Erythronolide synthase

The active site of this enzyme has a very broad specificity, which allows for the synthesis of long chains of carbon atoms by joining, via a thioester linkage, small organic acids, such as acetic and malonic acid.[4] The KS domain receives the growing polyketide chain from the upstream module and subsequently catalyzes formation of the C-C bond between this substrate and an ACP-bound extender unit that is selected by the AT domain.[5]

Each AT domain has an α-carboxylated CoA thioester (i.e. methylmalonyl-CoA) This specificity prevents non-essential addition of enzymes within the module. The AT captures a nucleophilic β-carboxyacyl-CoA extender unit and transfers it to the phosphopantetheine arm of the ACP domain.[6]

Functions via catalyzing acyl transfer from methylmalonyl-CoA to the ACP domain within the same module via a covalent acyl-AT intermediate. The importance of the AT to the stringent incorporation of specific extender unit in the synthesis of polyketide building blocks makes it vital that the mechanism and structure of these domains be well-elucidated in order to develop efficient strategies for the regiospecific engineering of extender unit incorporation in polyketide biosynthesis.[5]

The ACP is not substrate specific, which allows for the interaction with every domain present within its module. This protein collaborates with the ketosynthase (KS) domain of the same module to catalyze polyketide chain elongation, and subsequently engages with the KS domain of the next module to facilitate forward chain transfer.[7] The ACP first accepts the extender unit from the AT, then collaborates with the KS domain in chain elongation, and finally anchors the newly elongated chain as it undergoes modification at the β-keto position. In order to carry out their function, the ACP domains require post-translational addition of a phosphopantetheine group to a conserved serine residue of the ACP. The terminal sulfhydryl group of the phosphopantetheine is the site of attachment of the growing polyketide chain.[8]

Located at the C-terminus site of the furthest downstream module. It is terminated in a thioesterase, which releases the mature polyketide (either as the free acid or a cyclized product), via lactonization.[9]

Note: As stated above, the first module of DEBS contains an additional acyltransferase and ACP for initiation of the reactions

Additional components, may have any one or a combination of the following:

Ketoreductase- Uses NADPH to stereospecifically reduce it to a hydroxyl group[10]

Dehydratase- Catalyzes the removal of the hydroxyl group to create a double bond from organic compounds in the form of water

Enolreductase- Utilizes NADPH to reduce the double bond from the organic compound

The first reported instance of genetic engineering of DEBS came in 1991 from the Katz group[14] who deleted the activity of the KR in module 5 of DEBS which produced a 5-keto macrolide instead of the usual 5-hydroxy macrolide. Since then, deletion or inactivation (often via introduction of point mutations) of many active sites to skip reduction and/or dehydration reactions have been created. Such modifications target the various KR, DH, ER active sites seen on different modules in DEBS. In fact, whole modules can be deleted in order to reduce the chain-length of the polyketides and alter the cycle of reduction/dehydration normally seen.[13]

In one of the first reorganizations of DEBS, a copy of the terminal TE was placed at the end of each module in separate trials, which as predicted resulted in the cleavage and release of the correspondingly shortened products.[14] Following this, ever more complex methods were devised for the addition or substitution of single or multiple active sites to the DEBS complex.

The most common method of engineering DEBS as of 2005 is AT substitution, in which the native AT domain is replaced with an AT specific for a different primer or extender molecule.[12] Under normal circumstances, DEBS has a “loading” or priming AT specific for predominantly propionyl-CoA while all six subsequent AT are specific for the extender molecule, methylmalonyl-CoA. The native AT of DEBS have all been successfully substituted with AT from other modular PKS such as the PKS that produces rapamycin; which replaces the methylmalonyl-CoA specific AT with malonyl-CoA AT and produces a non-methylated erythromycin derivative.[12] This mode of engineering in particular shows the versatility that can be achieved as both the priming molecule and the extender molecule can be changed to produce many new products. In addition to the AT sites, any of the reductive/dehydrating enzyme active sites may be replaced with one or more additional reductive/dehydrating enzyme active sites. For example, in one study, the KR of module 2 of DEBS was replaced by a full set of reductive domains (DH, ER and KR) derived from module 1 of the rapamycin PKS as shown in Figure 2 FIGURE 2

There is at least one report of a whole module substitution, in which module 2 of DEBS was replaced with module 5 of the rapamycin PKS[15] The activities of the two modules is identical, and the same erythromycin precursor (6-deoxyerythronolide B) was produced by the chimeric PKS; however, this shows the possibility of creating PKS with modules from two or even several different PKS in order to produce a multitude of new products. There is one problem with connecting heterologous modules though; there is recent evidence that the amino acid sequence between the ACP domain and the subsequent KS domain of downstream modules plays an important role in the transfer of the growing polyketide from one module to another.[15] These regions have been labeled as “linkers” and although they have no direct catalytic role, any substitution of a linker region that is not structurally compatible with the wild-type PKS may cause poor yields of the expected product.

Using a semi-synthetic approach, a diketide intermediate may be added either in vitro or in vivo to a DEBS complex in which the activity of the first KS has been deleted.[14] This means that the diketide will load onto the second KS (in module 2 of DEBS) and be processed all the way to the end as normal. It has been shown that this second KS is fairly nonspecific and a large variety of synthetic diketides can be accepted and subsequently fully elongated and released. However, it has also been seen that this KS is not highly tolerant of structural changes at the C2 and C3 positions, especially if the stereochemistry is altered.[14] To date, this has been the most successful approach to making macrolides with potency equal to or greater than erythromycin.[16]

In modular PKS, KR active sites catalyze stereospecific reduction of polyketides. Inversion of an alcohol stereocenter to the opposite stereoisomer is possible via replacement of a wild-type KR with a KR of the opposite specificity.[13] This has rarely been done successfully, and only at the terminal KR of the DEBS complex. It has been theorized that changing the stereospecificity of a KR in an earlier module would also require the concurrent modification of all downstream KS.[12]

Recent studies of the amino acid sequence of the two types of stereospecificity in KR have determined a perfect correlation with these residues and the predicted stereochemical outcome.[12] This is particularly useful in situations where the gene sequence of a modular PKS is known but the final product structure has not yet been elucidated.

Enzymes that act on the macrolide after it has been released and cyclized by DEBS are called tailoring enzymes. Many such enzymes are involved in the production of erythromycin from the final product of unmodified DEBS, 6-deoxyerythronolide B. Such classes of enzymes include mainly oxidoreductases and glycosyl transferases and are essential for the antibiotic activity of erythromycin.[12][14][17]

Thus far, few attempts have been made to modify tailoring pathways, however, the enzymes which participate in such pathways are currently being characterized and are of great interest. Studies are facilitated by their respective genes being located adjacent to the PKS genes, and many are therefore readily identifiable.[17] There is no doubt that in the future, alteration of tailoring enzymes could produce many new and effective antimicrobials.