the detection limit at [V] = 7.8 × 10⁻¹⁰ M (Fig. 5A and B). With the possible exception of the cultures of the mutant at very low [V], which grow poorly, the catechol ligands excreted in the medium are in large excess of the V concentration in the medium.

FIG. 5.

Production of catechol siderophores by A. vinelandii. (A) Concentrations (± standard deviations; n = 3) of catechols produced by the mutant strain CA11.70 at a cell density of 4.8 × 10⁷ cells ml⁻¹ (OD₆₂₀ = 0.41 ± 0.05) as a function of the V concentration in the medium. n.d., not detected (i.e., below the detection limit). (B) Concentrations (± standard deviations; n = 3) of catechols produced by the wild-type strain (OP) at a cell density of 5.6 × 10⁷ cells ml⁻¹ (OD₆₂₀ = 0.48 ± 0.07) as a function of the V concentration in the medium. (C) Production of protochelin as a function of time and the optical density in a culture of the mutant strain CA11.70 grown at [V] = 10⁻⁸ M and 30°C.

When V becomes toxic at 10⁻⁵ M, the catechol concentrations in the medium increase sharply in both the wild-type and mutant strains, but the excess of total catechol concentrations above the V concentration decreases as the [V] increases; at [V] = 10⁻⁵ M, there is just enough (1.47 × 10⁻⁵ M) to complex all of the V, particularly so since some catechols become bound to Fe (Bellenger et al., unpublished). That might be the cause of the observed increase in V cellular quota and toxicity at [V] = 10⁻⁵ M. A similar series of experiments thus was performed across a range of V concentrations in a medium amended with 10⁻⁴ M azotochelin. Up to [V] = 10⁻⁶ M, the growth rates and V quotas in cultures with and without added azotochelin are identical (Fig. 2). Above [V] = 10⁻⁶ M, the V cellular quotas in the presence of azotochelin are significantly lower than those in its absence, and the growth rates are correspondingly higher.

DISCUSSION

Our results show that the model N₂-fixing bacterium A. vinelandii modulates its excretion of catechols and the uptake rate of the V-catechol complexes to precisely control its acquisition of V. The bacterium regulates its uptake of the V-catechol complexes (Bellenger et al., unpublished) such that, depending on conditions, the production of catechols can either increase or decrease the rate of V uptake.

At all V concentrations (except possibly at a very low [V] in the mutant strain), the bacteria excrete in their growth medium an excess of catechol compounds (azotochelin and protochelin) that bind vanadate. Unchelated vanadate, thus, is never the major species in the growth medium, except at the very beginning of the batch cultures. For V concentrations below 6.5 × 10⁻⁸ M, the bacteria take up the V-azotochelin and V-protochelin complexes at their maximum uptake rates. At these concentrations, the bacteria also deplete their medium of V, and their cellular quotas essentially reflect the amount of V initially available (Fig. 2B and 4). For V concentrations close to 6.5 × 10⁻⁸ M, the uptake of the catechol complexes decreases abruptly by a factor of three (Fig. 4). At about the same threshold, the cells stop taking up all of the V initially present in their medium, as reflected by their V quotas (Fig. 2B). For 6.5 × 10⁻⁸ M < [V] < 10⁻⁶ M, V uptake rates continue decreasing, and the intracellular V quotas are approximately constant. The bacteria, thus, are able to grow at maximum rates for V concentrations ranging from 10⁻⁸ to 10⁻⁶ M by maintaining maximum V uptake rates up to [V] = 6.5 × 10⁻⁸ M and reducing their V uptake rate above that threshold. At an even higher V concentration, somewhere between 10⁻⁶ and 10⁻⁵ M, V becomes toxic for growth, and lower growth rates correspond to increased cellular quotas despite a complete shutting down of the uptake system for the catechol complexes (Fig. 2 and 4). Concomitantly, the concentrations of azotochelin and protochelin increase sharply, confirming the large increase in protochelin concentrations at high [V] previously observed by Cornish and Page (10). At [V] = 10⁻⁵ M, the growth rate of the bacteria is significantly reduced. Based on the concentrations of azotochelin and protochelin in the growth medium at that V concentration and on their affinity for vanadate (log K_app^V-Az = 8.8), we calculate a concentration of unchelated vanadate close to 4 × 10⁻⁹ M. This corresponds to an uptake rate of 1.8 × 10⁻²¹ mol cell⁻¹ min⁻¹ (calculated on the basis of the constants of Table 1), which, considering a growth rate of 0.18 h⁻¹, is enough to sustain a cellular quota that is close to the toxic threshold (Q = 6 × 10⁻¹⁹ mol cell⁻¹). Thus, it appears that while the sharp increase in catechol concentrations between 10⁻⁶ and 10⁻⁵ M results in a reduced toxicity of the metal, it is not enough to alleviate its toxicity completely.

The management of V nutrition is the result of three complementary processes: (i) the release of catechols by the bacteria into the medium; (ii) the complexation of V by these catechols; and (iii) the regulation of uptake of the V-catechol complexes. Each of these processes is discussed below.

Catechol production.

In A. vinelandii, the biosynthesis of catechol siderophores was found to be under the dual control of the ferric uptake regulator (Fur) and another protein responding to superoxide stress (34). At high V levels, the excretion of large concentrations of catechol siderophores might result from a low Fe cellular quota, as V interferes with Fe uptake (1, 10). Alternatively, the availability of V itself might regulate catechol production. The increase in azotochelin and protochelin concentrations at very low V concentrations in cultures of the wild type is not observed for the mutant that has only the V-nitrogenase. This perhaps reflects its very poor growth at very low V concentrations or might result from a higher demand for iron in the wild type, which presumably grows with the Fe-only nitrogenase. Clearly, the regulation of siderophore production in A. vinelandii is a subject worthy of further investigation.

V complexation in the medium.

Azotochelin and protochelin are among the few ligands known to bind strongly to V in solution at circumneutral pH. Although the exact coordination of H₂VO₄⁻ to these catechols in aqueous solution is not known, that of MoO₄²⁻ is affected by the addition of one catechol moiety and the displacement by another of two oxygen atoms from the Mo coordination sphere. Because these catechols have high affinities for most transition metals, including Fe, it is difficult to predict which metals they bind when excreted in the external medium. At equilibrium, we calculate that Fe-catechol complexes are dominant because of the higher affinity of catechols for iron, but because of different kinetics of reactions, less stable complexes actually form during transient time scales. For example, in a solution at pH 7, Fe precipitates as iron oxides, while molybdate remains in solution. Azotochelin first complexes Mo, which then is slowly exchanged for Fe during a timescale of several hours to a few days (2, 8, 13). Similarly, in the growth medium of A. vinelandii, only a small fraction of iron is bound to protochelin (Bellenger et al., unpublished) because of the slow exchange kinetics between Fe-EDTA and protochelin (9, 16, 17). Complexation with protochelin thus dominates V speciation even as the total protochelin concentration in the medium is lower than the sum of the Fe and V concentrations. The effect of siderophores on metal speciation in soils and aquatic environments thus may be as dependent on kinetics as on equilibrium conditions.

Uptake of V complexes.

Understanding the mechanisms and regulation of the uptake of the V-catechol complexes requires the identification of the proteins and pathways involved. Nonetheless, our kinetic data provide preliminary insight.

The K_m for the V-azotochelin complex (K_m^V-Az ∼ 2 × 10⁻⁸ M) is lower than that for free vanadate (K_m^V ∼ 10⁻⁷ M), which suggests that azotochelin is used to acquire V in low-V environments. Both K_m^V-Az and K_m^V in A. vinelandii are higher than the K_m^V in Anabaena variabilis (Table 1), which may reflect the different levels of availability of V in the natural environments of these bacteria. Further, the K_m for vanadate transport usually is higher than that for molybdate, the other oxoanion used for dinitrogen fixation (Table 1).

The uptake system(s) for the V-azotochelin and V-protochelin complexes apparently is specific, since neither the V complexes with DFB and aerobactin nor the W complexes with azotochelin and protochelin are taken up (36). It also is regulated by the concentration of V (Fig. 4). We observed very fast down-regulation upon the exposure of cells to high [V]. However, cells grown at high [V] and resuspended in a low-V medium do not take up the V-azotochelin complex (Bellenger et al., unpublished), showing that the uptake system is not rapidly activated, either because it responds to the intracellular V status, because it requires the synthesis of the appropriate transporters, or both. In addition, the uptake of V in the wild type, as in the case of the expression of the V-nitrogenase, is regulated by the concentration of Mo (Fig. 3). This regulation, the genetic mechanism of which remains to be investigated, is not unique to A. vinelandii. In another N₂-fixing organism, the cyanobacterium Anabaena variabilis, the genes encoding a high-affinity ABC-type transport system for vanadate are repressed by molybdate (reference 27 and see below).

The V-azotochelin and V-protochelin complexes are probably, by analogy with the Fe-siderophore complexes, too large to cross the outer membrane of the bacteria through porins, and it is likely that there is a transporter located at the outer membrane to transport V into the periplasm at low [V]. What happens to V in the periplasmic space of A. vinelandii is not known, but the catechols azotochelin and protochelin have been shown to be involved in Mo uptake as well as V uptake (Bellenger et al., unpublished), and molybdate in the periplasm is transported by a well-characterized ABC-type transporter (19, 23, 28). Such transporters commonly are used for oxoanion uptake (21, 28, 30, 32), and one recently was identified for vanadate transport in A. variabilis (27). If such a transport system also exists in A. vinelandii, V uptake may be very similar to Mo uptake.

Multiple roles of catechols.

Together with previously published data, our results show that the catechol siderophores of A. vinelandii are used in multiple ways by the bacterium. Their role in iron acquisition has been extensively studied (9), and they are also used for Mo uptake (Bellenger et al., unpublished). In addition, azotochelin and protochelin can capture V from unavailable complexes with DFB or natural organic matter and, under V-limiting conditions, can promote vanadium uptake by the bacterium (Bellenger et al., unpublished). They can serve as vanadophores as well as siderophores and molybdophores. Further, our results showing a marked increase in siderophore production and a down-regulation of the uptake of the V complexes over a range of metal concentrations demonstrate that the apparent detoxification at high [V] is not incidental but part of an overall strategy in which A. vinelandii uses the unique coordination chemistry of catechols to carefully manage its metal homeostasis. The excretion of catechols and the accurate regulation of the uptake of the V-catechol complexes result in an efficient and controlled management of the vanadium supply.

There is evidence that Mo limits dinitrogen fixation in tropical and temperate ecosystems (29; A. R. Barron, A. M. L. Kraepiel, S. J. Wright, and L. O. Hedin, unpublished data), so that the alternative V-nitrogenase may play a more important role than previously recognized. The average V concentration in soils is about 100 times higher than that of Mo (35), but most of the V is complexed by natural organic matter or adsorbed at the surface of soil particles and is unavailable for biological uptake. The excretion of V-binding catechols should be particularly useful in natural environments that contain high concentrations of V-binding compounds, such as the O horizon of soils, in which heterotrophic N₂-fixing bacteria often are most active (12).