Chapter 38

Biosynthesis of Isoprenoids in Bacteria

ROBERT H. WHITE

INTRODUCTION

Much attention and effort have been focused on establishing the pathways leading to the biosynthesis of ubiquinone and menaquinone, the major isoprenoid-containing compounds in Escherichia coli and Salmonella typhimurium (official designation, Salmonella enterica serovar Typhimurium); however, little work has been focused on the biosynthetic pathways leading to the generation of isoprenoid side chains present in these molecules. This lack of interest likely resulted from the fact that the pathway to these isoprenoid side chains in eucarya (animals and yeasts) (5, 6, 22) and archaea (10, 19, 20) was very well established; it was assumed that the same pathway also occurred in bacteria. Basically, this pathway involves the condensation of three molecules of acetate to form mevalonate, which is then converted into isoprenyl pyrophosphate (IPP). The IPP is then polymerized to various isoprenoids, some of which are condensed (as the pyrophosphate esters) with a series of aromatic compounds to form ubiquinones and menaquinones. A number of observations over the years, however, have indicated that the "established" early steps in the biosynthesis of isoprenoids, i.e., the steps leading to the formation of IPP, are different from those involved in the biosynthesis of IPP in bacteria such as E. coli. In this chapter, I shall discuss these observations as well as our current knowledge of the biosynthesis of isoprenoids in E. coli.

BIOSYNTHESIS OF IPP

The first indication that the established steps in the pathway leading to IPP were not operating in E. coli, as well as several other bacteria, was published in 1965 (23). In this work, it was demonstrated that despite the fact that [2-¹⁴C]mevalonate and [1-¹⁴C]acetate were each readily incorporated into fatty acids by growing E. coli cells, these labeled precursors were not incorporated into ubiquinones. Thus, two of the central intermediates in the biosynthesis of IPP, acetate and mevalonate, were metabolized by E. coli but were not incorporated into IPP. Apparently this observation was not reconsidered until 1981, at which time, on the basis of the incorporation of [1-¹⁴C]acetate and [2-¹⁴C]acetate, another pathway for the biosynthesis of isoprenoids in E. coli was proposed (21). In this work, Pandian et al. used a chemical degradation scheme to establish specifically at which carbons of the ubiquinone side chain the labeled acetates were incorporated. The results of this work showed that both C-1 and C-2 of acetate were incorporated almost exclusively into C-1 of the presumed IPP unit used in the biosynthesis of ubiquinone. (This position of incorporation was inferred from the measured incorporation at C-2 of levulinic acid derived from the oxidative cleavage of the ubiquinone.) Since these data contradicted the established pathway, a new pathway termed the acetolactic pathway was proposed (21). One feature of this proposed pathway was the involvement of α-ketoisovaleric acid and α-ketoisocaproic acid as key intermediates. The role of these α-keto acids in the pathway, and thus in isoprenoid biosynthesis, could be easily tested by generating these acids in vivo and by measuring their incorporation into ubiquinone. Since the conversion of α-ketoisovaleric acid to α-ketoisocaproic acid is required for the conversion of valine to leucine in E. coli (24), cells converting labeled valine into labeled leucine would have to contain both of these labeled α-keto acids. Thus, by growing E. coli with [methyl-²H₆]valine and measuring the extent of labeling in the valine, leucine, and ubiquinone-8 in the cells, the involvement of these α-keto acids in isoprenoid biosynthesis could be established. The results of these experiments (30) showed that 76 and 81%, respectively, of the cellular valine and leucine were labeled with ²H₆, whereas no ²H was found to be incorporated into ubiquinone-8. This result, therefore, eliminated the possibility of the acetolactic acid pathway being involved in isoprenoid biosynthesis in E. coli.

Recent work by Rohmer et al. (25), measuring by ¹³C nuclear magnetic resonance spectroscopy the incorporation of ¹³C-labeled glucose, acetate, pyruvate, and erythrose, has enabled these authors to determine the biosynthetic origin of all carbons of the isoprenyl side chain of ubiquinone biosynthesized by E. coli as well as by several other bacteria. Data on the incorporation of [1-¹³C]acetate, which are not in agreement with those of Pandian et al. (21), show that C-1 of acetate is incorporated only at C-4 of IPP and that C-2 of acetate is incorporated equally into all of the carbons of IPP except C-4, which showed only one-half the incorporation found in the other positions. (The discrepancy in labeling patterns reported by these two groups may be attributed to the different growth conditions used. In the experiments of Rohmer et al. [25], acetate was the sole carbon source, whereas in the experiments of Pandian et al. [21], glucose was the principal carbon source.) These observed labeling patterns are both also inconsistent with the known positions of incorporation of labeled acetates into IPP in the established mevalonate pathway, where C-1 of acetate labels C-1 and C-3 of IPP and C-2 of acetate labels C-2, C-4, and C-5 of IPP.

In other experiments reported by Rohmer et al. using [1-¹³C]glucose and [6-¹³C]glucoses as precursors, the data showed each of these labeled glucose carbons to be incorporated equally well into C-1 and C-5 of the IPP precursor to the isoprenoids. Again, these results are completely inconsistent with the established pathway, where the ¹³C label of the glucoses would have been incorporated into C-2, C-4, and C-5 of IPP. These data and data on the incorporation of multiply labeled glucoses into ubiquinone by the other bacteria studied, as well as the data of Zhou and White (30) on the incorporation of [U-¹³C]glucose into ubiquinone by E. coli, demonstrate that C-1, C-2, and C-4 of IPP are derived as a unit from C-6, C-5, and C-4 of glucose and that C-3 and C-5 of the IPP are derived as a unit from C-5 and C-6 of glucose.

On the basis of this information, Rohmer et al. (25) proposed that the biosynthetic origin of the C₅ unit in IPP proceeds much like the biosynthesis of valine(4). Addition of a thiamine-activated acetaldehyde, derived from pyruvate, to the C-2 carbonyl group of dihydroxyacetone phosphate (or a similar triosephosphate) is followed by a migration of the hydroxymethyl group from the C-2 to the C-3 position, generating the carbon core structure of the isoprenyl unit. The core structure resulting from this initial condensation is then processed through an unknown series of reductions, isomerizations, eliminations, and/or phosphorylations to form IPP. Another possible scheme for a series of reactions leading to IPP is shown in Fig. 1. In this scheme, monohydroxyacetone phosphate and pyruvate condense, with the loss of CO_2, to produce the first intermediate in the pathway. It should be noted that in this scheme the monohydroxyacetone phosphate would have to be labeled in the reverse manner from that expected had it arisen directly from glycerol phosphate. Monohydroxyacetone phosphate with the required labeling pattern could be produced from methylglyoxal, a known metabolite of dihydroxyacetone phosphate in E. coli (9), by reduction of the C-1 aldehyde of methylglyoxal to an alcohol and subsequent phosphorylation.

Fig. 1One of several possible pathways for the biosynthesis of IPP from non-acetate-derived precursors. Each reaction is numbered and is catalyzed by the following enzymes or processes: 1, a thiamine-dependent decarboxylation of pyruvate followed by the condensation of the stabilized anion of 2-?-hydroxyethylthiamine pyrophosphate with monohydroxyacetone phosphate (this reaction is analogous to that used in the production of ?-acetolactate by acetolactate synthetase during the biosynthesis of valine); 2, carbon chain rearrangement (pinacol rearrangement) analogous to that catalyzed by acetohydroxy acid isomeroreductase of valine biosynthesis; 3, elimination of water in a manner analogous to that in the conversion of lactic acid to acrylic acid in the acrylic acid pathway (18), the conversion of 2-hydroxyglutaryl-coenzyme A (CoA) to glutaconyl-CoA (17), and the dehydration of 2-hydroxybutyryl-CoA to crotonyl-CoA (16); 4, reduction of a conjugated double bond in a manner analogous to that of enoyl-acyl carrier protein reductase in fatty acid biosynthesis or the bacterial reduction of acryloyl-CoA or ethyl vinyl ketone by acryloyl-CoA reductase (26); 5, NADPH-dependent reduction of a carbonyl; 6, elimination of water; 7, phosphorylation of a monophosphate ester to a pyrophosphate ester; 8, isomerization of isopentenyl pyrophosphate. The symbols on the compounds indicate the carbons of glucose from which the indicated carbon must be derived. ?, carbons 1 or 6; , carbons 2 or 5; and , carbons 3 or 4.

Source:

This recently discovered pathway to IPP in E. coli as well as in many other bacteria means that most bacteria probably do not contain the enzymes that are used in the established pathway for IPP production. The absence of these enzymes thus explains why many bacteria, even bacteria that can use mevalonate as their sole carbon source (28), are unable to incorporate mevalonate into the isoprenoids that they produce (13, 23). The often quoted statement that the most highly regulated enzyme of nature, hydroxymethylglutaryl coenzyme A reductase, is "universally distributed through-out nature" apparently is not correct (7). This enzyme may have just regulated itself out of existence in the competitive bacterial world.

POLYMERIZATION OF IPP TO POLYISOPRENOIDS

In contrast to our knowledge about IPP formation, our understanding about the sequence of reactions in E. coli leading from IPP to the polyisoprenyl phosphates is on a much firmer ground, largely as a result of the work of Fujisaki et al. (11, 12, 13). Initial work reported by Fujisaki et al. in 1984 (11) showed that lyophilized E. coli cells rehydrated with phosphate buffer containing ¹⁴C-IPP were able to incorporate ¹⁴C-IPP into various isoprenoids. They later showed these isoprenoids to be ubiquinone-8, demethylmenaquinone-8, and the phosphate ester of all-trans-octaprenol and cis,trans-polyprenyls. Cells incubated in like manner with buffer containing geranyl pyrophosphate (GPP) or farnesyl pyrophosphate in combination with ¹⁴C-IPP produced these same compounds containing even higher levels of radioactivity. These data are consistent with the established pathway of polyisoprenoid biosynthesis shown in Fig. 2. In this pathway, a portion of the IPP is isomerized to dimethylallyl pyrophosphate (DPP), which then condenses with another molecule of IPP to form the C₁₀ compound 9, GPP. The resulting GPP then combines with another IPP via the same basic reaction to form the C₁₅ compound 10, trans-farnesyl pyrophosphate (FPP) (Fig. 2). The FPP then serves as the starting point for the biosynthesis of compound 11, trans-octaprenyl pyrophosphate (OPP), which is used in the generation of ubiquinone, and the cis,trans-polyprenyls (the C₄₅, C₅₀, and C₅₅ compounds), which are the lipid carriers. In the synthesis of the octaprenyl compound, individual IPP molecules are inserted in the growing molecule with trans regiochemistry about the generated double bond, whereas in the case of the cis,trans-polyprenyls, a cis regiochemistry is generated (15).

Fig. 2Transformation of IPP to polyisoprenoids in E. coli. Each reaction is identified by an italic numeral and is catalyzed by one of the following enzymes: 7, isopentenyl pyrophosphate isomerase; 8, farnesyl pyrophosphate synthetase; 9, farnesyl pyrophosphate synthetase; 10, octaprenyl pyrophosphate synthetase; 11, 4-hydroxybenzoate octaprenyltransferase; 12, undecaprenyl pyrophosphate synthetase.

Source:

Proof that these reactions are in fact responsible for the biosynthesis of all of the isoprenyl-containing compounds in E. coli came from a study of the specificity and reaction products of four enzymes isolated from E. coli that are involved in isoprenyl biosynthesis (12). These enzymes are IPP isomerase, FPP synthetase, OPP synthetase, and undecaprenyl pyrophosphate (UPP) synthetase (Fig. 2). The FPP synthetase catalyzed the condensation of IPP with DPP as well as with GPP to yield FPP as the final product. OPP synthetase catalyzed the condensation of IPP with FPP to yield OPP, as shown in the top portion of Fig. 2. UPP synthetase catalyzed the condensation of IPP with FPP to yield cis,trans-polyprenyl pyrophosphates, as shown in the bottom portion of Fig. 2. The OPP synthetase from E. coli has been shown to be radiolabeled with the photolabile analog of FPP, [³H]2-diazo-3-trifluoropropionyloxy geranyl pyrophosphate, in the presence of IPP and a divalent cation (3). The photolabeled protein migrated on sodium dodecyl sulfate-polyacrylamide gel electrophoresis at a molecular weight of approximately 30,000. The UPP synthetase enzyme, in addition to requiring Mg²⁺ or Mn²⁺, also requires Triton X-100 for activity, as has been observed for UPP synthetases isolated from several other bacteria (1, 2, 27). Thus, these four enzymes would be the only ones required to generate, from IPP, all of the isoprenoids found in E. coli. The latter two enzymes, OPP synthetase and UPP synthetase, have also been isolated from Salmonella newington (8). These same enzymes also appear to be responsible for the biosynthesis of isoprenoids in all bacteria examined to date (see references 56 to 76 of reference 25).

FUTURE DIRECTIONS

From the information presented above, it is clear that two aspects of isoprenoid biosynthesis in E. coli must be studied in order to bring our understanding of this pathway to the same level as our understanding of the other biosynthetic pathways in E. coli: (i) the sequences of reactions leading to IPP and (ii) the genes responsible for the individual reactions. Since the present data indicate that the starting intermediates in the E. coli pathway to IPP are central metabolites, it will be difficult to define the pathway because it is very hard to obtain mutants that lack the required central metabolites. It may, however, be possible to determine the chemical nature of one or more of the intermediates leading to IPP by identifying the chemical structures of secreted phosphorylated compounds produced by E. coli mutants that require IPP or DPP for growth. In addition, as has been found with other biosynthetic pathways, genes for IPP biosynthesis may be clustered with the other genes involved in isoprenoid biosynthesis. Thus, the mapping of the gene (ispA) for the first enzyme involved in isoprenoid polymerization in E. coli (FPP synthetase) may be useful in this regard (14).

It is interesting to note that since the archaea appear to be specific relatives of the eucarya and both produce IPP from acetate, whereas bacteria produce IPP from sugar precursors, then this shift in pathway utilization may have occurred during the first branch point from the universal ancestor (29).