Pyruvate produced from the CBB cycle is converted to the key intermediate a-KIV by the BCAA biosynthetic pathway. AHAS catalyzes the first step in the biosynthe­sis of all three BCAAs (Fig. 4). It is capable of synthesizing (2S)-acetolactate, a precursor of valine and leucine from two molecules of pyruvate and synthesizing

(2S)-2-aceto-2-hydroxybutyrate, a precursor of isoleucine, from pyruvate and 2-keto-
butyrate. In most organisms, a single AHAS catalyzes both of the above-mentioned reactions, whereas in other organisms these reactions are catalyzed by separate enzymes [28]. The AHAS enzyme consists of two subunits, one being catalytic and one playing a regulatory role. No crystal structure with both subunits has been obtained to date, although individual subunits have been crystallized separately. Catalysis is believed to occur at the subunit interface, since the catalytic subunit alone has little to no activity [70]. Expression of AHAS is controlled by the amount of tRNABCAAs available in the cell. High levels of tRNABCAAs repress the transcription of AHAS. Additionally, AHAS activity is controlled allosterically at the activity level by its regulatory subunit, through binding of valine at the homodimer interface [71]. Combined site-directed mutagenesis and BCAA-binding studies have shown that valine binding at the homodimer interface potentially causes a conformational change resulting in a less stable complex with decreased enzymatic activity [72].

A promising approach to counter valine inhibition of AHAS can be found in

E. coli. E. coli has three AHAS isozymes, each with a different substrate specificity and regulation mechanism. The sequence of E. coli AHAS isozyme II differs from other AHAS, and its regulatory subunit is insensitive to direct feedback inhibition by valine [70]. R. eutropha AHAS (IlvBH in Fig. 1) shares most sequence similarity with E. coli AHAS isozyme III and is also subject to allosteric feedback inhibition by pathway intermediates (dihydroxyisovalerate and ketoisovalerate; Sinskey labo­ratory, unpublished data) and end products (valine, leucine and isoleucine; Sinskey laboratory, unpublished data). Thus, minimizing allosteric inhibition by products and intermediates is essential to optimize IBT production in R. eutropha. N-terminal amino acid residues that are conserved in all valine-sensitive AHAS could contrib­ute to the binding of valine or other BCAAs that cause allosteric feedback inhibi­tion. These residues can be mutated to the ones that are present in the valine-insensitive AHAS isozyme II from E. coli. C-terminal truncation studies on the valine-sensitive E. coli AHAS isozyme III show decreased valine inhibition. However, the exact mechanism of inhibition alleviation is unknown, since valine and other BCAAs are hypothesized to bind only at the N-terminus of the regulatory subunit [73] .

An AHAS enzyme’s selectivity (R) for aceto-2-hydroxybutyrate production over acetolactate production can be calculated by the following equation:

[AHB]/[2KB]

[AL]/[P]

where AHB, 2KB, AL, and P represent aceto-2-hydroxybutyrate, 2-ketobutyrate, acetolactate, and pyruvate, respectively. R values for E. coli AHAS isozymes I, II, III, and C. glutamicum AHAS are 2.0, 65, 40, and 20, respectively, all of which favor the formation of AHB over acetolactate. R. eutropha AHAS has R value of ~45 [28].

As actetolactate is a precursor of IBT, reducing the AHAS R value would help direct carbon flow towards a-KIV, and consequently IBT. Previous mutagenesis studies on the E. coli AHAS II catalytic subunit revealed a tenfold reduction in R when a tryptophan residue at position 464 was mutated to lysine, glutamine, or tyrosine [74]. It is suggested that the indole ring of tryptophan interacts with the
extra methyl group on 2-ketobutyrate and stabilizes it in the active site. Site-directed mutagenesis could also be used to decrease the R value for R. eutropha AH AS.

KARI, encoded by ilvC in R. eutropha, catalyzes the formation of 2,3- dihydroxyisovalerate from 2-acetolactate and the formation of 2,3-dihydroxy-3- methylvalerate from 2-aceto-2-hydroxybutyrate (Fig. 4). KARI has similar substrate preference towards both substrates. A unique feature of its reaction mechanism is that it simultaneously catalyzes both an isomerization and a reduction reaction. Mutations in active site residues that abolished the reductase activity also elimi­nated the isomerization reaction, suggesting that isomerization and reduction are coupled without any intermediate. KARI requires NADPH and a divalent metal ion, in most cases Mg2+, for catalysis. The metal cofactor is involved in the alkyl migra­tion isomerization step, whereas NADPH is the electron donor for the reduction step [28]. Since KARI has no substrate bias, simple overexpression of ilvC is likely sufficient to provide ample precursor amounts for the production of IBT.

DHAD catalyzes the formation of ketoacids from the products of KARI. The mechanism of action is unknown, but it likely involves the dehydration of vicinal diols to ketoacids via an enol intermediate. E. coli’s oxygen sensitive DHAD con­tains a [4Fe-4S]2+ cluster. The reaction mechanism is proposed to be similar to that of aconitase in the TCA cycle, which also involves an FeS cluster [28]. The activity, feedback inhibition and oxygen sensitivity of R. eutropha DHAD have not been studied. As shown in Figs. 1 and 4, the combined activities of DHAD, KARI, and AHAS convert pyruvate into the key IBT intermediate 2-KIV.