To overcome the inherent problems associated with use of S-prenylated peptides and proteins as Icmt substrates (expense in production, unwieldy separation techniques, etc), we synthesized a small molecule substrate for the enzyme that contains an appended biotin moiety (Fig. 1A) to facilitate the separation required for product analysis. The coupling of biotin moiety to the free amino group of the S-farnesylcysteine, FC, was readily achieved through the use of commercial biotin N-hydroxysuccinimide ester and the product BFC could be readily purified by precipitation or by reverse-phase chromatography (see Methods). BFC was characterized in terms of its ability to serve as a substrate for recombinant human Icmt using a standard in vitro methylation assay (Fig. 1B). The apparent K_m of BFC was 2.1 ± 0.4 μM, which is essentially identical to the K_m of 2.1 μM determined for farnesylated, Rce1-proteolyzed K-Ras (data not shown). Hence, BFC is a comparable to K-Ras as an Icmt substrate and has the added advantages of being a stable small molecule that can be readily captured on avidin resins.

Detailed knowledge of the kinetic mechanism of an enzyme is very important in providing the foundation for analysis of structure-function studies and in inhibitor design studies. For a two substrate enzyme reaction such as that catalyzed by Icmt, there are three basic mechanisms: i) random sequential, ii) ordered sequential, and iii) ping-pong 2829. Initial velocity studies can be used to distinguish between a ping-pong mechanism, in which a product is released between the addition of the two substrates, and the sequential mechanisms, in which both substrates must bind before any product is released. Hence, in the first series of studies, the initial velocity of the Icmt reaction at different substrates concentrations was determined. In a ping-pong mechanism for an enzyme with substrates A and B, the plots of 1/v versus 1/[A] at different fixed [B], and vice versa, should yield parallel lines 28. In a sequential mechanism, either ordered or random, the family of reciprocal plots should intersect above the X axis (if the binding of A increase the interaction with B), below the X-axis (if the binding of A decreases the interaction with B), or on the X-axis (if the binding of A has no effect on the interaction of B) 28. The data obtained in the experiments where AdoMet concentration was varied at a series of fixed BFC concentrations is shown in Fig 2A, and Fig. 2B shows the data obtained when BFC concentration is varied at fixed AdoMet concentrations. These results show that, in each case, the families of reciprocal plots are not parallel and intersect above the X-axis, ruling out the possibility of a ping-pong mechanism. Moreover, the data for the experiments in which BFC concentration was varied at fixed AdoMet concentrations show an intersection of the reciprocal plots on Y-axis (Fig. 2B), indicating that the reaction behaves as rapid-equilibrium bireactant system 28.

While the initial velocity studies described above can distinguish between a ping-pong mechanism and a sequential mechanism, they cannot be used to distinguish an ordered from a random mechanism. These two types of sequential mechanisms can be distinguished, however, through the use of so-called "dead-end" substrates that cannot go on to form products 30. Hence, we employed such a dead-end inhibitor of Icmt, farnesylthioacetic acid (FTA) 31, in kinetic studies. To distinguish between the two types of inhibition patterns this compound was examined for its inhibitory properties with respect to both the cognate (BFC) and noncognate (AdoMet) substrates under conditions where the fixed substrate was present at nonsaturating concentrations. The results obtained demonstrate that the FTA is a competitive inhibitor with respect to BFC, with a measured K_i of 1.2 ± 0.2 μM (Fig, 3A, see also Table I). In contrast, the experiments performed with varying AdoMet concentrations show a family of parallel lines (Fig. 3B), which is a characteristic profile for an uncompetitive inhibition 2829. If FTA was able to combine with the free enzyme (i.e. the condition where its cognate substrate BFC binds first), the plots in Fig. 3B would have yielded lines that intersect at negative 1/v and 1/[AdoMet] values, i.e. FTA should have shown a mixed-type inhibitor with respect to AdoMet 20. These results indicate that Icmt reaction occurs through an ordered mechanism in which AdoMet binds first to the enzyme followed by BFC binding to the AdoMet-Icmt binary complex.

Two products are formed in the reaction catalyzed by Icmt, a methylated prenylcysteine (in vivo on proteins, but in vitro also on small molecules or peptides) and AdoHcy. To determine the order of the dissociation of these two products, we performed product inhibition using the prenylcysteine methylester, AFCME, and AdoHcy. Examination of the initial rates of product formation at fixed BFC and varying AdoMet concentrations in presence of three concentrations of AdoHcy (Fig. 4A), revealed that AdoHcy is a competitive inhibitor with respect to AdoMet with an apparent K_i of 3.5 ± 1.0 μM (Table I). This finding is consistent with both AdoMet and AdoHcy binding to the same form of Icmt, ie. the free enzyme, suggesting that AdoMet binds first in an ordered mechanism and that AdoHcy is the last product released. When the same experiment was performed at varying BFC concentrations, a very similar pattern was observed (Fig. 4B), revealing that AdoHcy is also a competitive inhibitor with respect to BFC with a K_i of 3.6 ± 1.0 μM. While a classic ordered sequential mechanism in which AdoHcy was released last would result in this type of inhibition being non-competitive or mixed-type, this type of behavior in which a single product binds competitively with both substrates can also be observed in ordered bireactant systems that proceed via a rapid equilibrium process 28.

The product inhibition experiments were next repeated with AFCME as the second class of product inhibitor. An ordered sequential mechanism with initial departure of the methylated prenylcysteine product requires that noncompetitive or mixed-type inhibition be observed with respect to both of the substrates of the reaction 29. The data obtained with L-AFC did indeed demonstrate precisely this 20. Noncompetitive inhibition by AFCME was observed with respect to the BFC substrate (Fig. 5A) with an apparent K_i of 1.9 ± 0.6 μM, and mixed-type inhibition was observed with respect to AdoMet as a substrate with an apparent K_i of 2.4 ± 0.7 μM, respectively (Fig. 5B). These kinetic results are summarized in Table I. In addition, essentially identical results were observed with a second prenylcysteine-type product analog N-acetyl-S-farnesyl-L-cysteine methyl amide (AFCMA) (data not shown). Altogether, the results this kinetic analysis are completely consistent with the reaction catalyzed by Icmt proceeding through an ordered sequential mechanism with the AdoMet substrate binding first and the AdoHcy product being released last.

Streptavidin-sepharose beads were purchased from Amersham, trans, trans-farnesyl bromide, biotin N-hydroxysuccinimide ester and S-(5'-adenosyl)-L-methionine p-toluenesulfonate were purchased from Sigma-Aldrich, L-cysteine was purchased from Novabiochem, [³H-methyl]-S-adenosyl-L-methionine was purchased from Perkin Elmer Life Sciences, farnesylthioacetic acid (FTA) was purchased from Biomol, S-(5'-adenosyl)-L-homocysteine was purchased from Fluka. The N-acetyl-S-farnesylcysteine methyl ester (AFCME) and N-acetyl-S-farnesylcysteine methyl amide (AFCMA) were synthesized as previously described (242526). Recombinant human Icmt was prepared by infection of Sf9 cells with a recombinant baculovirus containing the entire open reading frame of the human Icmt cDNA as described 8. The membrane fraction of the infected Sf9 cells, isolated as described for studies with the Rce1 protease 27, was used as the source of Icmt for all studies described.

S-Farnesyl L-cysteine (FC) was prepared by reaction of farnesyl bromide with L-cysteine in methanol/ammoniac solvent as described 26. The resultant FC product was separated from L-cysteine by extraction with butanol/H₂O (1:1); the butanol phase was then evaporated under reduce pressure and the FC product washed several times with hexane to remove residual farnesyl bromide.

Two coupling procedures were used to attach the biotin moiety to the amino group of FC; both involved the use of biotin N-hydroxysuccinimide methylester (biotin-NHS) as the biotinylation agent. In the first method, an excess of biotin-NHS (75 μmol) to FC (7.5 μmol) was used. The two compounds were dissolved in 2.3 ml of DMSO and 200 μl of 1M Hepes, pH 12, was added. Following a 2 h incubation at room temperature, the DMSO was evaporated under reduced pressure and the resulting residue extracted with a solution of butanol/H₂O (1:1). The butanol phase was dried and the residue dissolved in 100 μl of methanol; this solution was then diluted to 1 ml by addition of 0.1% trifluoroacetic acid (TFA). The addition of the TFA resulted in the appearance of a white precipitate of essentially pure BFC as judged by chromatography and NMR analysis.

The second coupling method used an excess of FC (2.5 mmol) compared to biotin-NHS (0.73 mmol); these two compounds were dissolved in 18 ml of DMSO. To this solution was added 2 ml of 1M Hepes, pH 12, and the coupling allowed to proceed for 2 h at room temperature. As with the first method, the DMSO was evaporated under reduced pressure, the resulting residue extracted with a solution of butanol/H₂O (1:1), and the butanol phase dried and the residue dissolved in methanol. Because of the excess of FC present with the BFC product, specific precipitation of the BFC product was unsuccessful. Instead, the BFC was purified by preparative HPLC on a C₁₈ column developed in CH₃CN/H₂O (2:3). On this matrix, BFC was well-resolved from FC. The peak fractions were collected and solvent evaporated under reduced pressure; the resulting product was white solid that was >95% pure as judged by analytical reverse-phase HPLC [C₁₈ matrix developed in CH₃CN/H₂O/TFA (90:10:0.1). Proton NMR spectra of the products obtained by either method were completely consistent with that expected for authentic BFC.

The assay developed for Icmt activity involved quantitation of [³H]methyl incorporation into the small molecule substrate BFC. For the standard assay, reactions were initiated by addition of Sf9 membranes containing Icmt (0.5 μg protein) to an assay mixture containing BFC (4 μM) and [³H]AdoMet (5 μM, 1.3 Ci/mmol) in100 mM Hepes, pH 7.4 and 5 mM MgCl₂ in a total volume of 45 μl. Reactions were carried out for 20 min at 37°C, whereupon they were terminated by addition of 5 μl of 10% Tween 20. Following termination, streptavidin beads (10 μl of packed beads suspended in 500 μl of 20 mM NaH₂PO₄, pH 7.4, contaning 150 mM NaCl) were added, and the mixtures mixed by gentle agitation overnight at 4°C. The beads were harvested by centrifugation in a tabletop microcentrifuge at 10,000 rpm for 5 min and washed 3 times with 500 μl of 20 mM NaH₂PO₄, pH 7.4, containing 150 mM NaCl. The beads were then suspended in 100 μl of the same buffer, transferred to scintillation vials, and radioactivity determined. For the kinetic analyses, the concentrations of substrates (AdoMet, BFC) or additional ligands (eg. AdoHcy, AFCME, etc) were varied as detailed in the legends to the appropriate figures.