The tungsten metallome of Pyrococcus furiosus

Abstract

The tungsten metallome of the hyperthermophilic archaeon Pyrococcus furiosus has been investigated using electroanalytical metal analysis and native–native 2D-PAGE with the radioactive tungsten isotope ¹⁸⁷W (t_1/2 = 23.9 h). P. furiosus cells have an intracellular tungsten concentration of 29 μM, of which ca. 30% appears to be free tungsten, probably in the form of tungstate or polytungstates. The remaining 70% is bound by five different tungsten enzymes: formaldehyde ferredoxin oxidoreductase, aldehyde ferredoxin oxidoreductase, glyceraldehyde-3-phosphate ferredoxin oxidoreductase and the tungsten-containing oxidoreductases WOR4 and WOR5. The membrane proteome of P. furiosus is devoid of tungsten. The differential expression, as measured by the tungsten level, of the five soluble tungsten enzymes when the cells are subjected to a cold-shock shows a strong correlation with previously published DNA microarray analyses.

---

Introduction

Tungsten is the bioelement with the highest atomic number, but it is not a universal bioelement. For several microbiological species tungsten is essential for growth, e.g. the hyperthermophilic archaeon Pyrococcus furiosus¹. P. furiosus is a marine anaerobic , fermentative archaeon, which grows optimally at 100 °C. It can use peptides or polycarbohydrates as a carbon source, producing mainly acetate with the production of molecular hydrogen from protons as terminal electron acceptor.^2–4 It can also reduce elemental S⁰ to H₂S.

In the past few decades, five different tungsten-containing enzymes have been isolated from this organism, all members of the aldehyde oxidoreductase (AOR) family, namely: formaldehyde ferredoxin oxidoreductase (FOR), aldehyde ferredoxin oxidoreductase (AOR), glyceraldehyde-3-phosphate ferredoxin oxidoreductase (GAPOR), tungsten-containing oxidoreductase 4 (WOR4) and tungsten-containing oxidoreductase 5 (WOR5).^5–9 AOR has been reported to have the highest activity on aldehydes presumably derived from amino acids, and FOR has the highest activity on C1–C3 aldehydes and semi- and di-aldehydes.⁶ GAPOR is specific for the conversion of the glycolytic intermediate glyceraldehyde-3-phosphate.⁷ For WOR4 no substrate has yet been identified.⁸ WOR5 has broad substrate specificity and has been found to convert several substituted and non-substituted aliphatic and aromatic aldehydes.⁹

DNA microarray experiments have shown that the transcript levels of AORs such as WOR4, FOR and GAPOR are differentially regulated during growth on different carbon-sources and that the levels of WOR5, WOR4, and AOR change during cold-shock stress experiments.^10,11 Cold-shock here means that the cultures are shocked by rapidly dropping the temperature from 95 °C to 72 °C and keeping them at this temperature for 1 to 5 h. Alternatively, the cultures can also be gradually adapted to cold (72 °C). The transcription of WOR4 was up-regulated in cold-adapted cells that were grown for many generations at 72 °C. WOR5 mRNA levels were significantly up-regulated in the case of a short (1 to 2 h) or a prolonged shock (4 to 5 h), while AOR mRNA levels decreased.¹¹

In the present paper, we present a comprehensive study of the tungsten metallome in P. furiosus. A sensitive electroanalytical technique was used to determine the intracellular free and protein bound tungsten concentrations, and a native–native 2D-PAGE technique, which was recently developed, was used with the short-lived radioisotope ¹⁸⁷W to visualize the tungsten containing proteins.¹² We identify the tungsten metalloproteome of P. furiosus under different growth conditions, and compare these data to previous transcriptomics data.

Experimental

Growth of the organism

P. furiosus (DSM 3638) cultures were grown anaerobically at 95 °C using potato starch as carbon source as described previously.² Two batchwise growth conditions were used for comparison: (1) ‘normal’ growth at 95 °C for 14 h, (2) cold-shock. The cold-shock cultures were subjected to 72 °C for 5 h after normal growth. The growth was performed in 100-ml serum bottles, filled with 50 ml growth medium. A volume of 1 ml of the ¹⁸⁷W solution described below was added to the medium at the start of the cultivation period.

Preparation of the radioisotope solution

A volume of 7.5 μl of a 20 g l⁻¹ aqueous solution of (NH₄)₂WO₄ was pipetted into a quartz vial and was evaporated to dryness on a hot plate for 1 h. The vial containing (NH₄)₂WO₄ powder was sealed. Metal radioisotope ¹⁸⁷W (t_1/2 = 23.9 h) was produced by neutron irradiation (thermal neutron flux 5.0 × 10¹⁶ m⁻² s⁻¹) of the sealed quartz vial at the Reactor Institute Delft (Delft, The Netherlands) for 10 h. After irradiation, the vial was allowed to ‘cool down’ at room temperature for 10 h. The vial content was dissolved in 1 mL of 50 mM Tris–HCl pH 8.0, and was added to the growth medium. A small amount of the radioisotope-containing solution was measured in an automatic gamma counter (Wizard 1480-011; Wallac Turku, Finland) in order to calculate the amount of ¹⁸⁷W at the start of the experiment and the purity of this radioisotope.

Protein sample preparation for native 2D-PAGE

After cultivation, the P. furiosus cells were spun down using a Jouan CR 4.11 centrifuge at 3500 rpm at 5 °C for 1 h, and were resuspended in 1 ml 50 mM Tris–HCl pH 8.0 containing 0.1 mg l⁻¹DNase and 0.1 mg l⁻¹ RNase. The cells were broken by osmotic shock. A homogenized native soluble protein extract was obtained after vortexing with a home-made bead beater at 2500 rpm at 5 °C for 30 min, and broken cells were spun down in an Eppendorf centrifuge at 14 000 rpm at 5 °C for 1 h. The protein concentration was determined with the bicinchoninic (BC) acid assay kit (Uptima-Interchim) according to the manufacturer’s instructions.

Membrane proteinsample preparation for native 2D-PAGE

As described above, the cells were spun-down and resuspended in 10 ml 50 mM Tris–HCl pH 8.0 containing 0.1 mg l⁻¹DNase and 0.1 mg l⁻¹ RNase. Cell debris and unbroken cells were removed by centrifugation at 3000 rpm at 5 °C for 30 min, and supernatant was decanted and centrifuged at 100 000 × g at 5 °C for 2 h using a Beckman L8-70 ultracentrifuge. The membrane pellet was washed once with 10 ml of 1 M NaCl by resuspending the pellet using the bead beater at 2500 rpm at 5 °C for 1 h. A new membrane pellet was obtained by centrifugation at 100 000 × g at 5 °C for 2 h. The washed membrane pellet was resuspended and homogenized in a buffer containing 1% CHAPS and 750 mM aminocaproic acid in 50 mM Tris–HCl pH 7.0, and this formed the membrane protein fraction. The protein concentration was determined using a bicinchoninic acid assay kit. A Wallac automatic gamma counter (Perkin Elmer) was used to quantify the radioactivity retained in the membranes after each washing step.

Native 2D-PAGE

The first and second dimensions of the native–native polyacrylamideelectrophoresisgel (2D-PAGE) were carried out as previously described in ref. 12. In brief: native isoelectric focusing (IEF), was performed using a Multiphor II flatbed electrophoresis system (GE Healthcare) at 15 °C. Linear immobilized pH gradient (IPG) gel strips pH 3–11 and pH 4–7, 240 mm were used for the native protein separation based on iso-electric points (pI). In both cases the protein sample was mixed 1 : 1 with rehydration solution prior to IEF. The rehydration solution was composed of a 1 : 7 mixture of solution A (4% 3-[(3-cholamidopropyl)dimethylammonio] propanesulfonic acid (Chaps), 0.5% IPG buffer, 6.2 mg ml⁻¹1,4-dithiothreitol) and solution B (2% Chaps, 0.5% IPG buffer, 0.002% bromophenol blue). Subsequently the sample was applied at pH 4.2 by cup-loading. IEF of 50–550 μg protein sample was performed at 3500 V for 12 h (continuously) after an initial 1 min at 500 V ramping up to 3500 V in 1.5 h (gradient). After the first-dimension separation the IPG strips were equilibrated as previously described.¹² The second-dimension: blue native PAGE (BN 2D-PAGE), was performed on the same electrophoresis system, but at 5 °C using precast homogenous 12.5% gels (ExcelGel 2D, GE Healthcare). The BN-PAGE was run in two steps: 120 V for 1.5 h at 20 mA and 30 W, followed by 600 V for 2.5 h at 30 mA and 30 W. Autoradiograms were taken from the IPG strips before and after the second dimension, and from the second dimension BN-PAGE gel .

Autoradiography of the 2D-PAGE gels

Amounts of ¹⁸⁷W were quantified using a multipurpose phosphor screen (12.5 cm × 25.2 cm), a phosphor screen reader model Cyclone, and OptiQuant software, all purchased from Perkin Elmer. Phosphor screens were scanned with a resolution of 600 dots per inch (dpi). The image resolution was reduced by integration over 10 by 10 points resulting in a final resolution of 60 dpi using Matlab 7.4. Subregions were selected for three-dimensional (3D) visualization using Tecplot 360 2008. The phosphor screens were calibrated as previously described, the only difference being that ¹⁸⁷W was used instead of ⁶⁴Cu.¹² A calibration curve from the ¹⁸⁷W standard solution with a known specific activity was used to quantify the amount of W in spots produced in the IEF and native 2D-PAGE experiments. IEF, 2D-PAGE gels , and calibration spots were read with the same phosphor screen in order to allow for quantitative correction for the decay of ¹⁸⁷W. Following alignment of the autoradiogram and the original gel , the separated spots were excised, digested with trypsin and analyzed with mass spectrometry. After excision of the spots, the 2D-PAGE gel was analyzed again with a phosphor screen after 3 to 4 half-lives to confirm the efficiency of spot excision.

Quantification of the expressed tungsten proteins

From autoradiography, each of the separated spots was quantified based on the amount of intensity within its defined region in digital light units (DLU). Using the calibration curve of ¹⁸⁷W, DLU could be converted to tungsten amounts (pmol). By correcting for the protein losses during the electrophoresis steps, based on ¹⁸⁷W quantification, the absolute expression levels of the W proteins were estimated for the different conditions of growth.

Tandem MS analysis and protein identification

After the radioactivity had decayed for more than 15 half-lives, the excised protein spots were in-gel digested with trypsin (10 ng μl⁻¹ in 25 mM ammonium carbonate, pH 8.1) overnight at 37 °C.¹³ Before digestion, proteins were reduced with 5 mM 1,4-dithiothreitol in 25 mM ammonium carbonate, pH 8.1 and subsequently alkylated with 10 mM iodoacetamide in 25 mM ammonium carbonate, pH 8.1. In-gel digests were acidified with 5% formic acid and were analyzed using nano liquid chromatography (LC) electrospray ionization (ESI) quadrupole time-of-flight (QTOF) tandem MS (MS/MS). Nano-LC was performed using a nanoAcquity UPLC system (Waters, Manchester, UK) directly coupled to a QTOF premier (Waters, Manchester, UK). Peptides were delivered to a trap column packed in-house with ReproSil-Pur C18-AQ, 3 μm, 120 Å, (Dr Maisch, Ammerbuch, Germany) at a flow of 4 μl min⁻¹ in 100% solvent A (0.1 M acetic acid in water) for 10 min. After trapping, the trap column was switched in-line with the analytical column (Reprosil C18, 50 μm inner diameter × 200 mm length, 3 μm particles, packed in-house) and the flow was reduced to 150 nl min⁻¹. A linear gradient from 0 to 40% solvent B (0.1 M acetic acid in 8 : 2 v/v acetonitrile–water) at 1% min⁻¹ was used to analytically separate the content of the trap column. The column effluent was directly electrosprayed in the ESI-source of the mass spectrometer using a nano-ESI emitter (New Objective, Woburn, MA, USA). The mass spectrometer was set in a data dependent mode. Survey scans were acquired in positive ion centroid mode from m/z 400–1500 at a scan time of 1 s using an interscan time of 0.02 s. Up to three multiply charged precursors were allowed for low energy collision induced dissociation at a selection threshold of 1000 counts/s. MS/MSspectra were acquired in continuous mode for 1 s. ProteinLynx Global server 2.2.5 (Waters, Manchester, UK) was used to extract peak lists of the acquired MS/MSspectra. Proteins were identified by MS/MS ion search by using an in-house licensed Mascot 2.0 server (http://www.matrixscience.com). Searches were done in the Swiss-Prot NCBInr (NCBInr 071207) database using the following parameters: P. furiosus as taxonomy restriction, trypsin as enzyme (one miscleavage allowed), peptide tolerance 15 parts per million (ppm), fragment tolerance 0.05 Da, CAM-cysteine as fixed and ox-methionine as variable modifications.

FOR activity measurement after native–native 2D-PAGE

P. furiosus was cultivated on cold (NH₄)₂WO₄ and radioactive (NH₄)₂¹⁸⁷WO₄ in a one to one molar ratio. This ratio of radioactive and non-radioactive tungsten was used to make sure that at least 50% of the tungsten enzyme molecules still contain tungsten after radioactive decay. ¹⁸⁷W decays to ¹⁸⁷Re, which will most likely result in an inactive enzyme. Native–native 2D-PAGE was performed and autoradiography was used to visualize the tungsten-containing proteins. The FOR-containing spot was excised and, after the radioactivity had decayed, analysed for enzymatic activity as previously described.⁵ FOR activity was determined at 80 °C, under anaerobic conditions, with formaldehyde as the substrate and 3 mM methyl viologen (ε₆₀₀ = 10.6 mM⁻¹ cm⁻¹) as the electron acceptor in 50 mM EPPS buffer, pH 8.4.¹⁴

Determination of the intracellular free and protein-bound tungsten concentration

The tungsten content of proteins and free tungstate in the cell was determined by catalytic adsorptive stripping voltammetry.¹⁵ Three 100-ml serum bottles, filled with 50 ml growth medium were used for overnight (14 h) cultivation of P. furiosus at 95 °C. The cells (140 mg) were spun down as described above and the pellet was washed with 15% NaCl to remove extracellulartungstate, without disrupting the cells. The cells were introduced into an anaerobic glove box, and resuspended in 1.5 ml 50 mM Tris–HCl pH 8.0 containing 0.1 mg l⁻¹DNase and 0.1 mg l⁻¹ RNase. Cell debris and unbroken cells were removed by centrifugation at 14 000 rpm at 5 °C for 1 h using an Eppendorf centrifuge, and the supernatant was concentrated using a Microcon filter (Millipore) with 3 kDa cut-off. The protein-containing samples were digested as previously described with 10% (w/v) perchloric acid.¹⁵ Precipitated proteins were removed by centrifugation at 14 000 rpm for 10 minutes. The tungsten content was measured in the filtrate, the concentrate, and the original cell-free extract. The intracellular volume was calculated using the ratio of 4.5 μl volume mg⁻¹ of protein for P. furiosus as reported previously.¹⁶

Results

Method optimization

For each neutron irradiation of (NH₄)₂WO₄ 60 ± 5 MBq (Bq is Becquerel) ¹⁸⁷W was obtained. On average a specific activity of 7.8 × 10¹⁵ Bq/g tungsten was present in the growth media. An average half-life of 23.75 hours was found for ¹⁸⁷W, which is in a good agreement with the reported half-life of 23.90 h.¹⁷ In order to be able to convert the phosphorescence intensity from the screens in DLU to radioactivity and amounts of tungsten a calibration curve was made (see ESI Fig. S1† ). The linear correlation between the DLU intensity values and radioactivity was used to quantify the amount of radioactive tungsten present in total in each of the electrophoresis steps (i.e. transfer efficiency) and in each of the separated spots individually.

The recovery of ¹⁸⁷W after the different electrophoresis steps was a measure of the recovery of native tungsten-containing proteins. In the electrophoresis steps the 50 μl (the required volume for use with a 24 cm IPG strip) of the protein sample is defined as 100% at the beginning of IEF. Applying the cup loading overnight, 59 ± 11% was recovered on the IPG strips. Between the first and second dimension there is a 15 ± 3% loss of protein in the equilibration solution, and 20 ± 2% is not transferred and remains on the strip. From the initial amount added, 20 ± 2% ended up in the BN 2D-PAGE gel . These results are based on four independent experiments.

In order to assess whether the tungsten proteins, as separated above, still possessed catalytic activity, formaldehydeoxidation activity was measured in a separated spot from the native–native 2D-PAGEgel . From the same batch of soluble-protein extract three gels were run, and spots containing FOR were cut out from the 2D-PAGE gel . The formaldehyde oxidation activity of anaerobically purified FOR has previously been reported to be 42 U mg⁻¹ or 27 U mg⁻¹ at 80 °C.^5,18 In this study we measured an activity of 1.3 U mg⁻¹ at 80 °C for this oxygen sensitive enzyme isolated using native–native 2D-PAGE under aerobic conditions. FOR, like all other tungsten enzymes, is sensitive towards oxidation of the tungsten cofactor, rendering the enzyme irreversibly inactive, but the inactivated tungsten cofactor remains protein bound.⁶

The soluble tungsten metallome of P. furiosus

The P. furiosus cell has an intracellular concentration of free tungsten (<3 kDa) of 8.5 ± 0.2 μM and a protein bound (>3 kDa) tungsten concentration of 20.5 ± 0.5. Apparently, about 30% of the total tungsten is not bound to any protein and is most likely in the form of tungstate or possibly polytungstates. The distribution of the protein-bound tungsten in P. furiosus was determined using native–native 2D-PAGE. Fig. 1 shows the ¹⁸⁷W intensity profile of the 1D IEFgel of P. furiosus soluble protein extract using IPG strips with pH 3–11 and pH 4–7. From the autoradiography profiles of the IEFgel , pH 3–11, it is seen that all tungsten-containing proteins have a pI < 7. To increase the separation resolution all subsequent experiments were conducted using 240 mm IPG strips pH 4–7.

Fig. 1 Autoradiogram of native IEF of soluble protein extract of P. furiosus. The extract (280 μg protein) was obtained from cells grown under normal (see text) growth conditions for 14 h at 95° C. (a) IEFgel and autoradiogram for IPG pH 3–11. (b) IEFgel and autoradiogram for IPG pH 4–7.

Fig. 2 shows the autoradiogram of a native–native 2D-PAGE separation of P. furiosus soluble-protein extract obtained under two different growth conditions. Each of the separated spots was excised from the native 2D-PAGE, and proteins were identified after trypsin digestion and tandem MS analysis. The MS/MS results show that there can be numerous overlapping protein bands. Using the P. furiosusgenome annotation, the identities that are relevant to the tungsten metallome were selected (the detailed results of mass spectrometry are presented in ESI Tables S1 and S2† ). The results of the protein identification in the separated spots for the two growth conditions are presented in Tables 1 and 2 (Fig. 2 shows the spot positions on the gels ). From the MS data it can be concluded that all five known tungsten enzymes of P. furiosus were found under both growth conditions, as presented in Fig. 2. The amount of metal radioisotope was quantified in each of the separated spots and the data are presented in Table 3. The cause for the occurrence of multiple spots of certain proteins in the native gels is not known; partial inactivation of these oxygen-sensitive enzymes is a plausible factor of relevance.

Fig. 2 Native–native 2D-PAGE of soluble protein extract of P. furiosus. (a) Soluble protein extract (515 μg protein) obtained from normal growth conditions for 14 h at 95 °C. (b) Soluble protein extract (280 μg protein) obtained from cold-shock growth conditions for 14 h at 95 °C plus 5 h at 72 °C.

Table 1 Protein identification using mass spectrometry for native–native 2D-PAGE separation of P. furiosus soluble protein extract, growth 14 h at 95 °C

Spot	Accession	Protein	Mass	Number of unique peptides	Mascot^aprotein score
a Mascot identity threshold p < 0.05. The positions of numbered spots are indicated in Fig. 3A.
1	PF1203	FOR	69 072	7	269
2	PF1480	WOR5	65 025	19	740
3	PF0346	AOR	66 931	9	228
4	PF0346	AOR	66 931	12	605
5	PF0346	AOR	66 931	9	294
6	PF1961	WOR4	69 610	9	251
	PF0464	GAPOR	74 089	2	65
7	PF1961	WOR4	69 610	2	76
8	PF0464	GAPOR	74 089	2	60
	PF1961	WOR4	69 610	2	43
9	PF0464	GAPOR	74 089	14	563
	PF1203	FOR	69 072	5	199
10	—	—	—	0	—

---

Table 2 Protein identification with mass spectrometry after native–native 2D-PAGE separation of P. furiosus soluble protein extract obtained from cold-shock growth conditions. Cells were grown for 5 h at 72 °C in addition to 14 h at 95 °C

Spot number	Accession	Protein	Mass	Number of unique peptides	Mascot^aprotein score
a Mascot identity threshold p < 0.05. The positions of numbered spots are indicated in Fig. 3B.
1	PF1203	FOR	69 072	17	791
2	PF1480	WOR5	65 025	25	1330
	PF1203	FOR	69 072	3	90
3	PF0346	AOR	66 931	33	2184
	PF1480	WOR5	65 025	3	207
4	PF0346	AOR	66 931	10	476
5	PF0346	AOR	66 931	9	405
6	PF1961	WOR4	69 610	27	2332
	PF0464	GAPOR	74 089	14	792
	PF1203	FOR	69 072	15	498
7	PF1961	WOR4	69 610	28	1,803
	PF0464	GAPOR	74 089	10	424
8	PF1961	WOR4	69 610	32	1616
9	PF0464	GAPOR	74 089	12	530
	PF1961	WOR4	69 610	3	207
10	PF0464	GAPOR	74 089	2	83
11	PF0464	GAPOR	74 089	15	479
12	PF1480	WOR5	65 025	15	637
13	—	—	—	0	—
14	—	—	—	0	—

---

Table 3 Metal quantification from a native–native 2D of soluble protein extract (280 μg protein)

Enzyme	Quantification of tungsten based on ^187W/pmol^a
	Normal growth	Cold-shock
a The error value was defined as the standard deviation of the quantification from the spots of two independent gels for each condition.
FOR	22.7 ± 0.8	22.6 ± 1.4
AOR	16.5 ± 1.6	12.7 ± 0.2
GAPOR	2.5 ± 0.15	3.9 ± 0.1
WOR4	3.5 ± 0.15	7.3 ± 0.9
WOR5	3.2 ± 0.01	5.5 ± 0.2

---

The autoradiograms of the native–native 2D-PAGE of the two growth conditions employed here, showed the presence of all tungsten enzymes in a reproducible pattern of spots as presented in Fig. 2. Only some features in the acidic part of the native–native 2D-PAGE could not be attributed to a tungsten enzyme based on MS/MSprotein identification. A control experiment was therefore performed by running a native–native 2D-PAGE using the ¹⁸⁷W solution without any protein extract and indeed the same features were found at a pH around 4.5 (see ESI Fig. S3† ). Apparently, these features are due to inorganic tungsten species formed during the preparation of the ¹⁸⁷W solution. For example, ammonium paratungstate is known to form in concentrated ammonium tungstate solutions subject to evaporation, and it is slightly soluble in water at high temperature (i.e. 100 °C).¹⁹

The membrane tungstoproteome of P. furiosus is empty

After cell disruption, removal of the soluble protein extract and cell debris, not more than 40 Bq ¹⁸⁷W was left in the membrane pellet after two washing steps. This amount of radioactivity corresponds to, less than 1 fmol tungsten (135 μg protein). This means that the amount of tungsten associated with the membrane proteins is at least three orders of magnitude lower than that associated with the soluble proteins. The membrane proteins from P. furiosus were separated using native–native 2D-PAGE. As expected, no spots containing ¹⁸⁷W were detected on the native–native 2D-PAGEgel in three independent experiments (see ESI Fig. S2† ).

Differential expression of tungsten enzymes under different growth conditions

3D images of the native–native 2D-PAGE for two growth conditions are presented in Fig. 3. By integrating the intensity over the area of each individual spot, the amount of ¹⁸⁷W in that spot was quantified under normal growth and cold-shock conditions as presented in Table 3. From these data we conclude that FOR and GAPOR have approximately invariant expression levels in normal growth versus cold-shock conditions, while AOR is down-regulated 1.3 fold under cold-shock conditions and both WOR4 and WOR5 are up-regulated approximately 2 fold. The experiments for both growth conditions were carried out in duplicate. The obtained data show a strong correlation with the results of previous studies using DNA microarray analysis on cold-shock versus normal growth (see Table 4).¹¹ It has been reported that for prolonged cold-shock (5 h) the mRNA levels of WOR5 and WOR4 are up-regulated 4.2 fold compared to normal growth conditions, while AOR is down-regulated 4.2 fold.

Fig. 3 3D images of native–native 2D-PAGE of soluble protein extract (280 μg protein) of P. furiosus. (a) Soluble protein extract obtained from normal growth conditions. (b) Soluble protein extract obtained from cold-shock growth conditions.

Table 4 P. furiosus tungsten enzyme expression. Comparison of the tungsten enzyme relative expression levels (cold-shock versus normal growth) determined with native–native 2D-PAGE and previously published DNA microarray experiments

Enzyme	Tungsten enzyme relative expression levels (cold-shock versus normal)
	mRNA ^a	Tungsten protein
a Data previously published.^11
FOR	1.1	1
AOR	0.2	0.8
GAPOR	0.8	1.1
WOR4	4.2	2.0
WOR5	4.2	2.2

---

Identification of PF1479 as a putative binding partner of WOR5

Blue-native PAGE is known to keep protein complexes intact during the electrophoresis. Analysis of the WOR5-containing spot by MS/MS revealed the co-localization of PF1479 in four independent gels (see ESI Tables S1 and S2† ). PF1479 is next to WOR5 (PF1480) on the genome of P. furiosus and the two genes have been identified to form an operon together.²⁰ PF1479 is predicted to be a 20 kDa protein with cysteine-motifs for the coordination of four [4Fe-4S] clusters based on a significant sequence homology with the β-subunit of the structurally characterized E. coliformate dehydrogenase (FDH-N). The co-localization of PF1479 with WOR5 on the native–native PAGEgels strongly suggests that these proteins form a complex and that WOR5 is a heteromeric enzyme with a catalytic α-subunit (PF1480) and an electron-transfer β-subunit (PF1479). This is a clear difference between WOR5 and the other tungsten enzymes (AOR, FOR, and GAPOR), which use a ferredoxin (PF1909) as a redox partner protein and presumably only exhibit transient binding.

The data presented in ESI Tables S1 and S2† reveal several additional putative protein complexes: PF1837–PF1838 (hypothetical proteins), PF0182–PF0183 (V-type ATP synthase subunit a and b) and PF1076–PF1077 (Gyrase modulator). Of these only PF0182–PF0183 has not been previously recognized to be part of an operon.

Discussion

The P. furiosus cell contains a substantial amount of ‘free tungsten’, presumably in the form of tungstate and/or polytungstates. This is different from the metallome of other metal ions, such as copper and zinc with, for example, in E. coli a cytoplasmic concentration of the free metal ion in the fmol range.²¹ This observation indicates that these high levels of tungstate are apparently not toxic to P. furiosus. We did not find any evidence for a tungstate storage protein in P. furiosus, and the high level of free tungstate makes it unlikely that such a system would be necessary.

It has taken two decades of classical biochemical research to identify and isolate all five soluble tungsten enzymes from P. furiosus.^5–9 Here, we have demonstrated that all five of these tungsten enzymes can be identified and quantified in a single experiment using no more than 200 μg soluble protein extract. In addition to the five soluble tungsten enzymes already known, namely; AOR, FOR, GAPOR, WOR4 and WOR5, no other protein that contains tungsten was found.

From microarray data, it is known that the five tungsten enzymes are all expressed at finite levels under the different growth conditions. Previously, however, only three of these enzymes could be identified and isolated after growth on polycarbohydrate under normal conditions (95 °C batchwise for 14 h). Here, the presence of the other two, i.e. WOR4 and WOR5 was shown under the same conditions, due to the superior detection limit of the method used.

The levels of the five tungsten enzymes based on ¹⁸⁷W content are compared to the protein levels based on the original enzyme purifications in Table 5. The quantification based on the radioactivity is in good agreement with the amount of specific tungsten enzyme that can be obtained after proteinpurification. The published values of the purification factor for the tungsten enzymes, however, give an overestimation of the tungsten enzyme content. This is most likely due to inactivation of the enzymes by oxygen, and to the overlapping substrate specificities of AOR, FOR and WOR5. The differential expression levels of the five tungsten enzymes under normal growth and cold-shock show a strong correlation with previously published DNA microarray data.¹¹

Table 5 Comparison of the absolute expression levels from a native–native 2D gel with the expression levels based on the original enzyme purification tables

Enzyme	Purification fold^a	Quantification based on
		Isolated protein^a	^187W 2D–PAGE^b
a Ref. 5–9. b This work; n.d. (not determined).
FOR	6%	0.4%	0.56%
AOR	4.2%	0.3%	0.39%
GAPOR	3.3%	n.d.	0.09%
WOR4	n.d.	0.06%	0.09%
WOR5	6.8%	0.15%	0.06%

---

Genomic analysis indicated that, besides the tungsten-containing enzymes of the AOR family, there are two genes with significant sequence homology to genes of tungsten- or molybdenum-containing formate dehydrogenases (FDH).²² Despite the fact that transcription of the two putative FDH encoding genes has been shown in DNA microarray experiments, no evidence for the expression as a soluble or membrane-bound tungsten-containing protein has been found here. Possibly, the expression of these FDHs is too low under the growth conditions used here, or they contain molybdenum instead of tungsten.

P. furiosus has an active transport system for tungstate: WtpABC.²³ WtpA has been shown to bind tungstate with high affinity. This protein delivers the tungstate to the transmembrane part of the transport system. It has a significant off-rate in vitro and loses all its tungstate within a few hours in an environment without tungstate. Therefore, it is probably not possible to visualize this protein on its ¹⁸⁷W content using the native–native 2D-PAGE procedure, even if it would be expressed under the cultivation condition used (notably: non-limiting tungstate concentrations).

In conclusion, the tungsten metallome of the tungsten-dependent hyperthermophilic archaeon P. furiosus, grown on polycarbohydrate, consists of approximately 30% intracellular free tungstate (and/or polytungstate), a small fraction of which may be transiently bound to the tungstate transport system Wtp and to the machinery for tungstopterin cofactor biosynthesis and insertion.²⁴ The remaining 70% is present as tungstopterin cofactor in the five soluble aldehyde oxidoreductase enzymes FOR (33%), AOR (23%), GAPOR (6%), WOR4 (5%), WOR5 (4%). The expression levels of the tungsten enzymes as % of total cellular protein are: FOR (0.56%), AOR (0.39%), GAPOR (0.09%), WOR4 (0.09%), WOR5 (0.06%). No other soluble or membrane-bound tungstoproteins are present above the detection limit (1 fmol) of the method of native–native gel electrophoresis in combination with the radioisotope ¹⁸⁷W.

Acknowledgements

We thank J. J. Kroon (Department of Radiation, Radionuclides and Reactors) for help with radioisotope ¹⁸⁷W preparation. We thank R. de Kat (Department of Aerodynamics) for help with visualizing data using Tecplot 360 2008. This research was supported by grant NWO-CW 700.55.004 from the Council for Chemical Sciences of The Netherlands Organization for Scientific Research. PV and MP acknowledge financial support from The Netherlands Proteomics Centre and the Kluyver Centre for Genomics of Industrial Fermentation.

References

1. R. N. Schicho, L. J. Snowden, S. Mukund, J. B. Park, M. W. W. Adams and R. M. Kelly, Arch. Microbiol., 1993, 145, 380–385.
2. G. Falia and K. O. Stetter, Arch. Microbiol., 1986, 145, 56–61.
3. P. J. Silva, E. C. D. v. d. Ban, H. Wassink, H. Haaker, B. d. Castro, F. T. Robb and W. R. Hagen, Eur. J. Biochem., 2000, 267, 6541–6551.
4. R. Sapra, K. Bagramyan and M. W. W. Adams, Proc. Natl. Acad. Sci. U. S. A., 2003, 100, 75457550.
5. R. Roy, S. Mukund, G. J. Shut, D. M. Dunn, R. Weiss and M. W. W. Adams, J. Bacteriol., 1999, 181, 1171–1180.
6. S. Mukund and M. W. W. Adams, J. Biol. Chem., 1991, 266, 14208–14216.
7. S. Mukund and M. W. W. Adams, J. Biol. Chem., 1995, 270, 8389–8392.
8. R. Roy and M. W. W. Adams, J. Bacteriol., 2002, 184, 6952–6956.
9. L. E. Bevers, E. Bol, P. L. Hagedoorn and W. R. Hagen, J. Bacteriol., 2005, 187, 7056–7061.
10. G. J. Schut, S. D. Brehm, S. Datta and M. W. W. Adams, J. Bacteriol., 2003, 185, 3935–3947.
11. M. V. Weinberg, G. J. Schut, S. D. Brehm, S. Datta and M. W. W. Adams, J. Bacteriol., 2005, 187, 336–348.
12. A. M. Sevcenco, G. C. Krijger, M. W. Pinkse, P. D. E. M. Verhaert, W. R. Hagen and P. L. Hagedoorn, J. Biol. Inorg. Chem., 2009, 14, 631–640.
13. M. Wilm, A. Schevchenko, T. Houthaeve, S. Breit, L. Schweigerer, T. Fotsis and M. Mann, Nature, 1996, 379, 466–469.
14. J. L. Corbin and G. D. Watt, Anal. Biochem., 1990, 186, 30–40.
15. P. L. Hagedoorn, P. Van 't Slot, H. P. Van Leeuwen and W. R. Hagen, Anal. Biochem., 2001, 297, 71–78.
16. L. O. Martins and H. Santos, Appl. Environ. Microbiol., 1995, 61, 3299–3303.
17. R. van der Linden, F. de Corte and J. Hoste, J. Radioanal. Chem., 1974, 20, 695–706.
18. E. Bol, L. E. Bevers, P. L. Hagedoorn and W. R. Hagen, J. Biol. Inorg. Chem., 2006, 11, 999–1006.
19. L. Yunjiao, S. Peimei, L. Honggui, S. Pengtuan, Z. Zhongwei and L. Maosheng, J. Cent. South Univ. Technol., 1997, 4, 32–35.
20. T. T. Tran, P. Dam, Z. Su, F. L. Poole, M. W. W. Adams, G. T. Zhou and Y. Xu, Nucleic Acids Res., 2007, 35, 69–78.
21. L. A. Finney and T. V. O'Halloran, Science, 2003, 300, 931–936.
22. F. T. Robb, D. L. Maeder, J. R. Brown, J. DiRuggiero, M. D. Stump, R. K. Yeh, R. B. Weiss and D. M. Dunn, Methods Enzymol., 2001, 330, 134–157.
23. L. E. Bevers, P. L. Hagedoorn, G. C. Krijger and W. R. Hagen, J. Bacteriol., 2006, 188, 6498–6505.
24. L. E. Bevers, P. L. Hagedoorn and W. R. Hagen, Coord. Chem. Rev., 2009, 253, 269–290.

---

Footnote

† Electronic supplementary information (ESI) available: ¹⁸⁷W calibration curve, native–native 2D-PAGE of P. furiosus membrane proteins, native–native 2D-PAGE of ¹⁸⁷W solution and protein identification tables. See DOI: 10.1039/b908175e

---