A subunit of decaprenyl diphosphate synthase stabilizes octaprenyl diphosphate synthase in Escherichia coli by forming a high-molecular weight complex

2.1. Materials 

DNA markers and restriction enzymes were obtained from TOYOBO (Osaka, Japan). Protein markers were obtained from Fermentas Life Sciences (Ontario, Canada) and Oriental Yeast (Tokyo, Japan). Antibodies were obtained from Santa Cruz Biotechnology (Santa Cruz, CA, USA). Isopentenyl diphosphate (IPP) and all-E-farnesyl diphosphate (FPP) were obtained from Sigma Chemical Co. (St. Louis, MO, USA). [1–14C]IPP (1.96TBqmol−1) was obtained from Amersham (Little Chalfont, UK). Kieselgel 60 F254 TLC plates were purchased from Merck (Rahway, NJ, USA). Reversed-phase LKC-18 thin-layer plates were obtained from Whatman (Maidstone, UK). The Blue Native-PAGE NOVEX Bis-Tris Gel System and the NativeMark Unstained Protein Standard were obtained from Invitrogen (Osaka, Japan). Blue Native-PAGE was performed according to the manufacturer’s instructions.

2.2. Plasmid construction 

To construct pBQ-His-ispB or pBQ-His-ispBR321A, the primers 5′-CGGATCCGATGAATTTAGAAAAAATC-3′ (forward) and 5′-CGAAGCTTGGCCATGGGCGCG-3′ (reverse) were used to amplify ispB from pKO56 [3] and pBRA(R321A). The amplified fragments were first cloned into the BamHI and HindIII sites of a pQE-31 vector (Qiagen), and the fragments containing His6-ispB and His6-ispBR321A were digested with EcoRI and HindIII, before cloning into the same restriction sites in pBluescript II KS+ (Stratagene). To construct pSTVK-msps1, mSPS1 was released from the pBmSPS1 plasmid using EcoRI and KpnI [11], and cloned into the same sites in pSTVK28. To construct pSTVK-hdps1, hDPS1 was digested from pGEX-hdps1 using BamHI and XbaI [11], and cloned into the same sites in pSTVK28 (Table 1).

Table 1. Plasmids used in this study.

	Plasmids	Characteristics	Sources
	pSTVK-msps1	Km, full-length mouse SPS1 in pSTV28K	This study
	pSTVK-hdps1	Km, full-length human DPS1 in pSTV28K	This study
	pBQ-His-ispB	Ap, His6 with full-length ispB in pBluescript II KS+	This study
	pBQ-His-ispBR321A	Ap, His6 with full-length ispBR321A in pBluescript II KS+	This study
	pBRA(R321A)	Ap, 2.5-kb fragment including full-length ispBR321A in pBluescript	[13]
	pSTVK-His-dps1	Km, His6 with full-length dps1 in pSTVK28	[9]
	pSTVK-His-dlp1	Km, His6 with 1.1-kb fragment including full-length dlp1 in pSTVK28	Lab stock
	pSTVmDLP1	Km, full-length mouse DLP1 in pSTVK28	[11]
	pSTVhDLP1	Km, full-length human DLP1 in pSTVK28	[11]
	pGEX-1	Amp, tac promoter, GST tag, high expression vector	Amersham
	pGKO56	Amp, full-length ispB gene in pGEX-1X	[13]
	pGKO56-R321A	Amp, full-length ispBR321A gene in pGEX-1X	This study
	pSTVK-His-hDPS1	Km, His6 with full-length human DPS1 in pSTVK28	This study
	pSTVK-HIS-hDLP1	Km, His6 with full-length human DLP1 in pSTVK28	[11]

Ap, ampicillin; Km, kanamycin.

To construct pGKO56-R321A, site-directed mutagenesis was performed using a QuickChange II XL Site-Directed Mutagenesis Kit (Stratagene, Tokyo, Japan) as previously described [12]. Briefly, a pair of reverse-complementary primers (5′-CATCGCTGTTCAAGCCGATCGTTAATCC-3′, only the forward primer shown) was used to amplify pGKO56. The amplified PCR products were self-ligated, and the plasmid was recovered from E. coli to obtain pGKO56-R321A. The sequences of the mutants were confirmed using an ABI3100 DNA sequencer (Applied Biosystems, Foster City, CA, USA).

2.3. Purification of GST-IspB from E. coli 

GST-IspB, GST-IspBR321A and GST were purified as previously described [9].

2.4. Ubiquinone extraction and PDS assay 

Ubiquinone was extracted and measured as previously described [12]. PDS activity was measured as previously described [11].

3.1. Expression of dps1 or dlp1 reversed the thermo-sensitivity of the E. coli ispBR321A mutant 

It is known that the ispB gene is essential for E. coli growth [7] and that Arg321 is important for the thermo-stability of IspB; so ispBR321A mutants grow very slowly at 43°C [13]. Surprisingly, when we expressed S. pombe dps1 in the ispBR321A mutant, the growth of the co-expressed cells was similar to wild-type cells (Fig. 1). S. pombe dlp1 also reversed the thermo-sensitivity of the ispB mutant in a similar manner (Fig. 1). Also, we observed that both human hDLP1 and mouse mDLP1, but not hDPS1 or mSPS1, rescued the growth of the ispBR321A mutant at 43°C (Fig. 1), indicating the functional conservation of Dlp1 in a broad range of organisms. Because neither Dlp1, nor Dps1 alone can function as an active PDS, these observations suggest that Dps1, and Dlp1 are capable of supporting the IspB activity in E. coli. We then focused on the role of S. pombe Dps1 and Dlp1 in the E. coli ispBR321A mutant.

View Large Image Download to PowerPoint

Fig. 1. Dps1 or dlp1 reverses the thermo-sensitivity of an E. coli ispBR321A mutant. An overnight culture of the indicated strains was diluted to OD600=0.02, and the cells were incubated with vigorous shaking at 43°C. Cell mass was calculated every 30min. E. coli strain KO229 (ΔispB harboring an ispB-containing plasmid) was used as the wild-type control. The plasmid pKA3 in the KO229 E. coli strain was replaced by pBRA (R321A) in order to express the mutant ispB. Expression of dps1, mSPS1, or hDPS1 from S. pombe, mouse, or human, respectively, was achieved by introduction of pSTVK-His-dps1, pSTVK-msps1, or pSTVK-hdps1 into the mutant ispB cells. Expression of dlp1, mDLP1, or hDLP1 from S. pombe, mouse, or human, respectively, was achieved by the introduction of pSTVK-His-dlp1, pSTVmDLP1, or pSTVhDLP1 into the mutant ispB cells.

3.2. IspB plays a major role in ubiquinone synthesis in the dps1 or dlp1 co-expressed cells 

We then analyzed the Q-species in the co-expressed cells to ascertain the role played by Dps1 or Dlp1 within the complex, as dps1 and dlp1 are responsible for the synthesis of Q-10, and ispB for the synthesis of Q-8. The expression of dps1 or dlp1 by the ispBR321A mutant did not change the ubiquinone species at either 37°C (data not shown), or 43°C (Fig. 2C–F); also, the Q-8 levels in all the samples were similar (Fig. 2). This suggests that IspB plays a major role in Q-8 synthesis within the co-expressed cells, and that the contribution of Dps1 or Dlp1 may be to assist the function of IspB.

View Large Image Download to PowerPoint

Fig. 2. Expression of dps1 or dlp1 does not alter the Q species in E. coli. Both the wild-type and the ispBR321A E. coli strains were incubated at the indicated temperatures and ubiquinone was extracted. To test the effect of dps1 and dlp1 on ubiquinone synthesis, the plasmid harboring dps1 or dlp1 was introduced into the mutant ispBR321A cells.

3.3. Low PDS activity in the early growth phase of the ispBR321A mutant leads to its thermo-sensitivity, and is complemented by dps1 or dlp1 

To further analyze the roles of Dps1 and Dlp1, and the reason for defects in the IspB mutant, we looked at the enzymatic activities of poly-PDSs at both 30°C and 43°C, and at several points in the growth phase, by measuring the intensity of incorporated radioactive products. As expected, neither dps1, nor dlp1 affected the activity of the mutant IspB at 30°C (Fig. 3A, left). However, the expression of dps1 or dlp1 led to a fivefold increase in activity at 43°C at OD600=0.4 (Fig. 3A, right). Because neither dps1, nor dlp1 changed the level of IspB expression (Fig. 3B), it is reasonable to suppose that dps1 or dlp1 promotes the stability of the mutant IspB via a physical interaction. Interestingly, at the later point in the growth phase, the PDS activity of the mutant cells increased very quickly and was similar to that of the co-expressed cells at 43°C (Fig. 3A right). This may have been due to increased IspB levels after OD600=0.4 (Fig. 3B) and the limited availability of substrate. Consistently, the thermo-sensitivity of the IspB mutant was only observed earlier in the growth phase, and the mutant cells showed a wild-type-like generation time in the exponential phase (data not shown). Thus, these data indicate that higher protein levels can compensate for the structural defects within the IspB mutant. To the contrary, the low activity of the mutant IspB during the early phase of growth could not support normal growth, and so led to the thermo-sensitivity of the mutant cells.

View Large Image Download to PowerPoint

Fig. 3. Analysis of PDS activity and IspB protein levels. (A) Growth phase-specific activity of PDSs in the E. coli cells. The E. coli cells were grown at both 30°C and 43°C and collected at the indicated OD600 values. The PDS activity of E. coli was assayed at both 30°C and 43°C using [1–14C]IPP and FPP as substrates. After removal of excess [1–14C]IPP, the amount of radioactivity in the products was measured to give an indication of PDS activity. The strains used were: WT, ΔispB harboring pBQ-His-ispB; R321A, ΔispB harboring pBQ-His-ispBR321A; R321A+dps1, ΔispB harboring pBQ-His-ispBR321A and pSTVK-His-dps1; and R321A+dlp1, ΔispB harboring pBQ-His-ispBR321A and pSTVK-His-dlp1. The data represent the means±S.D. of five separate measurements. (B) Phase-specific IspB levels in the E. coli cells. E. coli cells were grown at 43°C, and collected at the indicated OD600 values. 5μg of the crude extracts obtained from the cells were used for immunoblot analysis with an anti-penta-His antibody (diluted 1:1000). The bands corresponding to His6-IspB or His6-IspBR321A were then compared. The same strains were used as in (A).

3.4. Physical interaction between IspB and Dps1, or Dlp1, in E. coli 

The above results suggest that Dps1 and Dlp1 might directly interact with and stabilize IspB in E. coli. Next, we performed Western blot analysis to check whether Dps1 or Dlp1 was co-precipitated with mutant IspB at 43°C. As expected, the purified GST-IspBR321A bound to either His6-Dps1 or His6-Dlp1 in the co-expressed cells (Fig. 4A and B, lane 1), and the purified GST (the negative control) did not (Fig. 4A and B, lane 4). The binding was not limited to the IspB mutant. The wild-type GST-IspB bound to His6-Dps1 or His6-Dlp at both 43°C and 37°C. This confirms that Dps1 and Dlp1 reverse the thermo-sensitivity of the IspB mutant cells by increasing the stability of IspB.

View Large Image Download to PowerPoint

Fig. 4. Physical interactions of Dps1–IspB and Dlp1–IspB in E. coli. (A) Dps1 binds to IspB in E. coli. E. coli DH5α cells harboring the corresponding plasmids were grown at the indicated temperatures. The purified GST-IspB, GST-IspBR321A and GST samples were then subjected to Western blot analysis using an anti-penta His antibody to examine the co-precipitation of Dps1 with IspB. pGKO56 and pGKO56-R321A were used to express GST-IspB and GST-IspBR321A, respectively. pGEX-1 was used to express the GST protein and pSTVK-His-dps1 for His6-Dps1. (B) Dlp1 binds to IspB in E. coli. As in (A), except that the dps1 expression plasmid was replaced by the dlp1 expression plasmid, pSTVK-His-dlp1.

3.5. IspB–Dps1 or IspB–Dlp1 form high-molecular weight complexes in E. coli 

As shown previously, both the native Dps1–Dlp1 complex and the artificial Coq1–Dps1 complex form a heterotetramer in S. pombe [9], [10]. To investigate how Dps1 and Dlp1 interact with IspB, we employed Blue Native PAGE to analyze these complexes. First, we looked at the homomeric IspB complex, which is known to form a homodimer in vitro [13]. Unexpectedly, both GST-IspB and GST-IspBR321A formed tetramers, dimers and polymers, as judged by the molecular size of the proteins (Fig. 5A). Because Blue Native PAGE maintains proteins in a more ‘native’ state compared with the cross-linking technique, these data imply that tetramers might be the main form of IspB found in E. coli. It also indicates that mutation of R321A does not prevent IspB from forming these high order structures. We also analyzed the size of the IspB–Dps1 and IspB–Dlp1 complexes. Because the expression levels of GST-IspB were much higher than those of His6-Dps1 or His6-Dlp1 in the co-expressed cells, the amounts of the IspB–Dps1 or IspB–Dlp1 complex are not comparable with those of the IspB complex. Therefore, immunoblotting must be used to detect the heteromeric complexes. To our surprise, the molecular weight of both the IspB–Dps1 and the IspB–Dlp1 complexes was approximately 800kDa (Fig. 5B, lanes 2 and 3). Clearly, these were novel heteromeric poly-PDSs, and were quite different from the heterotetramers seen previously. In addition, we found that a human Dps1 homolog did not form such a complex with IspB (Fig. 5B, lane 4). These data suggest a structural difference between S. pombe Dps1 and its mammalian homologs, which is supported by a report showing that the S. pombe Dps1 cannot form a complex with mammalian Dlp1.

View Large Image Download to PowerPoint

Fig. 5. High-molecular weight complexes formed by Dps1–IspB and Dlp1–IspB. (A) IspB forms a tetramer in E. coli. ΔispB harboring pGKO56 and ΔispB harboring pGKO56-R321A were grown at 43°C and used to purify wild-type and mutant IspB, respectively. The purified GST-IspB was then analyzed using the Blue Native PAGE system on 4–16% Bis-Tris gels. Molecular weight markers range in size from 20 to 1200kDa. (B) Dps1 and Dlp1 form high-molecular weight complexes with IspB. GST-IspB was prepared as described in (A) from ΔispB harboring pGKO56 (lane 1), ΔispB harboring pGKO56 and pSTVK-His-dps1 (lane 2), ΔispB harboring pGKO56 and pSTVK-His-dlp1 (lane 3), ΔispB harboring pGKO56 and pSTVK-His-hDPS1 (lane 4), and ΔispB harboring pGKO56 and pSTVK-HIS-hDlp1 (lane 5), and analyzed by Blue Native PAGE. The native gel was then transferred onto a PVDF membrane and an anti-penta-His antibody (diluted 1:500) was used to detect the complexes.