Biochemical characterization of the water-soluble squalene synthase from Methylococcus capsulatus and the functional analyses of its two DXXD(E)D motifs and the highly conserved aromatic amino acid residues Kana Ohtake, Naoki Saito, Satoshi Shibuya, Wakako Kobayashi, Ryosuke Amano, Takumi Hirai, Shinji Sasaki, Chiaki Nakano and Tsutomu Hoshino Department of Applied Biological Chemistry, Faculty of Agriculture and Graduate School of Science and Technology, Niigata University, Japan

Keywords farnesyl diphosphate; Methylococcus capsulatus; NADPH; squalene; squalene synthase Correspondence T. Hoshino, Department of Applied Biological Chemistry, Faculty of Agriculture, Niigata University, Ikarashi 2-8050, Nishi-ku, Niigata 950-2181, Japan Fax: +81 25 262 6854 Tel: +81 25 262 6637 E-mail: [email protected] (Received 8 April 2014, revised 12 September 2014, accepted 30 September 2014) doi:10.1111/febs.13090

Information regarding squalene synthases (SQSs) from prokaryotes is scarce. We aimed to characterize the SQS from Methylococcus capsulatus. We studied its reaction mechanism by kinetic analysis and evaluated the structure of the substrate/inhibitor-binding sites via homology modeling. The cloned M. capsulatus SQS was expressed in Escherichia coli and purified by nickel-nitrilotriacetic acid column chromatography. Interestingly, M. capsulatus SQS was water-soluble and did not require any detergent for its higher activity, unlike other SQSs studied previously; supplementation of any type of detergent inhibited enzyme activity. The specific activity and the kinetic values (Km and kcat) for the substrate farnesyl diphosphate and NADPH are reported. The substrate analog farnesyl methylenediphosphonate showed potent inhibition toward the enzyme. We prepared the sitespecific mutants directed at potential active-site residues 58DXX61E62D (S1 site) and 213DXX216D217D (S2 site), which were assumed to be involved in the binding of the substrate farnesyl diphosphate through the Mg2+ ion. We first demonstrated that the S1 site and the two basic residues (R55 and K212) were responsible for the binding of farnesyl diphosphate. Furthermore, we examined the catalytic roles of the highly conserved aromatic residues and demonstrated that the Y164 residue abstracts the proton of cation 5, which is produced during the first half-reaction (Scheme 1), to afford presqualene diphosphate, and that the W224 residue stabilizes the intermediary cation 5 via the cation–p interaction. Furthermore, we confirm for the first time that the F32 and the Y51 residues also stabilize the carbocation intermediate(s) generated during the second half-reaction.

Introduction The biosynthesis of sterols by prokaryotic organisms is very rare, with the methanotrophic bacterium Methylococcus capsulatus being one of the few examples that

produces such steroid alcohols [1,2]. Squalene epoxidase and lanosterol synthase are prerequisite enzymes to cyclic sterol biosynthesis. In a previous study [3], we

Abbreviations CrtM, dehydrosqualene synthase from Staphylococcus aureus; DSQ, dehydrosqualene; DTT, dithiothreitol; FMDP, farnesyl methylenediphosphonate; FPMP, farnesyl (phosphinylmethyl)phosphonate; FPP, farnesyl diphosphate; GGOH, geranylgeraniol; mcSQS, Methylococcus capsulatus squalene synthase; OMY, O-methyltyrosine; PDB, Protein Data Bank; PPi, inorganic pyrophosphate; PSOH, presqualene; PSPP, presqualene diphosphate; SQS, squalene synthase; teSQS, Thermosynechococcus elongatus squalene synthase.

FEBS Journal 281 (2014) 5479–5497 ª 2014 FEBS

5479

Squalene synthase from M. capsulatus

K. Ohtake et al.

C11

A 1st Half-reaction: Ionization/Condensation C11

FPP (donor)

B

OPP

S1

H3C H(R)

R1

H(S)

S1

H(SH)

OPP

S2

R 1= H(R)

S1

OPP H(R)

H(R)

R1 PPO

S2 H(R) H(S)

H

H

H(R)

R1 FPP (acceptor)

R1 S2

H(S)

C14

H(S)

5: Cationic intermediate

6: Presqualene Diphosphate (PSPP)

B 2nd Half-reaction: Rearrangement/NADPH-dependent Reduction

H

H H2NOC

H(S) S1

H(R)

R1 H 2O

7a

H R1 HO

H(R)

H(R)

H(R)

CH3

R1

H

H(S)

H(R)

CH3

R1

S2

8a H

H

H 2O

H 2O

H

H

H CH3 R1

7b: HBO (Hydroxybotryococcene)

R1

CH3 HO R 1

8b: ROH (Rellingol)

R1

HO

H(R) R1

R1 CH3

9b: HSQ (Hydroxysqualene)

R1

H

+ NADPH R1

R1

9a

R1

NADPH

N R

H(S)

H(R)

CH3

CH3 H(S) H(R) 2: Squalene (SQ)

CrtM (dehydrosqualene synthase) - NADPH R1

deprotonation

R1

10: Dehydrosqualene (DSQ)

Scheme 1. Reaction mechanism from the substrate FPP (1) to the end-product squalene (2), which consists of two half-reaction steps: (A) the condensation reaction between two molecules of FPP to give PSPP (6), which proceeds at a fast rate even under the absence of NADPH, and (B) the reaction process of PSPP to yield squalene, consisting of two rearrangements and the nucleophilic attack of the hydride ion from NADPH. Under the absence of NADPH or the presence of the dihydro analog (NADPH3), the rearrangement reactions to afford the solvolysis and deprotonation products (7b–9b and 10) are substantially slower. On the other hand, in the presence of NADPH, the second half-reaction is significantly fast, without the cationic intermediates being trapped [16]. The DSQ synthase (CrtM) found in Staphylococcus aureus affords DSQ (10) from squalene. The reaction mechanism is analogous to that of squalene synthase (SQS), differing only in the final reaction: the proton elimination for CrtM and the nucleophilic attack of the hydride ion for SQS.

identified the genes encoding these two enzymes, which were functionally expressed in Escherichia coli. Structural analyses of the enzymatic products confirmed that the reactions proceed in a completely regio- and stereospecific fashion to afford (3S)-2,3-oxidosqualene from squalene and lanosterol from (3S)-2,3-oxidosqualene, in full accordance with the process found in eukaryotes. In addition to sterols, M. capsulatus also generates hopanoids [4], the carbocyclic skeleton of which is produced by the enzymatic action of squalene–hopene cyclase on squalene. By the catalytic action of squalene synthase (SQS; EC 2.5.1.21), a head-to-head condensation reaction (10 -2,3-linked) between two molecules of farnesyl diphosphate (FPP, 1; C15) forms a cyclopropylcarbinyl intermediate, presqualene diphosphate (PSPP, 6). The subsequent conversion of PSPP to squalene (2) involves an extensive rearrangement of the carbon skeleton and an NADPH-dependent reduction reaction. The chemical 5480

structures of FPP and squalene are shown in Fig. 1. In the case where NADPH is unavailable, PSPP is accumulated without yielding the end product, squalene. This reaction mechanism is depicted in Scheme 1. Until now, many SQSs have been cloned from eukaryotes and functionally expressed in E. coli as a truncated and soluble form [5–13]. However, investigations of prokaryotic SQSs are very limited, and only one example from Thermosynechococcus elongatus BP1 has been reported [14]. In the present study, we describe the detailed characterization of the SQS from M. capsulatus (mcSQS), which is exclusively water-soluble as an intact full-length form and does not require any type of detergent for its high activity. This SQS lacks the C-terminal sequence that is predicted to form a membrane-spanning a-helix [8,14,15]. All of the SQSs identified previously, including the bacterial T. elongatus SQS, require a detergent for high activity, although mcSQS was severely inhibited by any type of FEBS Journal 281 (2014) 5479–5497 ª 2014 FEBS

Squalene synthase from M. capsulatus

K. Ohtake et al.

O O

O

P –

X

O

P

O– O–

1. FPP X=O 3. FMDP X=CH2

Fig. 1. Chemical structures of FPP (1), squalene (2), FMDP (3) and FPMP (4).

C H2

O

P

P O– C – H – 2 O

O

4. FPMP

detergent. Thus, this is the first report of an SQS that does not require a detergent for its activity. Jarstfer et al. [16] and Blagg et al. [17] revealed that the incubation of FPP with SQS in buffer lacking NADPH results in the rapid conversion of FPP to PSPP, followed by a substantially slower enzyme-catalyzed solvolysis of PSPP to afford a mixture of hydroxybotryococcene (HBO, 7b), rellingol (ROH, 8b), hydroxysqualene (HSQ, 9b) and dehydrosqualene (DSQ, 10), with these latter intermediary products having been trapped by using NADPH3 (a dihydroanalog of NADPH) instead of NADPH. NADPH3 resembles NADPH in its chemical structure but does not possess the hydride transfer activity. The isolation of these compounds unambiguously demonstrated that the reaction from PSPP to squalene proceeds via two rearrangement steps, as shown in Scheme 1B. Staphylococcus aureus, a virulent bacterium, produces a carotenoid pigment (staphyloxanthin) that is biosynthesized through DSQ [18]. The mechanism of formation of DSQ is analogous to that of squalene. Although the DSQ synthase from S. aureus (CrtM) catalyzes the same reaction as SQS, it does not require NADPH. Recently, Liu et al. [19] and Lin et al. [20] reported the X-ray structures of the co-crystals of inhibitors complexed to CrtM, by which the catalytic mechanism of CrtM was proposed. However, reports describing the activities of site-directed mutated SQSs are lacking, with the exception of a report by Gu et al. [21]. Thus, details of the catalytic mechanism of SQSs are still uncertain. In the present study, we report the enzymatic activities of mutants targeted for the motifs 58 DXX61E62D (S1 site; Fig. S1) and 213DXX216D217D (S2 site), which were assumed to be involved in the binding of FPP through the Mg2+ ion. Gu et al. [21] studied the enzyme activity of some site-directed mutants targeted for the S2 site, although the functional analysis of the S1 site has not been reported previously. Based on site-directed mutagenesis experiments, we provide the first demonstration that the residues D58, D62 and E61 in the S1 site are crucial for

FEBS Journal 281 (2014) 5479–5497 ª 2014 FEBS

2. squalene(SQ) O

the catalytic action of SQSs. Similar to the report by Gu et al. [21], our mutational experiments support that residues D213 and D217 of the 213DXX216D217D sequence (S2 site) are critical, although D216 is not important for the catalysis. In addition to the conserved acidic motifs, we identified the basic amino acid residues R55 and K212 as new active sites, which possibly bind to the ionized inorganic pyrophosphate (PPi) moiety of FPP near the S1 site and S2 site, respectively. Furthermore, we checked whether farnesyl methylenediphosphonate (FMDP, 3) could inhibit SQS activity because its structure is very similar to that of FPP (Fig. 1). FMDP showed a strong inhibition effect in a mixed noncompetitive fashion, with an IC50 value of 0.37 lM and KiFPP of 0.10 lM. In addition, we aimed to identify the active-site residues that participate in the proton abstraction and/or stabilization of the cationic intermediate during the first half-reaction step (Scheme 1). We propose that Y164 acts as the proton abstractor, although it does not have a role in stabilizing the cation via the cation– p interaction, in contrast to the report by Gu et al. [21]. We demonstrate here that, instead of Y164, W224 has a function in stabilizing the intermediary cation 5, which is generated during the first half-reaction, via the cation–p interaction. We report the experimental evidence suggesting that the cation intermediates 7a–9a generated in the second half-reaction are also stabilized by the aromatic p-electrons of F32 and by the hydroxyl group in the Y51 residue, possibly via an unshared electron pair or a phenoxide anion. These aromatic residues are highly conserved in various SQSs (Fig. S1).

Results and Discussion Expression of mcSQS in E. coli and the enzyme purification The cloning, expression and purification of mcSQS were conducted as described in the Materials and methods.

5481

Squalene synthase from M. capsulatus

The amplified sqs gene was ligated into the pET22b vector and the pET22b-sqs construct was then transformed into E. coli BL21 (DE3). The expressed SQS protein was then purified by nickel-nitrilotriacetic acid column chromatography. b-Mercaptoethanol (12 mM; 0.1%) was used as the sulfhydryl reagent for purification through the nickel-nitrilotriacetic acid resin. According to the manufacturer’s protocol, the concentration of bmercaptoethanol used should be < 20 mM. After successive binding and washing, the solution of pure SQS (eluted with elution buffer containing 300 mM imidazole) was dialyzed against 50 mM Tris-HCl (pH 8.0) containing 0.5 mM DTT to remove the imidazole. DTT reagent was used because it had a better activity than bmercaptoethanol. The expressed amount of SQS was estimated to be approximately 25 mgL1 culture. Figure 2 shows the SDS/PAGE results for each of the purification steps. The molecular mass of the expressed SQS on the gel was approximately 40 kDa, which corresponded to the calculated mass of 41.1 kDa. The purity of the mcSQS was > 96%.

Fig. 2. Purification of the recombinant SQS from M. capsulatus. Lane 1, molecular weight markers. Lane 2, total protein. Lane 3, soluble protein. Lane 4, insoluble protein. Lane 5, the supernatant fraction that was not adsorbed on the His-tag resin with a batch protocol, with a binding buffer containing 0.1% 2-mercaptoethanol having been used. Lane 6, the SQS-adsorbed His-tag resin that had been placed into a short column, using a binding buffer containing 0.1% 2-mercaptoethanol. The eluted buffer solution was applied to the SDS/PAGE. Lane 7, the eluted fraction washed with a wash buffer containing 0.1% 2-mercaptoethanol. Lane 8, purified SQS eluted with the elution buffer [300 mM imidazole, 300 mM NaCl and 100 mM NaH2PO4-NaOH (pH 8.0) containing 0.1% 2mercaptoethanol]. The detailed protocol for the purification is described in the Materials and methods.

5482

K. Ohtake et al.

Unexpectedly, there was no SQS band from the insoluble fraction (lane 4) on the gel, confirming that the SQS is a water-soluble enzyme but is not membrane-bound. M. capsulatus SQS lacks the C-terminal domain responsible for a membrane-spanning a-helix [8,15] (Fig. S1). Recombinant SQSs truncated at the N- and/or C-terminus region are functionally expressed in E. coli and prepared as a soluble and active form by using a detergent [6–13]. All of the SQSs identified previously require a detergent for their high activity, with CHAPS and Tween 80 having been used mainly for activity enhancement [6–13]. Recently, Lee and Poulter [14] reported the first characterization of a prokaryotic SQS from T. elongatus BP-1 (teSQS; Fig. S1). This teSQS expressed in E. coli initially presented as inactive inclusion bodies. Tween 20, Tween 80 and polyethylene glycol could not render the soluble active teSQS, although glycerol was effective in doing so. Even after the glycerol was removed, this teSQS remained soluble and active. However, teSQS also required the detergent Tween 80 (2%, v/v) for its high activity. We thus investigated whether the mcSQS activity could be enhanced by various detergents (Fig. 3). The neutral detergents, Triton X-100 and CHAPS; the anionic detergent, sodium deoxycholate; and the cationic detergent, hexadecyltrimethyl ammonium bromide, all strongly inhibited the enzyme activity. In the case of Tween 80 and glycerol, the activity decreased to 20–40% at a high detergent concentration. The enzymatic activity in the presence of Tween 20 was almost the same as that in buffer without the detergent (Fig. 3), indicating that Tween 20 does not enhance the enzyme activity and its inhibition effect is minor. Other detergents, such as n-dodecyl-b-D-maltoside, noctyl-b-D-glucopyranoside and IGEPAL CA-630, also did not enhance the mcSQS activity but, instead, inhibited it. Thus, it can be concluded that mcSQS is water-soluble and does not require any detergent for enhancement of its activity. Therefore, the optimal catalytic conditions and kinetic constants for mcSQS were determined in buffer solution without detergents. Optimal catalytic conditions and kinetic constants The graphs in Fig. 4 show the optimal incubation conditions for the SQS: temperature 33 °C, pH 8.0, and 5 mM MgCl2 and 10 mM DTT. Under these optimal conditions, the enzyme activity increased linearly up to 15 min after incubation, and the specific activity of the wild-type was determined to be 800  36 nmolmin1mg1. The Michaelis–Menten plot is FEBS Journal 281 (2014) 5479–5497 ª 2014 FEBS

Squalene synthase from M. capsulatus

120 100 80

Triton X-100 Tween 80 CHAPS Tween 20

60 40 20 0 0

5

Relative specific activity (%)

Relative specific activity (%)

K. Ohtake et al.

120 sodium deoxycholate

100

hexadecyltrimethylammonium bromide

80

glycerol

60 40 20 0

10

0

Detergent concentration (w/v%)

10

5

Detergent concentration (w/v%)

Fig. 3. Effects of detergent concentration on the SQS activity. None of the detergents employed in the present study (neutral, cationic and anionic) enhanced the enzyme activity; instead, they acted as inhibitors. The maximum specific activity of the wild-type SQS was 800  36 nmolmin1mg1 under optimal conditions. The detergent concentrations (%, w/v) of Triton X-100, Tween 80, CHAPS and Tween 20 were 0, 0.05, 0.1, 0.2, 0.3, 0.5, 0.75, 1, 2, 3, 5 and 7.5; the concentrations (%, w/v) of sodium deoxycholate, hexadecyltrimethyl ammonium bromide and glycerol were 0, 0.1, 0.5, 1, 2, 4 and 6.75.

B 120

Relative specific activity (%)

Relative specific activity (%)

A 100 80 60 40 20 0 0

20

40

120 100 80 60 40 20 0 5

60

7

9

D 120 MgCl2

100

MnCl2 80 60 40 20 0 0

20

40

MgCl2,MnCl2 [mM]

60

Relative specific activity (%)

C Relative specific activity (%)

11

pH

Temperature (°C)

120 100 80 60 40 20 0 0

20

40

60

DTT [mM]

Fig. 4. Determination of the optimal conditions for squalene formation by SQS. (A) Effect of incubation temperature on the enzymatic activity. (B) Effect of pH on the enzymatic activity. The buffer solutions used were: MES-NaOH (pH 5.5–6.0), MOPS-NaOH (pH 6.5–7.3), Tris-HCl (pH 7.5–9.0) and Na2CO3-NaHCO3 (pH 9.5–10.5). (C) Effect of divalent cations (Mg2+ and Mn2+) on the specific activity. The concentrations of MgCl2 were 0.025, 0.5, 1, 2, 3, 5, 10, 20 and 40 mM, whereas those of MnCl2 were 0.3125, 0.625, 1.25, 2.5, 5, 10, 20 and 40 mM. The maximum activity observed at 2.5 mM of MnCl2 corresponds to 42% of the maximum activity in the 5 mM MgCl2 solution. (D) Determination of the optimal concentration of DTT required. The concentrations of DTT used were 0, 0.5, 1, 2, 3, 5, 10, 20 and 40 mM. To determine the optimal concentration of Mg2+ or DTT, they were omitted from the standard conditions described above, and then their concentrations were varied. The specific activity corresponding to 100% relative activity was determined to be 800  36 nmolmin1mg1 under optimal catalytic conditions: pH 8.0 (0.1 M Tris-HCl), 33 °C, 5 mM MgCl2 and 10 mM DTT. Fifty micromolar farnesyl diphosphate, 1.0 lg of the purified M. capsulatus squalene synthase (10 nM) and 2.5 mg of BSA were used.

FEBS Journal 281 (2014) 5479–5497 ª 2014 FEBS

5483

Squalene synthase from M. capsulatus

K. Ohtake et al.

shown in Fig. 5. Steady-state parameters were determined by fitting the curve to v = Vmax([S]/(KM + [S]) using SOLVER in EXCEL (Microsoft Corp., Redmond, CA, USA). The Km and kcat values for FPP were determined by using 10 nM of mcSQS and by incubating for 15 min (Fig. 5A) to be 13.6  0.92 lM and 0.549  0.01 s1, respectively, and kcat/Km was 0.04 lM1s1. Another experiment (1.0 nM mcSQS; 10-min incubation) was conducted to validate the kinetic values; the quantity of squalene produced was significantly decreased compared to that produced with 10 nM mcSQS, although the peak area of squalene was sufficiently detected to quantify the squalene product (see the detailed description in the Materials and meth0.35

B 0.35

0.3

0.3

0.25

0.25

V0 (μ/min)

V0 (μ/min)

A

ods). Under these incubation conditions, the Km and kcat values were estimated to be 12.2  0.75 lM, kcat = 0.548  0.01 s1 (Fig. S2), respectively, which were almost the same as those estimated with 10 nM mcSQS. To validate the kinetic data measured by the GC analyses, we further conducted a radioassay using [1-3H]FPP, which was essentially the same as the protocol reported by Lee and Poulter [14]; the buffer contained 0.25–50 lM [1-3H]FPP and 1 nM mcSQS (see Materials and methods), giving the kinetic values: Km = 13.1  0.99 lM, kcat 0.407  0.02 s1 and kcat/ Km = 0.031 lM1s1 (Table 1 and Fig. S3A). The kcat was somewhat decreased compared to that obtained by the GC method. For NADPH (Fig. 5B), the

0.2 0.15

0.2 0.15

0.1

0.1

0.05

0.05 0

0 0

20

40

0

60

250

FPP (μ)

500

750

1000

NADPH (μ)

Fig. 5. Michaelis–Menten plots of the mcSQS activity on FPP (A) and NADPH (B), as determined by GC analyses. Kinetic parameters for FPP were obtained by fitting the data to the Michaelis–Menten equation, using the SOLVER function in EXCEL. The results obtained were: Km = 13.6  0.92 lM, kcat = 0.549  0.01 s1 and kcat/Km = 0.04 lM1s1. For NADPH, the parameters were: Km = 75.4  1.02 lM, kcat = 0.33  0.0015 s1 and kcat/Km = 0.004 lM1s1. On the other hand, the Michaelis–Menten plot of mcSQS for NADH showed the values: Km = 2899  285 lM, kcat = 0.38  0.02 s1 and kcat/Km = 0.00013 lM1s1 (Fig. S4), indicating that the NADH binding to mcSQS was significantly looser and the catalysis slower than that of NADPH. The error bars show the differences between two independent experiments.

Table 1. Kinetic constant values for SQSs from different species and sources.

Species and/or source Methylococcus capsulatus Thermosynechococcus elongtus BP-1 Yeast Trypanosoma cruzi TcSQS24/36 Leishmania donovani Leishmania mexicana Rat Human

kcat (s1)

13.6  0.92 13.1  0.99b 0.97  0.10

75.4  1.02 105.5  9.5b 241  13

0.549  0.01 0.407  0.02b 1.74  0.04

2.5  0.46 2.3 5.25  1.2 3.8 2.8 1.8 2.8

530  77 33 23.34  4.5 43.23 57

KmFPP (lM)

Full-length Full-length C-terminal truncated Microsomal/mitochondrial C- and N-double-truncated Microsomal/mitochondrial Microsomal Double-truncated

kcat/KmFPP (lM1s1)

KmNADPH (lM)

SQS type

a

a

a

a

References

0.04 0.031b 1.80

Present study Present study Lee and Poulter [14]

0.53  0.03

0.21

1.05  0.16

0.20

LoGrasso et al. [7] Urbina et al. [11] Sealey-Cardona et al. [10] Bhargava et al. [12] Urbina et al. [11] Thompson et al. [9] Thompson et al. [9]

1.44

0.51

a

Determined by GC methods. Two separarte experiments were conducted. Another experiment (1.0 nM mcSQS, 10-min incubation) was conducted to validate the kinetic values; KmFPP = 12.2  0.75 lM, kcat = 0.548  0.01 s1 (see Materials and methods). b Determined by radioassay. Two independent experiments were conducted.

5484

FEBS Journal 281 (2014) 5479–5497 ª 2014 FEBS

Squalene synthase from M. capsulatus

K. Ohtake et al.

Inhibition of mcSQS activity by the substrate analog FMDP 3 In 2000, the X-ray crystal structures of human SQS bound with the inhibitors CP-320473, CP-458003 and CP-424677 were resolved [22]; the structures of these inhibitors are apparently different from that of the substrate FPP, intermediate PSPP and product squalene. We tested the inhibitory effect of the substrate-like compound FMDP 3 on mcSQS. The substrate analog 3 was chemically synthesized according to the method described in Fig. S5. FMDP 3 strongly inhibited the enzyme activity, with an IC50 value of 0.37 lM and KiFPP of 0.1 lM, indicating that the affinity of FMDP 3 to mcSQS was approximately 130-fold higher than that of the substrate FPP (Ki