Báo cáo khoa học: Geranylgeranyl reductase involved in the biosynthesis of archaeal membrane lipids in the hyperthermophilic archaeon Archaeoglobus fulgidus potx

Hemmi, Department of Applied Molecular Bioscience, Graduate School of Bioagricultural Sciences, Nagoya University, Furo-cho, Chikusa-ku, Nagoya, Aichi 464–8601, Japan Fax: +81 52 7894120

of archaeal membrane lipids in the hyperthermophilic

archaeon Archaeoglobus fulgidus

Motomichi Murakami1, Kyohei Shibuya2, Toru Nakayama2, Tokuzo Nishino2, Tohru Yoshimura1 and Hisashi Hemmi1

1 Department of Applied Molecular Bioscience, Graduate School of Bioagricultural Sciences, Nagoya University, Aichi, Japan

2 Department of Biomolecular Engineering, Graduate School of Engineering, Tohoku University, Miyagi, Japan

The structure of membrane lipids is the most striking

characteristic of the Archaea (one of the three domains

of life), which includes many extremophiles, such as

thermophiles, halophiles and methanogens [1–3] The

archaeal membrane lipids are different from the

typ-ical glycerolipids in organisms of the other domains –

Bacteria and Eucarya – in the following respects

First, Archaeal lipids have fully reduced prenyl chains, whereas glycerolipids typically have fatty acyl chains Almost all archaea produce membrane lipids that contain phytanyl groups (i.e fully saturated C20prenyl groups) Second, the connection of the hydrocarbon chains with the glycerol moiety occurs via an ether bond in archeal lipids, not via the ester bond generally

Keywords

archaea; geranylgeranyl reductase;

isoprenoid; lipid; oxidoreductase

Correspondence

H Hemmi, Department of Applied

Molecular Bioscience, Graduate School of

Bioagricultural Sciences, Nagoya University,

Furo-cho, Chikusa-ku, Nagoya, Aichi

464–8601, Japan

Fax: +81 52 7894120

Tel: +81 52 7894134

E-mail: hhemmi@agr.nagoya-u.ac.jp

(Received 27 September 2006, revised

29 November 2006, accepted 5 December

2006)

doi:10.1111/j.1742-4658.2006.05625.x

Complete saturation of the geranylgeranyl groups of biosynthetic interme-diates of archaeal membrane lipids is an important reaction that confers chemical stability on the lipids of archaea, which generally inhabit extreme conditions An enzyme encoded by the AF0464 gene of a hyperthermophi-lic archaeon, Archaeoglobus fulgidus, which is a distant homologue of plant geranylgeranyl reductases and an A fulgidus menaquinone-specific prenyl reductase [Hemmi H, Yoshihiro T, Shibuya K, Nakayama T, & Nishino T (2005) J Bacteriol 187, 1937–1944], was recombinantly expressed and puri-fied, and its geranylgeranyl reductase activity was examined The radio HPLC analysis indicated that the flavoenzyme, which binds FAD noncova-lently, showed activity towards lipid-biosynthetic intermediates containing one or two geranylgeranyl groups under anaerobic conditions It showed a preference for 2,3-di-O-geranylgeranylglyceryl phosphate over 3-O-geranyl-geranylglyceryl phosphate and geranylgeranyl diphosphate in vitro, and did not reduce the prenyl group of respiratory quinones in Escherichia coli cells The substrate specificity strongly suggests that the enzyme is involved

in the biosynthesis of archaeal membrane lipids GC-MS analysis of the reaction product from 2,3-di-O-geranylgeranylglyceryl phosphate proved that the substrate was converted to archaetidic acid (2,3-di-O-phytanyl-glyceryl phosphate) The archaeal enzyme required sodium dithionite as the electron donor for activity in vitro, similarly to the menaquinone-specific prenyl reductase from the same anaerobic archaeon On the other hand, in the presence of NADPH (the preferred electron donor for plant homo-logues), the enzyme reaction did not proceed

Abbreviations

DGGGP, 2,3-di-O-geranylgeranylglyceryl phosphate; DGGGPS, 2,3-di-O-geranylgeranylglyceryl phosphate synthase; GGGP,

3-O-geranylgeranylglyceryl phosphate; GGGPS, 3-O-3-O-geranylgeranylglyceryl phosphate synthase; GGR, geranylgeranyl reductase; GGPP,

geranylgeranyl diphosphate; GGPS, geranylgeranyl diphosphate synthase; IPP, isopentenyl diphosphate; PR, prenyl reductase.

formed in lipids Third, the stereochemistry of the

gly-cerol moiety is enantiomeric between the archaeal and

the typical lipids Fourth, most thermophilic and

meth-anogenic archaea also contain bipolar cyclic lipids

(‘tetraether’ lipids), which are probably formed by the

dimerization of two ‘diether’ lipids The biosynthesis

of the archaeal membrane lipid has been studied

previ-ously (Fig 1) The precursor of the glycerol moiety,

sn-glycerol-1-phosphate, is formed from dihydroxy

acetone phosphate [4,5] On the other hand, the

pre-cursor of the prenyl moiety is synthesized from active

C5 isoprene units, for example, isopentenyl

diphos-phate (IPP) and dimethylallyl diphosdiphos-phate, usually by

geranylgeranyl diphosphate (GGPP) synthase, which

yields the C20 precursor [6,7], although a few archaea

are known to utilize geranylfarnesyl diphosphate

syn-thase to synthesize the C25 precursor [8,9] The

gera-nylgeranyl chains thus produced are then transferred

to the sn-3 position of sn-glyceryl-1-phosphate by

3-O-geranylgeranylglyceryl phosphate (GGGP)

syn-thase [10–12] and subsequently to the sn-2 position by 2,3-di-O-gerenylgeranylglyceryl phosphate (DGGGP) synthase [13] The fundamental carbon-oxygen skeleton

of archaeal membrane lipids is formed at this point, fol-lowed by various processes, such as modification of polar head groups [14], saturation or cyclization of pre-nyl chains, and the creation of a bipolar cyclic structure [15] However, although a few enzymes that catalyze polar head modification (i.e CTP:DGGGP cytidyl-transferase [16] and archaetidylserine synthase [17]) have been found, enzymes catalyzing the other process have not been examined in detail

The complete saturation of prenyl chains would con-fer chemical stability on archaeal membrane lipids Therefore, the reduction of prenyl chains is generally thought to play an important role in the survival of archaea under extreme conditions, such as high tem-perature or salinity, although partially saturated prenyl chains have been found in some archaea and, interest-ingly, the number of unsaturated double bonds is known to be related to the temperature at which the organism grows [18,19] Saturated prenyl groups are also found in compounds, other than membrane lipids,

in archaea [2,20] Many archaea produce respiratory quinones (i.e menaquinone, caldariellaquinone, sulfo-lobusquinone, thermoplasmaquinone, etc.) that contain fully or partially saturated prenyl side-chains We recently identified an enzyme that catalyzes the satura-tion of the prenyl side-chain of menaquinone in a hyperthermophilic archaeon, Archaeoglobus fulgidus [21] The enzyme, prenyl reductase (PR), is a distant homologue of geranylgeranyl reductase (GGR) from plants [22,23] and cyanobacteia [24], which catalyzes the saturation of the geranylgeranyl group to produce chlorophyll, tocopherol and probably phyloquinone The menaquinone-specific PR is a FAD-dependent flavoenzyme Sodium dithionite, as an electron donor, is required for the reducing reaction to proceed

in vitro, but NADPH does not function, although plant homologues are dependent on this reducing agent In addition, the enzyme does not require a diva-lent metal ion for reaction We isolated three other genes of GGR homologues from A fulgidus, none of which have known functions

In this article, we report on the function of an archaeal GGR homologue encoded in the ORF AF0464 (chlP-1), which is efficiently expressed in Escherichia coli, as evidenced by our previous report [21] The recombinant enzyme was affinity purified, and its activity was assessed in vitro under anaerobic conditions The enzyme, whose properties are very similar to those of the menaquinone-specific PR, cata-lyzes the conversion of geranylgeranyl groups of

Fig 1 Biosynthetic pathway of archaeal membrane lipids X

denotes a polar head group When X is phosphate, phosphoserine

or phosphoethanolamine, the archaeal lipid with two phytanyl

chains is denoted as archaetidic acid, archaetidylserine or

archaeti-dylethanolamine, respectively DGGGP,

2,3-di-O-geranylgeranyl-glyceryl phosphate; DGGGPS, DGGGP synthase; GGGP,

3-O-geranylgeranylglyceryl phosphate; GGGPS, GGGP synthase;

GGPP, geranylgeranyl diphosphate; GGPS, GGPP synthase; IPP,

isopentenyl diphosphate Enzymes are indicated in boxes.

DGGGP into phytanyl groups It also acts, although

weakly, on the geranylgeranyl group of GGGP and

GGPP These facts strongly suggest the involvement of

the enzyme, A fulgidus GGR, in the biosynthesis of

membrane lipids

Results

Recombinant expression and purification

of an archaeal GGR homologue

In our previous study, we reported on the recombinant

expression of archaeal GGR homologues encoded in

ORFs, namely AF0464 (chlP-1), AF1023 (chlP-2),

AF1637 (chlP-3) and AF0648 [21] The expression of

AF0648 resulted in a change in the quinone profile of

the host E coli, which led us to conclude that the

menaquinone-specific PR is encoded in AF0648

How-ever, the functions of the other three GGR

homo-logues were not clear at that time Among them, only

the recombinant expression product of AF0464 could

be obtained from the soluble fraction after

centrifuga-tion of the cell lysate, whereas the other two were

found in the precipitate Therefore, in this study, we

attempted to purify the recombinant protein encoded

in AF0464 in order to understand its function in

greater detail To express the archaeal protein as a

fusion with a polyhistidine-tag at its N terminus, a

gene fragment, containing AF0464, was cut from the

pET3a–AF0464 vector, constructed previously, and

inserted into the pET15b vector E coli cells were

transformed with the resultant plasmids and then

cul-tured with the appropriate induction The expressed

protein was purified using a Ni-chelating affinity

col-umn chromatography after heat treatment The purity

of the protein was verified by SDS⁄ PAGE As shown

in Fig 2, a strong protein band was observed in the

crude extract, as well as in the heat-treated enzyme

solution However, the molecular weight of the protein

estimated from SDS⁄ PAGE data seems to be slightly

smaller than the calculated value, 46978.37 The

pro-tein did not specifically bind to a Ni-chelating affinity

column and was recovered in the flow-through

frac-tion Edman degradation of the protein gave us the

N-terminal amino acid sequence, MYDVVVGA,

which clearly showed that the protein arose from the

translation from Met at the 24th position of the

expec-ted full-length of the archaeal enzyme (except for the

polyhistidine-tag) In contrast, a distinct protein of

slightly lower mobility, which corresponds reasonably

well with the calculated molecular weight, was purified

using the affinity column We used the later,

affinity-purified, protein for further characterization because

we needed purer protein solution The UV-visible spec-trum of the concentrated, purified enzyme solution is shown in Fig 3 In this spectrum, specific peaks for flavin coenzymes were observed at  380 and 440 nm, like the spectrometric analysis of recombinant A fulgi-dusPR Thus, we attempted to extract the flavin cofac-tor from the protein by heating in methanol The extracted compound, which had a yellow color and emits fluorescence under UV light, comigrated with FAD, but not with FMN, on chromatography paper (data not shown) This fact proved that the archaeal protein, at least when recombinantly expressed in

E coli, contains noncovalently bound FAD By refer-ring to the absorption coefficient for free FAD, 87%

of the purified enzyme was estimated to bind FAD

Fig 2 SDS ⁄ PAGE of the recombinant geranylgeranyl reductase (GGR) homologue from Archaeoglobus fulgidus Lane 1, standard molecular marker; lane 2, crude extract from BL21(DE3) ⁄ pET15b-AF0464; lane 3, supernatant fraction after heat treatment; lane 4, recombinant GGR homologue purified using Ni-chelating affinity col-umn chromatography The strongly expressed protein bands in lanes 2 and 3, indicated by an asterisk, are shown to arise from the archaeal protein expressed in a truncated form.

Fig 3 UV-visible spectrum of a geranylgeranyl reductase (GGR) homologue from Archaeoglobus fulgidus.

GGR assay using radio HPLC

We performed enzyme assays under anaerobic

condi-tions because A fulgidus is an obligate anaerobe and

because the menaquinone-specific PR from the archaeon

requires anaerobic conditions [21] Oxygen was

removed from the reaction mixture by bubbling with

N2 gas, and sodium dithionite was added to eliminate

oxygen completely in the reaction mixture and also to

donate electrons for reducing reactions The

radio-labeled products synthesized with the three

recombin-ant prenyltransferases [i.e GGPP synthase (GGPS),

GGGP synthase (GGGPS), and DGGGP synthase

(DGGGPS)], which mainly contained DGGGP, were

first used as substrates for the reductase assay After the reaction, compounds that contained a phosphate

or diphosphate group were hydrolyzed with acid phos-phatase and then extracted with n-pentane to be ana-lyzed by radio HPLC The elution profiles of the pentane extracts showed the appearance of a new peak, with an elution time longer than that of the hydrolyzed product from DGGGP (Fig 4A) When 0.1% Triton X-100 was added, several new peaks appeared between the new peak and that of dephos-phorylated DGGGP To confirm this finding, various concentrations of Triton X-100 were added to the reaction mixture, in which radioactive substrates, syn-thesized from fourfold greater quantities of [1-14C]IPP

D C

Fig 4 Radio HPLC analysis of the products from the geranylgeranyl reductase (GGR) assay (A) Elution profiles of the radiolabeled com-pounds extracted from the GGR assay mixture, which mainly contained radiolabeled 2,3-di-O-geranylgeranylglyceryl phosphate (DGGGP) as the substrate The compounds were dephosphorylated with acid phosphatase prior to HPLC analysis Digeranylgeranylglycerol arose from the dephosphorylation of unreacted DGGGP eluted at  18 min An asterisk indicates peaks eluted at  10 min that are probably derived from dephosphorylated alcohols from the intermediates of substrate production (i.e geranylgeraniol and ⁄ or geranylgeranylglycerol) (B) Elu-tion profiles of the radioactive compounds from the GGR assay mixture to which various concentraElu-tions of Triton X-100 were added For the synthesis of the substrates for these assays, a fourfold greater amount of [1- 14 C]IPP was used An asterisk indicates peaks derived from radioactive compounds brought into the GGR reactions other than DGGGP Peaks at  10 min are probably derived from geranylgeraniol and ⁄ or geranylgeranylglycerol, whereas the compounds corresponding to peaks at  7 min are unidentified (C and D) The radio HPLC profiles of the compounds from the GGR assay mixture, in which radiolabeled 3-O-geranylgeranylglyceryl phosphate (GGGP) (C) and geranyl-geranyl diphosphate (GGPP) (D) were the main substrates.

and (all-E) farnesyl diphosphate, were used (Fig 4B).

The elution profiles clearly showed that higher

deter-gent concentrations increased the intensity of the peaks

with shorter elution times These new peaks were very

similar to those observed in the reduction of

menaqui-none by A fulgidus PR, which arise from

menaquin-ones with a partially saturated prenyl group The

compounds corresponding to the new peaks appeared

to be produced from DGGGP because they eluted

suc-cessively after the peak of DGGGP and because the

decline in the DGGGP peak corresponded to the

appearance of the new peaks Furthermore, a total of

eight new peaks, derived from DGGGP, appeared,

strongly suggesting that the peaks correspond with

reaction products that have different numbers of

dou-ble bonds remaining unsaturated If so, the peak with

the longest elution time, observed in the absence of

detergent, would be expected to arise from archaetidic

acid, the final product with two phytanyl chains The

addition of 2 mm NADPH, instead of sodium

dithio-nite, and also the removal of sodium dithiodithio-nite, failed

to produce such new peaks, suggesting that NADPH

does not act as a specific electron donor for the

enzyme This hypothesis was also supported by the

facts that the addition of NADPH did not diminish

the absorption peak of the enzyme at 440 nm, which

is derived from the oxidized-form of FAD, even under

anaerobic conditions, and that the enzyme did not

reduce NADPH at 55C in the presence of oxygen

(data not shown) The supplemental addition of

0.5 mm FAD did not significantly enhance the

reac-tion, probably because the enzyme is already saturated

with FAD, as described above The addition of 10 mm

EDTA to the reaction mixture did not inhibit the

reac-tion, indicating that the enzyme does not require a

divalent metal ion, as observed in the reaction of the

menaquinone-specific PR from A fulgidus

We next carried out GGR assays, using other

radio-active substrates, to determine the substrate specificity

of the enzyme The reaction products of GGPS, and of

both GGPS and GGGPS, which mainly contained

GGPP and GGGP, respectively, were used as the

sub-strates As shown in Fig 4C,D, new peaks with longer

elution times were also observed in the elution profiles

of the reaction with both substrates These

observa-tions suggested that the enzyme is able, at least

parti-ally, to reduce both of the substrates, which contain a

geranylgeranyl group However, the enzyme activity

for these substrates seemed not to be as high as that

for DGGGP because such new product peaks, arising

from GGPP or GGGP, were not as obvious as those

from DGGGP when the reaction mixture contained

both DGGGP and the other substrates (Fig 4A,B)

The product specificity of the enzyme strongly suggests that the enzyme preferentially catalyzes reducing reac-tions to produce archaetidic acid from DGGGP

Product analysis by GC-MS The butanol-extracted products from the GGR assay,

in which a nonlabeled substrate, such as DGGGP, GGGP and GGPP, was used, were dephosphorylated, trimethylsilylated and subjected to GC-MS analysis When DGGGP was used as the substrate, a small peak, with the same retention time at 31.5 min as that of tri-methylsilylated archaeol, extracted from Halobacterium salinarum as an authentic sample, was observed on the chromatogram (Fig 5) Such a peak could not be found when the enzyme was not present in the reaction mixture Although only a slight ion peak identical to [M+H]+ was observed at m⁄ z 726, strong peaks

at m⁄ z 710, 621 and 426 were considered to corres-pond with [M-CH3]+, [M-CH3OSi(CH3)3]+ and [M-C20H41OH]+, respectively Moreover, the mass spectrum of the compound corresponding to the peak was almost identical to those of trimethylsilylated authentic archaeol we prepared and previously reported

by Teixidor & Grimalt [25], strongly suggesting that DGGGP was converted to archaetidic acid, a common component of the archaeal membrane DGGGP and partially saturated intermediate products were not detected by GC-MS, which can be explained by a scen-ario in which the production of such intermediates might be negligible because detergent was not added to the reaction mixture and that compounds with double bonds were comparatively labile and therefore decom-posed under the severe conditions used for the detection

of archaeol On the other hand, such new products were not detected when GGGP or GGPP was used in the assay (data not shown), probably because the sat-uration of the geranylgeranyl group did not proceed well

Discussion

In this article, we characterized the function of an archaeal homologue of plant GGR and report here that the enzyme is able to reduce geranylgeranyl groups of archaeal membrane lipid precursors The properties of the enzyme from A fulgidus are very similar to those of the recently reported menaquinone-specific PR from the same organism [21], which is also homologous to plant GGR, in the following respects First, these reductases, at least when recombinantly expressed in E coli, are flavoenzymes that non-covalently bind FAD Second, they require sodium

dithionite, not NADPH, the preferable reducing agent

for plant homologues, for in vitro activity Third, they

do not require a divalent metal ion A fulgidus GGR

prefers DGGGP as the substrate and can convert it to

a general component of the archaeal membrane

(arch-aetidic acid), which contains two phytanyl chains The

enzyme also accepts GGGP and GGPP as substrates,

but the activity for them seems to be much weaker

than that for DGGGP On the other hand, the

expres-sion of the enzyme in E coli was reported to have no

effect on the quinone profile of the host, clearly

indica-ting that menaquinone and ubiquinone are not

prefer-able substrates for the enzyme [21] These facts

strongly suggest the involvement of the enzyme, GGR,

in the biosynthesis of membrane lipids in the

hyper-thermophilic archaeon A fulgidus, as predicted in the

previous publication It should be noted here that the

reduction of lipid precursors was not catalyzed by

the menaquinone-specific PR from A fulgidus (data

not shown)

The archaeal enzyme is distinct from plant GGR

because the final product of the enzyme contains a

phytanyl group, whereas plant GGR cannot saturate

all the double bonds of a geranylgeranyl group and

finally yields a phytyl group, which retains a double bond at position 2 [22–24] This difference is very important because the double bond is responsible for the formation of an allylic carbocation during the pre-nyltransfer reaction, which means the phytanyl group cannot be transferred to acceptors by prenyltransf-erases, whereas the phytyl group can Therefore, the substrate specificity of A fulgidus GGR is reasonable:

if phytanyl diphosphate is produced as a result of the complete reduction of GGPP, it cannot be utilized for the biosynthesis of isoprenoid compounds, such as archeal membrane lipids and respiratory quinones On the other hand, phytyl diphosphate produced by plant GGR is actually used in the biosynthesis of chloro-phyll, tocopherol and phylloquinone

Morii et al reported on a CTP:DGGGP cytidyl-transferase from Methanothermobacter thermoautotro-phicus, which catalyzes the modification of the polar head group of archaeal phospholipid [16] The enzyme can accept DGGGP as a substrate, but cannot utilize archaetidic acid On the other hand, an archaetidyl-serine synthase from M thermoautotrophicus accepts both substrates with geranylgeranyl and phtanyl groups (i.e CDP-2,3-di-O-geranylgeranylglycerol and

Fig 5 GC-MS analysis of the product from the geranylgeranyl reductase (GGR) assay (A) Chromatogram of butanol-extracted compounds from the GGR assay, in which 2,3-di-O-geranylgeranylglyceryl phosphate (DGGGP) was used as the substrate Reac-tion products were dephosphorylated and then trimethylsilylated The peak with an arrowhead had the same retention time as that of an authentic sample,

trimethylsilylat-ed archaeol (B and C) Mass spectra of the peak in (A) and the authentic sample, respectively.

CDP-2,3-di-O-phytanylglycerol), respectively, for the

formation of phosphatidylserine [17] (Here, the term

‘phosphatidyl’ denotes phosphoglycerolipids in a broad

sense, including archaetidyl phospholipids and their

analogues with geranylgeranyl groups in this case.)

Thus, the authors conclude, based on the specificities

of the enzymes, that the saturation of the

geranylgera-nyl groups of archaeal phospholipids occurs at least

after the transfer of the cytidyl group A fulgidus was

reported to contain no detectable phosphatidylserine,

but does contain phosphatidylethanolamine [3] So, if

a similar situation exists in the cells of A fulgidus,

GGR from the archaeon should catalyze the saturation

of the geranylgeranyl groups of various phospholipid

precursors, such as

CDP-2,3-di-O-geranylgeranylglyc-erol, 2,3-di-O-geranylgeranylglyceryl phosphoserine

and 2,3-di-O-geranylgeranylglyceryl

phosphoethanol-amine, as well as DGGGP However, the A fulgidus

genome still encodes two more homologues, with

unknown functions, of GGR [21] If these homologues

also catalyze the reduction of a geranylgeranyl or

prenyl group, they and A fulgidus GGR might have

distinct substrate specificities and physiological roles

The activity of the enzymes on the unknown

precur-sors of bipolar cyclic lipids (‘tetraether’ lipids), which

A fulgidusalso produces, is particularly interesting

While we were writing this article, Nishimura &

Eguchi reported on the purification of GGR, which is

specific for DGGGP and some other precursors of

archaeal phospholipids, from a thermoacidophilic

archeaon, Thermoplasma acidophilum [26] They

deter-mined the partial amino acid sequence of the enzyme

and concluded that the enzyme is encoded in an ORF,

Ta0516m, which is homologous to AF0464 Their

results strongly support our findings, although they

did not confirm the enzyme activity of the gene

expres-sion product However, they purified the enzyme from

the membrane fraction of T acidophilum, which

indi-cates that T acidophilum GGR is tightly associated

with the membrane On the other hand, A fulgidus

GGR seemed to be soluble, at least when expressed in

E coli, because it could be purified from the

superna-tant fraction after heat treatment, which usually makes

recombinant membrane proteins precipitate with the

membrane fractions of E coli, even though it had not

been solublized with detergents This characteristic of

A fulgidus GGR is also similar to that of the

mena-quinone-specific PR In fact, these enzymes were

pre-dicted to be soluble by sosui, a program for

classification and secondary structure prediction of

membrane proteins [27] (data not shown)

Further-more, T acidophilum GGR can utilize NAD(P)H as

an electron donor, whereas the enzyme from A

fulgi-duscannot This fact strongly suggests that A fulgidus GGR accepts electrons from other specific reducing agents (e.g cofactor F420or redox proteins such as fer-redoxin), in the living cells

Experimental procedures

Materials (All-E)-farnesyl diphosphate was donated by K Ogura and

T Koyama (Tohoku University) Nonlabeled IPP was donated by C Ohto (Toyota Motor Co, Toyota, Japan) [1-14C]IPP was purchased from GE Healthcare (Piscataway,

NJ, USA) All other chemicals were of analytical grade

General procedures Restriction enzyme digestions, transformations and other standard molecular biology techniques were carried out as described by Sambrook et al [28]

Expression and purification of the recombinant enzyme

The NdeI–BamHI fragment, containing the ORF AF0464, was cut from the pET3a-derived expression vector, reported previously [21], and inserted into the pET15b vector (Novagen, Darmstadt, Germany) E coli BL21(DE3), transformed with the new vector, was cultured aerobically

in Luria–Bertani broth supplemented with 50 mgÆL)1 ampi-cillin When the attenuance (D), at 600 nm, of the culture reached 0.6, the transformant cells were induced by treat-ment with 1.0 mm isopropyl thio-b-d-galactoside After

18 h of additional culture, the cells were harvested and dis-rupted by sonication in HisTrap binding buffer, containing

20 mm potassium phosphate buffer, pH 7.6, 0.5 m NaCl and 100 mm imidazole The homogenate was centrifuged at

15 000 g for 15 min, and the supernatant was recovered as

a crude extract The crude extract was heated at 55C for

1 h, and the denatured proteins were removed by centrifu-gation at 15 000 g for 15 min The supernatant fraction was recovered as a heat-treated enzyme The heat-treated enzyme, after filtration through a 0.45 lm membrane, was loaded onto a HisTrap column (GE Healthcare), previously equilibrated with binding buffer The column was washed with binding buffer, and specifically bound proteins were then eluted with an elution buffer, containing 20 mm potas-sium phosphate buffer, pH 7.6, 0.5 m NaCl and 500 mm imidazole, and used for characterization as purified GGR The level of protein expression was determined by electro-phoresis on a 12% SDS polyacrylamide gel UV-visible analysis of the purified enzyme solution (containing 0.25 lgÆlL)1 of the enzyme) was conducted with a Shim-adzu UV-2450 spectrophotometer (Shimadzu, Kyoto,

Japan) The absorption coefficient of FAD used to

calcu-late the concentration was 11 300 at 450 nm [29] The

con-centration of the protein was quantified by the Bradford

method [30] Prediction of transmembrane regions was

per-formed using the sosui program (http://bpnuap.nagoya-u

ac.jp/sosui/)

Protein sequencing

We performed Edman degradation for sequencing

N-ter-minal amino acids of the protein After SDS⁄ PAGE,

pro-teins were transferred onto a poly(vinylidene difluoride)

membrane, and the transferred protein band was clipped

out and brought into Edman degradation with a

Procise-TM HT protein sequencing system (Applied Biosystems,

Framingham, MA, USA)

Flavin analysis

We concentrated 500 lL of the purified enzyme solution

into a volume of  100 lL using a Centricon YM-10 spin

filter (Millipore, Billerica, MD, USA), replacing the buffer

with water To the concentrated enzyme solution, 1 mL of

methanol was added The mixture was heated at 100C for

15 min, and then centrifuged at 20 000 g for 10 min The

recovered supernatant was evaporated to 10 lL and then

spotted onto ADVANTEC 51A chromatography paper

(ADVANTEC, Tokyo, Japan) and developed with

n-buta-nol⁄ methanol ⁄ 5% Na2HPO4 (60 : 15 : 30, v⁄ v ⁄ v)

Authen-tic FMN and FAD were chromatographed on the same

paper Spots corresponding to flavins were detected by UV

illumination

Preparation of hypothetical substrates for

A fulgidus GGR

Enzymatic synthesis of GGPP, GGGP and DGGGP was

performed as reported previously [13] The standard

reac-tion mixture contained, in a final volume of 100 lL,

430 pmol of [1-14C]IPP (2.1 GBqÆmmol)1), 1 nmol of (all-E)

farnesyl diphosphate, 0.2 lmol of a-glycerophosphate,

2 lmol of MgCl2, 2 lmol of sodium phosphate buffer,

pH 5.8, and suitable amounts of recombinant enzymes (i.e

Sulfolobus acidocaldarius GGPS, S solfataricus GGGPS

and S solfataricus DGGGPS) The mixture was incubated

at 55C for 1 h and then used directly in the GGR assay as

the substrate mixture that mainly contains DGGGP To

synthesize the substrate mixtures containing mainly GGPP

and GGGP, DGGGPS or both DGGGPS and GGGPS

were removed, respectively

For the analysis of the GGR reaction products by mass

spectrometry, nonlabeled compounds were enzymatically

synthesized and purified The standard reaction mixture

contained, in a final volume of 3 mL, 600 nmol of

nonlabe-led IPP, 600 nmol of (all-E) farnesyl diphosphate, 80 lmol

of a-glycerophosphate, 30 lmol of MgCl2, 300 lmol of 2-molpholinoethanesulfonic acid-NaOH buffer, pH 5.8, and suitable amounts of recombinant prenyltransferases (i.e GGPS, GGGPS and DGGGPS) The mixture was incuba-ted at 55C for 2 h and then extracted with 3 mL of 1-but-anol saturated with H2O After evaporation, the butanol layer was loaded onto a COSMOSIL 5C4-AR-300 reverse-phase column (4.6· 150 mm; Nacalai Tesque, Kyoto, Japan), interfaced with an HPLC system, to purify DGGGP To recover GGGP and GGPP, DGGGPS and both DGGGPS and GGGPS were removed from the mix-ture, respectively The compounds were eluted from the col-umn with eluent A (25 mm NH4HCO3) isocratically for the first 2.5 min, and then with a linear gradient from 100% eluent A to 100% eluent B (acetonitrile) through 15 min, and finally with eluent B for 12.5 min, at a flow rate of

1 mLÆmin)1 Elution of the products was detected by UV absorption at 210 nm

Radio HPLC assay of GGR All manipulations for the GGR assay were carried out in

an anaerobic chamber until the reaction was complete The standard reaction mixture contained (in a volume of

350 lL) the substrate mixture from the prenyltransferase reaction described above, 200 lmol of 3-molpholinopro-panesulfonic acid-NaOH buffer, pH 7.5, and an appropri-ate amount of purified GGR Various amounts of Triton X-100 were added, as required The solutions of all con-tents, except for the enzyme and detergent, were bubbled with N2 gas to remove oxygen To the mixture, 50 lmol

of sodium dithionite in 50 lL of N2-bubbled water was added The mixture was then incubated at 55C for 1 h, and the reaction was stopped by adding 200 lL of a cold, saturated NaCl solution The mixture was extracted with

600 lL of 1-butanol saturated with H2O, and the butanol-extracted compounds were hydrolyzed with potato acid phosphatase (Sigma-Aldrich, St Louis, MO, USA) by the method of Fujii et al [31] The resulting alcohols were extracted with n-pentane and analyzed by HPLC with

a YMC Pack ODS-A C18 reverse-phase column (4.6·

250 mm, 5 lm; YMC Co., Ltd, Kyoto, Japan) The alco-holic compounds were isocratically eluted from the column with methanol⁄ 2-propanol (7 : 3, v ⁄ v) at a flow rate of 0.5 mLÆmin)1 Elution of the products was detected by radioactivity measured using a ramona Star radio-HPLC analyzer (Raytest, Straubenhardt, Garmany) The flow rate of the scintillation cocktail was 0.5–2 mLÆmin)1

Extraction of archaeol from H salinalium

H salinarum was cultured in 1 L of culture medium, con-taining 5 g of casamino acids, 5 g of yeast extract, 3 g of

trisodium citrate, 20 g of MgSO4)7H20, 2 g of KCl and

200 g of NaCl, at 37C for 3 days, and then harvested by

centrifugation Lipid extraction was performed by the

method of Bligh & Dyer [32] The chloroform layer,

con-taining the total lipids, was concentrated by evaporation,

and acetone was added to > 20-fold excess The mixture

was stored at 4C overnight to precipitate the polar lipids

Excision of polar head groups was performed according to

the method of Demizu et al [33] The precipitated polar

lipids were subjected to acetolysis at 165C for 20 h in

5 mL of a mixture of acetic acid⁄ acetic anhydride (3 : 2,

v⁄ v) After evaporating the solution to dryness, the

acetyl-ated lipids were hydrolyzed, by acid methanolysis, at

100C for 10 h in 3.5 mL of 5% HCl in methanol solution

After evaporation, the lipids, which mainly contained

archaeol, were recovered by partitioning with

chloro-form⁄ methanol ⁄ water (10 : 10 : 9, v ⁄ v ⁄ v)

GC-MS analysis

The nonlabeled substrates, purified as described above,

were used for reaction with GGR The substrate solution in

a glass tube was concentrated by evaporation and then

placed in an anaerobic chamber All manipulations

des-cribed below were carried out under anaerobic conditions

until the reaction was complete In the tube of the

sub-strate, the standard reaction mixture, in a volume of

2.7 mL, containing 1.5 mmol of

3-molpholinopropanesulf-onic acid-NaOH buffer, pH 7.5, and an appropriate

amount of purified GGR, was added Various amounts of

Triton X-100 were added, as required All solutions, except

for the enzyme and detergent, were bubbled with N2gas to

remove oxygen To the mixture, 300 lmol of sodium

dithi-onite, dissolved in 300 lL of N2-bubbled water, was added

The mixture was then incubated at 55C for 2 h, and the

reaction mixture was extracted with 3 mL of 1-butanol

sat-urated with H2O The compounds in the butanol layer were

enzymatically dephosphorylated by the method of Fujii

et al [31] The resulting alcohols were extracted with

n-pen-tane, and the pentane layer was completely evaporated The

residual lipids, or the authentic archaeol, were dissolved

with 90 lL of anhydrous pyridine After mixing the

pyrid-ine solution with 10 lL of 1-trimehylsilylimidazole (Wako

Pure Chemical Industries, Osaka, Japan) for more than

15 min at room temperature, part of the solution was

sub-jected to a GC-MS analysis performed with a

Hewlett-Packard 6890 gas chromatograph interfaced with a

MStation JMS-700 mass spectrometry system (JEOL,

Tokyo, Japan) A J&W DBTM-1 capillary column

(30 m· 0.25 mm, d.f ¼ 0.25 lm) was used for the GC

Samples were injected at 70C, and the temperature was

increased to 220C, at a rate of 50 CÆmin)1, and then to

320C, at 4 CÆmin)1, and held constant for 6 min

HOURSelium was used as the carrier gas The electron

impact-MS was performed at 70 eV with a mass range from

m⁄ z 50–750 and a cycle time 1 s in the positive ion mode

Acknowledgements

This work was supported by Grants-in-Aid from the Ministry of Education, Culture, Sports, Science, and Technology of Japan We are grateful to Dr K Ogura and Dr T Koyama, Tohoku University, for providing farnesyl diphosphate We wish to thank Dr C Ohto, Toyota Motor Co., for donating IPP and dimethyl-ally diphosphate We are grateful to S Kitamura, Nagoya University, for his technical assistance with the GC-MS analyses We also thank Dr Y Sakagami and

Dr M Ojika for helpful discussions on mass spectro-metry

References

1 De Rosa M & Gambacorta A (1988) The lipids of archaebacteria Prog Lipid Res 27, 153–175

2 Gambacorta A, Trincone A, Nicolaus B, Lama L & De Rosa M (1994) Unique features of lipids of Archaea Syst Appl Microbiol 16, 518–527

3 Koga Y & Morii H (2005) Recent advances in struc-tural research on ether lipids from archaea including comparative and physiological aspects Biosci Biotechnol Biochem 69, 2019–2034

4 Nishihara M & Koga Y (1995) sn-Glycerol-1-phos-phate dehydrogenase in Methanobacterium thermoauto-trophicum: key enzyme in biosynthesis of the

enantiomeric glycerophosphate backbone of ether phospholipids of archaebacteria J Biochem (Tokyo)

117, 933–935

5 Nishihara M & Koga Y (1997) Purification and proper-ties of sn-glycerol-1-phosphate dehydrogenase from Methanobacterium thermoautotrophicum: characteriza-tion of the biosynthetic enzyme for the enantiomeric glycerophosphate backbone of ether polar lipids of archaea J Biochem (Tokyo) 122, 572–576

6 Chen A & Poulter CD (1994) Isolation and characteri-zation of idsA: the gene for the short chain isoprenyl diphosphate synthase from Methanobacterium thermo-autotrophicum Arch Biochem Biophys 314, 399–404

7 Ohnuma S-i, Suzuki M & Nishino T (1994) Archaebac-terial ether-linked lipid biosynthetic gene Expression cloning, sequencing, and characterization of geranylger-anyl-diphosphate synthase J Biol Chem 269, 14792– 14797

8 Tachibana A (1994) A novel prenyltransferase, farnesyl-geranyl diphosphate synthase, from the haloalkaliphilic archaeon, Natronobacterium pharaonis FEBS Lett 341, 291–294

9 Tachibana A, Yano Y, Otani S, Nomura N, Sako Y &

Taniguchi M (2000) Novel prenyltransferase gene

encoding farnesylgeranyl diphosphate synthase from a

hyperthermophilic archaeon, Aeropyrum pernix

Molecu-lar evolution with alteration in product specificity Eur

J Biochem 267, 321–328

10 Chen A, Zhang D & Poulter CD (1993)

(S)-Geranylger-anylglyceryl phosphate synthase Purification and

char-acterization of the first pathway-specific enzyme in

archaebacterial membrane lipid biosynthesis J Biol

Chem 268, 21701–21705

11 Soderberg T, Chen A & Poulter CD (2001)

Geranylger-anylglyceryl phosphate synthase Characterization of the

recombinant enzyme from Methanobacterium

thermoau-totrophicum Biochemistry 40, 14847–14854

12 Nemoto N, Oshima T & Yamagishi A (2003)

Purifica-tion and characterizaPurifica-tion of geranylgeranylglyceryl

phosphate synthase from a thermoacidophilic archaeon,

Thermoplasma acidophilum J Biochem (Tokyo) 133,

651–657

13 Hemmi H, Shibuya K, Takahashi Y, Nakayama T &

Nishino T (2004) (S)-2,3-Di-O-geranylgeranylglyceryl

phosphate synthase from the thermoacidophilic

archaeon Sulfolobus solfataricus Molecular cloning and

characterization of a membrane-intrinsic

prenyltransfer-ase involved in the biosynthesis of archaeal ether-linked

membrane lipids J Biol Chem 279, 50197–50203

14 Daiyasu H, Kuma K, Yokoi T, Morii H, Koga Y &

Toh H (2005) A study of archaeal enzymes involved in

polar lipid synthesis linking amino acid sequence

infor-mation, genomic contexts and lipid composition

Archaea 1, 399–410

15 Kon T, Nemoto N, Oshima T & Yamagishi A (2002)

Effects of a squalene epoxidase inhibitor, terbinafine, on

ether lipid biosyntheses in a thermoacidophilic

archaeon, Thermoplasma acidophilum J Bacteriol 184,

1395–1401

16 Morii H, Nishihara M & Koga Y (2000) CTP:

2,3-di-O-geranylgeranyl-sn-glycero-1-phosphate

cytidyltransfer-ase in the methanogenic archaeon Methanothermobacter

thermoautotrophicus J Biol Chem 275, 36568–36574

17 Morii H & Koga Y (2003)

CDP-2,3-Di-O-geranylgera-nyl-sn-glycerol: 1-serine O-archaetidyltransferase

(archaetidylserine synthase) in the methanogenic

archaeon Methanothermobacter thermautotrophicus

J Bacteriol 185, 1181–1189

18 Nichols DS, Miller MR, Davies NW, Goodchild A,

Raftery M & Cavicchioli R (2004) Cold adaptation in

the antarctic archaeon Methanococcoides burtonii

involves membrane lipid unsaturation J Bacteriol 186,

8508–8515

19 Gibson JA, Miller MR, Davies NW, Neill GP, Nichols

DS & Volkman JK (2005) Unsaturated diether lipids in

the psychrotrophic archaeon Halorubrum lacusprofundi

Syst Appl Microbiol 28, 19–26

20 Collins MD & Jones D (1981) Distribution of isopre-noid quinone structural types in bacteria and their taxo-nomic implication Microbiol Rev 45, 316–354

21 Hemmi H, Takahashi Y, Shibuya K, Nakayama T & Nishino T (2005) Menaquinone-specific prenyl reductase from the hyperthermophilic archaeon Archaeoglobus fulgidus J Bacteriol 187, 1937–1944

22 Keller Y, Bouvier F, d’Harlingue A & Camara B (1998) Metabolic compartmentation of plastid prenyllipid bio-synthesis – evidence for the involvement of a multifunc-tional geranylgeranyl reductase Eur J Biochem 251, 413–417

23 Tanaka R, Oster U, Kruse E, Rudiger W & Grimm B (1999) Reduced activity of geranylgeranyl reductase leads to loss of chlorophyll and tocopherol and to par-tially geranylgeranylated chlorophyll in transgenic tobacco plants expressing antisense RNA for geranyl-geranyl reductase Plant Physiol 120, 695–704

24 Addlesee HA, Gibson LC, Jensen PE & Hunter CN (1996) Cloning, sequencing and functional assignment

of the chlorophyll biosynthesis gene, chlP, of Synecho-cystissp PCC 6803 FEBS Lett 389, 126–130

25 Teixidor P & Grimalt JO (1992) Gas chromatographic determination of isoprenoid alkylglycerol diethers in archaebacterial cultures and environmental samples

J Chromatogr A 607, 253–259

26 Nishimura Y & Eguchi T (2006) Biosynthesis of archaeal membrane lipids: digeranylgeranylglyceropho-spholipid reductase of the thermoacidophilic archaeon Thermoplasma acidophilum J Biochem (Tokyo) 139, 1073–1081

27 Mitaku S, Hirokawa T & Tsuji T (2002) Amphiphilicity index of polar amino acids as an aid in the characteriza-tion of amino acid preference at membrane–water inter-faces Bioinformatics 18, 608–616

28 Sambrook J, Fritsch EF & Maniatis T (1989) Molecular Cloning: a Laboratory Manual, 2nd edn Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY

29 Whitby LG (1953) A new method for preparing flavin-adenine dinucleotide Biochem J 54, 437–442

30 Bradford MM (1976) A rapid and sensitive method for the quantitation of microgram quantities of protein util-izing the principle of protein-dye binding Anal Biochem

72, 248–254

31 Fujii H, Koyama T & Ogura K (1982) Efficient enzy-matic hydrolysis of polyprenyl pyrophosphates Biochim Biophys Acta 712, 716–718

32 Bligh EG & Dyer WJ (1959) A rapid method of total lipid extraction and purification Can J Biochem Physiol

37, 911–917

33 Demizu K, Ohtsubo S, Kohno S, Miura I, Nishihara M

& Koga Y (1992) Quantitative determination of methano-genic cells based on analysis of ether-linked glycerolipids

by high-performance liquid chromatography J Ferment Bioeng 73, 135–139