Negative Feedbacks by Isoprenoids on a Mevalonate Kinase Expressed in the Corpora Allata of Mosquitoes

Background Juvenile hormones (JH) regulate development and reproductive maturation in insects. JHs are synthesized through the mevalonate pathway (MVAP), an ancient metabolic pathway present in the three domains of life. Mevalonate kinase (MVK) is a key enzyme in the MVAP. MVK catalyzes the synthesis of phosphomevalonate (PM) by transferring the γ-phosphoryl group from ATP to the C5 hydroxyl oxygen of mevalonic acid (MA). Despite the importance of MVKs, these enzymes have been poorly characterized in insects. Results We functionally characterized an Aedes aegypti MVK (AaMVK) expressed in the corpora allata (CA) of the mosquito. AaMVK displayed its activity in the presence of metal cofactors. Different nucleotides were used by AaMVK as phosphoryl donors. In the presence of Mg2+, the enzyme has higher affinity for MA than ATP. The activity of AaMVK was regulated by feedback inhibition from long-chain isoprenoids, such as geranyl diphosphate (GPP) and farnesyl diphosphate (FPP). Conclusions AaMVK exhibited efficient inhibition by GPP and FPP (Ki less than 1 μM), and none by isopentenyl pyrophosphate (IPP) and dimethyl allyl pyrophosphate (DPPM). These results suggest that GPP and FPP might act as physiological inhibitors in the synthesis of isoprenoids in the CA of mosquitoes. Changing MVK activity can alter the flux of precursors and therefore regulate juvenile hormone biosynthesis.


Introduction Insects
A. aegypti of the Rockefeller strain were reared at 28°C and 80% relative humidity under a photoperiod of 16 h light: 8 h dark. A cotton pad soaked in 3% sucrose solution was provided to adults.

Sequence analysis and homology modeling
Sequences similarity searches were performed using the alignment tool BLAST [23]. MVK amino acid sequences were obtained from the National Center of Biotechnology Information and Vector Base. Analyses of degrees of similarity among sequences were performed using the ClustalW tool [24]. AaMVK secondary structure was predicted using PDBsum [25]. Amino acid sequence alignments were performed using Muscle [26]. Motifs from aligned sequences were selected, and consensus sequence logos were built using Weblogo [27]. AaMVK tertiary structure was modeled using the protein structure homology-modeling server Swiss v.8.05 and rat MVK (PDB code 1KVK) as template.

Expression of recombinant A. aegypti mevalonate kinase
The AaMVK cDNA was expressed in E. coli cells as described by Nyati et al. [28]. Recombinant His-tagged proteins were purified using HiTrap affinity columns and PD-10 desalting columns (Amersham Pharmacia, Piscataway, NJ). Glycerol was added to the enzyme solution (final concentration 50%), and samples were stored at -20°C until used. Protein concentrations were determined using the bicinchoninic acid (BCA) protein assay reagent (Pierce, Rockford, IL). Bovine serum albumin was used as a standard.

Enzyme assays
The catalytic activity of AaMVK was measured indirectly using a spectrophotometric assay that couples ADP formation to pyruvate synthesis and reduction to lactate [17,18]. The disappearance of NADH (measured at 340 nm) serves as a measurement for the phosphorylation of MA by MVK. Samples were incubated for 10 min at 30°C. Each 100 μl reaction mixture contained 0.5 mM phosphoenolpyruvate, 0.01 mM DTT, 0.35 mM NADH, 10 mM MgCl 2 , 2 units of LDH, and 2 units of PK in 100 mM Tris-HCl pH 7.6. Phosphorylation of MVA was analyzed in reactions containing ATP (250 μM) and MA (200 μM). Assays were performed in triplicate in 96-well plates (BioTek, Winooski, VT).
Reaction products from the catalytic activity of AaMVK were evaluated by reverse-phase HPLC (RP-HPLC). Briefly, recombinant protein (150 ng) was incubated for 1 h in the reaction buffer (100 mM Tris-HCl pH 7.5, 10 mM MgCl 2 , 0.5 mM DTT) containing: MA (200 μM) and ATP (250 μM). Reactions were terminated by adding 500 μl of acetonitrile, vortexed for 1 min, and centrifuged at 14,000 rpm for 5 min. The organic phase containing PM was recovered, filtered and analyzed by RP-HPLC on a Dionex Summit System (Dionex, Sunnyvale, CA) as previously described [28]. Water and glycerol were used in place of recombinant enzyme as negative controls.

Kinetic parameters
The Michaelis-Menten constant for MA (K m-MA ) was determined at a saturating concentration of ATP (5 mM), with MA concentrations ranging from 0.005 to 2.5 mM. Reactions were initiated with the addition of 150 ng of recombinant AaMVK. The K m-ATP was determined using saturating concentrations of MA (1.25 mM) and ATP concentrations ranging from 0.005 to 5 mM. The amount of NADH oxidized to NAD + was monitored at 340 nm. To determine steady-state kinetic parameters, data were subjected to nonlinear regression fits to the Michaelis-Menten equation using the GraphPad Prism software (San, Diego, CA).
Inhibition studies were performed in triplicate by adding different MVAP intermediates (DPM, DMAPP, IPP, GPP and FPP) to the reaction mix, as well as GGPP at various concentrations (0-1 μM). Inhibition constants (K i ) for GPP, FPP and GGPP were calculated after multicurve fits using the GraphPad Prism software.

AaMVK activity in extracts of mosquito thoraces
Mevalonate kinase activities in thoraces of female adult mosquitoes were measured by monitoring the production of PM using RP-HPLC. Thoraces from 24h old 3% sugar-fed females were dissected in Aedes physiological saline (APS) (138 mM NaCl, 8.4 mM KCl, 4 mM CaCl 2 , 2 mM MgCl 2 , 12 mM Na 2 HPO 4 and 42.5 mM sucrose), and transferred to a buffer solution (100 mM Tris-HCl pH 7.5, 10 mM MgCl 2 , 0.01 mM DTT). Thoraces were homogenized for 1 min, sonicated 3 min and centrifuged at 10,000 g for 10 min at 4°C. Supernatants were recovered and used as crude extract (CE) for activity assays as previously described [29]. The protein contents of the CE were measured using the BCA assay. Enzymatic assays were performed using 4 mg of protein as previously described. Boiled crude extract and reactions without enzyme were included as controls. A standard curve was constructed for the quantification of PM.
IPP in CE were measured by RP-HPLC after conversion into the corresponding alcohol by treating samples with 50 μL of 2.5 N HCl for 10 min [30]. Afterward, 500 μL of hexane were added, samples were vortexed for 1 min and centrifuged at 14, 000 rpm for 10 min at 4°C. Organic phases (containing the alcohols) were recovered, filtered through a 0.2 μm nylon filter and quantified by RP-HPLC.

Statistical analysis
Statistical analyses were performed using the GraphPad Prism Software (San Diego, CA, USA). The results are expressed as means ± S.E.M. Significant differences (P < 0.05) were determined with a one-tailed Student's t-test or one-way ANOVA followed by a pair-wise comparison of means (Tukey's test).

Molecular characterization of A. aegypti MVK
The full-length AaMVK open reading frame is 1818 bp long (AAEL006435) [31,32], and encodes a 397-aa protein with a calculated molecular mass of 43.27 kDa and a pI of 5.8. Amino acid sequence alignments of MVKs from insect and vertebrate species revealed 30-44% similarities (S1 Fig). The three conserved motifs that characterize MVKs were highly conserved among all the sequences analyzed, corroborating the functional role of AaMVK as a kinase ( Fig  1). Motif I: containing part of the active site (PGKVILXGEHSVVXXXPA); motif II: a conserved glycine-rich motif (SIGXGLGSSAG) that forms a phosphate-binding loop in all GHMP kinases, and motif III: a conserved amino acid sequence (KLTGAGGGGC) that stabilizes the phosphate binding loop [5,9].

Structural analysis of the active site of AaMVK
The molecular model of AaMVK was built by homology modeling using the rat MVK (PDB: 1kvk), which exhibited 32.7% identity to AaMVK, as template (Fig 2). The analysis of the AaMVK structure revealed a fold consisting of a mixture of α-helices and β-sheets (S2 Fig). The N-terminal domain is composed of ten β sheets and eight αhelices and the C-terminal domain is composed of four helices and two β sheets. The larger N-terminal (include amino acids 1 to 246, 358 to 397) and the smaller C-terminal (include amino acids 247 to 357) domains are arranged in a V-shape that creates a central cleft, with the AaMVK ligand binding pocket, composed by Lys 14 , Ser 159 , Glu 208 , and Asp 219 , located at the cleft between the two domains. Similar structures have been previously described for other MVKs [4,5,33].

Functional characterization of AaMVK
The recombinant AaMVK was isolated from E.coli extracts by affinity chromatography, and identified using an anti-His antibody (S3 Fig) as previously described [28]. Optimal catalytic conditions were initially established for AaMVK. The recombinant enzyme displayed activity The catalytic activity of recombinant AaMVK increased in a dose response manner when Mg 2+ was used as a cofactor. Mn 2+ and Co 2+ also enhanced MVK activity to a lesser degree than Mg 2+ (Fig 3A).
The nucleotide specificity of AaMVK was measured in the presence of different triphosphates phosphoryl donors. Relative rates of MVK activity when ATP, GTP, TTP and CTP were used were 100, 15, 11 and 5 percent respectively (Fig 3B).

Kinetic analyses of AaMVK
AaMVK showed normal Michaelis-Menten kinetics, which were obtained using a range of ATP concentrations and a fixed mevalonic acid (MA). The apparent K m ATP was 140 ± 28 μM. Measurements using a fixed ATP concentration and MA levels that ranged from 0.005 to 2.5 mM indicated that the apparent K m MA was 90 ± 18 μM. K m values for MA were comparable to those previously described in archaea, bacteria and eukaryotes (Table 1).

Inhibition of AaMVK by long-chain isoprenoids
The sensitivity of AaMVK towards several phosphorylated isoprenoids is shown in Fig 4 and Table 1. AaMVK activity was strongly inhibited by long chain isoprenoids. Our results demonstrated that GGPP, FPP and GPP are competitive inhibitors for the binding of ATP to AaMVK. Their inhibitory capacities (Ki) were: GGPP (0.93 ± 0.19 μM), FPP (0.44 ± 0.2 μM) and GPP (0.55 ± 0.28 μM) (S5 Fig). Short chain isoprenoids, such as DMAPP and IPP inhibited only in the micromolar range, with a K i value greater than 10 μM; while C 6 compounds, such as PM and DPM, were not inhibitory.

AaMVK activity in extracts of mosquito thoraces
To further assess the inhibitory feedback of isoprenoid on AaMVK, we analyzed the MVK activity in homogenates of mosquito thoraces that contained the CA (24h old sugar-fed females), in the presence of 100 μM FPP. Addition of FPP resulted in a 25% reduction of MVK activity ( Fig  5A). When we analyzed the reactions products from the activity of enzymatic extracts from mosquito thoraces on MA, we observed a significant increase in isopentenyl pyrophosphate (IPP) concentration. Since MVK, phosphomevalonate kinase and mevalonate diphosphate decarboxylase share similar reaction conditions [34], these results suggested that in our in vitro assay using mosquito homogenates the catalytic transformation of MA into PM continue to generate IPP via diphosphomevalonate (DPM). Changes in the IPP concentration were therefore used as a proxy to study the effect of FPP on MVK activity in mosquito thoraces homogenates. An inhibitory effect of FPP on MVK was confirmed, with the levels of IPP significantly reduced when FPP was added to the extracts, compared to those that were not exposed to FPP (Fig 5B).

Discussion
The JHs are synthesized through the MVAP, an ancient metabolic pathway present in the three domains of life [7]. The MVAP consists of a main trunk followed by sub-branches that generate a diverse range of essential biomolecules required for cell signaling, membrane integrity, energy homeostasis, protein prenylation and glycosylation [35][36][37][38]. Insects lack the cholesterol-synthetic branch present in vertebrates [39], but in the CA the MVAP branches into the synthesis of JH.
Mevalonate kinase is one of three consecutive ATP-dependent enzymes in the MVAP. The primary structure of AaMVK revealed all the characteristic GHMP kinase domains (Pfam: 00288 and Pfam08544), as well as the key amino acids involved in substrate binding and catalytic activity (Lys 14 , Ser 159 , Glu 208 and Asp 219 ) [16,40]. The structural analysis confirmed that these   residues are situated at the active site [5,17], suggesting that catalysis in insect is also mediated by a base mechanism [5], in which Asp 219 makes a salt bridge with Lys 14 , with the penta-coordinated γ-phosphate transition state stabilized by Mg 2+ and the amino acids Glu 208 , Ser 159 and Lys 14 . The residue Asp 219 acts as a general base, abstracting a proton from the hydroxyl group of MA, therefore converting MA into an excellent nucleophile that attacks the γ-phosphorus of ATP. The residue Lys 14 is believed to maintain the aspartate residue in the deprotonated state to facilitate the proton transfer. Several insertions (loops) lying on the surface of the globular structure were identified in the AaMVK. These loops do not contain any catalytic amino acid, but as it has been suggested that might be important conferring protein stability [5]. Although disulfide bridges confer thermostability to prokaryote MVKs [41], the absence of disulfide bridges is another important feature that the AaMVK shares with other eukaryotic MVKs. Like many ATP-dependent reactions, AaMVK requires divalent metal cations for catalysis. The function of the divalent metal cation is to anchor the diphosphate moieties and to facilitate ionization of allylic substrates [18]. Although in insects it seems that the essential cation in vivo is Mg 2+ , our results shown that AaMVK can replace Mg 2+ in vitro by other divalent cations such as Mn 2+ and Co 2+ . Similarly, in the process of phosphorylation, although other nucleotide triphosphates including GTP, CTP and TTP can partially substitute for ATP as phosphoryl donors in vitro, most likely ATP is the preferential in vivo phosphoryl donor. The results for the analysis of cofactor requirements, phosphoryl source and optimal pH of AaMVK were in agreement with those described for previously characterized MVKs [42,43]. Our kinetic studies revealed that the V max for the formation of PM in mosquitoes was comparable to that described for other MVK's, ranging from 12 to 50 μmol min -1 mg -1 ( Table 1). The AaMVK Michaelis-Menten constants for mevalonate (Km MA ) and ATP (Km ATP ) were in the range of those previously described for other MVKs (Table 1), which also have higher affinity for MA than ATP. Comparisons of kinetics between purified and recombinant enzymes are not always straightforward; much more difficult is to compare their activities with those of crude extracts. Discrepancies between the kinetic properties of purified and recombinant enzymes from the same species have been reported for other MVP enzymes [44]. Conclusive evidence linking the activity of the recombinant AaMVK with the activity detected in extracts is missing; but the fact that AaMVK is a highly conserved protein, encoded by a single annotated gene in the A. aegypti genome, and with transcripts enriched in the CA [20], suggest that both activities correspond to the same protein.
AaMVK mRNA expression levels in the CA are concurrent with JH biosynthesis titers in female mosquitoes [20,31]. MVK transcripts in Bombyx mori also correlate with JH synthesis [19], suggesting an important role of this enzyme in the regulation of the JH pathway. The mevalonate pathway is subject to multivalent transcriptional and post-transcriptional regulation, primarily at the level of HMG-CoA reductase [35]; however, it is becoming increasingly clear that regulation of MVK catalysis plays also an important modulatory role. A regulatory mechanism for controlling MVK activity is feedback inhibition by the presence of isoprenoids [17,42,45]. The competitive inhibition results from the interaction of the isoprenoid binding site of the phosphoryl group of ATP [17].
AaMVK is a class I enzyme, exhibiting efficient inhibition by GPP and FPP (Ki less than 1 μM), and none by IPP and DPPM. It is interesting that the two products of a single enzyme (FPP synthase) are specific in their inhibition of AaMVK. The possibility that GPP and FPP act as physiological inhibitors (in vivo) in the synthesis of JH in mosquito is strengthened when considering the inhibition exerted by these two metabolites on the MVK activity present in crude extracts of female mosquito thoraces containing the CA. However, further analysis are required to evaluate their significance as regulators in vivo.
The activity of AaMVK in the CA of female mosquito shows dramatic changes during the gonotrophic cycle that correlate well with changes in JH biosynthesis [34]. AaMVK activity is very low in newly emerged adult females (30 fmol/CA/h), it increases more than 200 folds by 12 h after adult eclosion (4500 fmol/CA/h), and markedly decreases with the decline of JH synthesis by 24 h after blood feeding (20 fmol/CA/h) [34]. Although rate limiting bottlenecks have been proposed at single specific steps in both the MVAP and JH-branch in the CA of different insects, our previous studies suggested that there are multiple regulatory points and they change in different physiological stages [34,46]. Further studies will be necessary to reveal if AaMVK plays a key role restricting the flux into JH III at specific physiological conditions. Two different buffers were used to generate the pH gradient: MES at pH 5.5 to 7 and Tris-HCl at pH 7 to 9. The optimum pH was found to be 7.5 to 8.0; with the enzyme exhibiting 60-70% of its optimum activity over a rather broad pH range (7 to 8.5). Activities are expressed as μmol of product produced by min per mg of enzyme. Each value represents the means ± S.E. of three replicate assays.