Amides as Bioisosteres of Triazole-Based Geranylgeranyl Diphosphate Synthase Inhibitors

Abstract

Geranylgeranyl diphosphate synthase (GGDPS) inhibitors are of potential therapeutic interest as a consequence of their activity against the bone marrow cancer multiple myeloma. A series of bisphosphonates linked to an isoprenoid tail through an amide linkage has been prepared and tested for the ability to inhibit GGDPS in enzyme and cell-based assays. The amides were designed as analogues to triazole-based GGDPS inhibitors. Several of the new compounds show GGDPS inhibitory activity in both enzyme and cell assays, with potency dependent on chain length and olefin stereochemistry.

Graphical Abstract

1. Introduction

Multiple myeloma is a malignancy characterized by proliferation of clonal plasma cells that produce monoclonal protein. As a result of this activity, these cells are inherently susceptible to agents that disrupt protein homeostasis and trafficking pathways.^1–2 Our efforts to therapeutically exploit this potential vulnerability have focused on developing inhibitors of the enzyme geranylgeranyl diphosphate synthase (GGDPS),³ a key component of the isoprenoid biosynthesis pathway (IBP, Figure 1). Inhibition of this enzyme leads to a diminished pool of geranylgeranyl diphosphate (GGDP), which is the isoprenoid donor used in protein geranylgeranylation reactions. The Rab family of small GTPases require geranylgeranylation in order to function as regulators of intracellular trafficking events.⁴ Mutated Rab proteins that are not geranylgeranylated do not localize properly in the cell and are essentially non-functional.⁵ Our work has demonstrated that impairment of Rab geranylgeranylation, either as a result of depletion of GGDP or via direct inhibition of the Rab geranylgeranyl transferase, leads to disruption of intracellular monoclonal protein trafficking in myeloma cells.⁶ In turn, this disruption in protein trafficking leads to induction of the unfolded protein response and apoptosis.^6–7 While the two classes of clinically used IBP inhibitors (statins (HMG-CoA reductase inhibitors) and nitrogenous bisphosphonates (farnesyl diphosphate synthase (FDPS) inhibitors)) disrupt Rab geranylgeranylation in vitro,⁶ these agents are not ideal from an anti-myeloma therapeutic perspective given their off-target effects at high doses (statins^8–9) and limited systemic biodistribution (nitrogenous bisphosphonates¹⁰). We therefore have focused on the development of GGDPS inhibitors as a more clinically relevant means by which to selectively disrupt protein geranylgeranylation in vivo.^{7, 11}

Figure 1.

The mammalian isoprenoid biosynthesis pathway (IBP) with key enzymes and inhibitors.

During the course of our efforts to develop GGDPS inhibitors, we reported a series of diand monoalkyl isoprenoid bisphosphonates that selectively inhibit GGDPS over other enzymes in the IBP.^12–17 Subsequently, our efforts have focused on the development of triazole-based GGDPS inhibitors,^18–24 prepared using “click” chemistry²⁵ to introduce the isoprenoid unit. One benefit of the click strategy is that it allows reasonably efficient preparation of groups of compounds from a common intermediate. Structure-activity studies with triazole bisphosphonates have revealed that both chain length and olefin stereochemistry have significant impact on inhibitor potency.^{3, 18–20, 22} For example, those studies established that while triazoles bearing C₁₀ monoterpenoid chains were more active than either the C₅ or C₁₅ analogues, the homologated C₁₁ terpenoids were more active than either normal-length terpenoids or bishomologated terpenoids.^{18–20, 22} Given the predominance of trans isoprenoids in both protein prenylation²⁶ and plant metabolism,²⁷ it was surprising that the cis-olefin isomers were consistently more active than trans-olefin isomers.^{18, 20, 22} In fact, one of the most potent single isomer triazole-based GGDPS inhibitors is compound (1) that contains a homoneryl substituent i.e. a C₁₁ substituent with a cis-olefin (IC₅₀ of 75 nM in an enzyme assay with GGDPS).²⁰ Finally, the triazole moiety might interact favorably with the metal ions known to be held in the enzyme active site^{15, 20} but this has not been established rigorously.

One disadvantage of the click strategy is that when allylic azides are employed, they undergo a facile [3,3] sigmatropic rearrangement that isomerizes the alkene stereochemistry and affords a mixture of olefinic products.²⁸ We have developed a synthetic sequence that circumvents this issue through use of an epoxide as an olefin protecting group,²⁹ although this necessarily extends the synthetic sequence. The most direct synthesis of the homogeranyl/homoneryl compounds is based on addition of cyclopropylmagnesium bromide to a methyl ketone, and also generates a mixture of olefin isomers.¹⁹ Strategies to connect an isoprenoid chain to a bisphosphonate head group that do not afford mixed olefin isomers would be welcome, particularly if they could afford products that also might allow complexation with metal ions. The triazole ring system shares many features with an amide functional group and can be viewed as a bioisosteric replacement.^30–31 From that perspective, it is reasonable to question whether an amide could be employed in place of a triazole moiety and still preserve the biological activity as a GGDPS inhibitor. To explore this concept, we have prepared a set of new bisphosphonates linked to an isoprenoid chain through an amide bond. The synthesis of these new compounds and their biological activity are the subjects of this report.

2. Chemical Synthesis.

Amides of the general structure 2 can be seen to arise from the amine 3 and a variety of carboxylic acids (4) via amide formation followed by hydrolysis of the resulting phosphonate esters (Scheme 1), as long as the amide bond proves stable to the conditions of the standard McKenna hydrolysis.³² The amine 3 already has been synthesized through a 3-step process^33–35 from tetraethyl methylenebisphosphonate (5). The necessary carboxylic acids 4 can be obtained from readily available terpenoid alcohols (6) through various means depending on the degree of homologation. The normal-length carboxylic acids could be derived by oxidation of the alcohol to the corresponding aldehydes³⁶ followed by Pinnick oxidation to the carboxylic acids.³⁷ Homologated aldehydes are known intermediates in a synthetic route which was employed in our synthesis of some homologated triazoles.²¹ Alternatively, these carboxylic acids can be seen as a product of oxidation of the known homologated terpenoid alcohols, achievable through a sequence including oxidation to the aldehyde, Wittig condensation with methylene triphenylphosphorane, and hydroboration-oxidation,³⁸ or by hydrolysis of geranyl and neryl nitriles.³⁹ Finally, the bishomologated acids 4h, 4i, and 4j are known to be available through a malonic ester synthesis,²² again starting with an isoprenoid alcohol 6.

Scheme 1.

Retrosynthesis of amide bisphosphonates.

Our initial synthetic efforts were focused on optimizing amide formation with the commercially available carboxylic acid 4a, a mixed geranyl/neryl (E/Z) olefin isomer that affords the amide 10a (Scheme 2). While there is some precedent for amide formation in similar systems,³³ in our hands the reaction of amine 3 with the commercial acid 4a in the presence of N,N’-dicyclohexylcarbodiimide (DCC) afforded a complex mixture. Attention then turned to amide formation from the acyl halide 9a. Both traditional Schotten-Baumann conditions and the direct reaction of the acyl halide and amine 3 afforded the amide 10a in only moderate yield. Significant improvements were found with a protocol from Zhang et al.⁴⁰ in which the acyl halide is added slowly to a cooled solution of the amine in the presence of an excess of anhydrous potassium carbonate. Subsequent hydrolysis of the phosphonate esters in the presence of trimethylsilyl bromide and collidine,³² followed by brief treatment with NaOH, afforded the terpenoid amide bisphosphonate salt 2a.

Scheme 2.

Synthesis of the amide bisphosphonate 2a.

Once acceptable conditions for formation of the terpenoid amides were in hand, a series of analogues was prepared by parallel reactions (Scheme 3). Single isomers of both geranic acid (4b) and nerolic acid (4c) were obtained in good to excellent yield by oxidation of the alcohols 6b and 6c, first with TEMPO to afford the aldehydes 11b and 11c³⁶ and then by Pinnick oxidation to afford the respective carboxylic acids 4b and 4c.³⁷ Commercial (S)-citronellol (6d) was oxidized to (S)-citronellal (11d) under Swern conditions.⁴¹ Subsequent Pinnick oxidation of both this (S)-citronellal (11d) and commercial (R)-citronellal (11e) afforded the citronellic acids 4d and 4e respectively. Homologation of geranial and neral (aldehydes 11b and 11c) could be accomplished by Wittig reaction with methylene triphenylphosphorane to afford the respective olefins,³⁸ which were then subjected to a known hydroboration-oxidation scheme with 9-BBN in tetrahydrofuran.^{24, 42} The resulting alcohols (6f and 6g) were oxidized first with PCC to the aldehydes, which were then immediately subjected to Pinnick conditions to yield the carboxylic acids 4f and 4g. Finally, the bishomologated carboxylic acids 4h, 4i, and 4j were obtained through a known malonic ester synthesis starting from alcohols geraniol (6b), nerol (6c), and prenol (6j).²²

Scheme 3.

Synthesis of the carboxylic acids 4b-j.

Subsequent conversion of the carboxylic acids 4 to their respective amides 10 through the acid chlorides 9 proceeded in good to fair yield (Table 1). Hydrolysis of the bisphosphonate esters upon treatment with trimethylsilyl bromide and collidine in dichloromethane afforded the terpenoid amide bisphosphonate salts 2 in moderate to good yield, with the exception of the homogeranyl amide 10f. In all attempts to hydrolyze this specific amide, either incomplete hydrolysis or decomposition⁴³ was observed via NMR or ESIMS analysis. The other amide bisphosphonates were then subjected to a variety of bioassays.

Table 1.

Yields for conversion of carboxylic acids 4 to amide bisphosphonates 2.

RCO2H	9	10	2
4a	97%	60%	94%
4b	90%	48%	87%
4c	66%	35%	79%
4d	62%	61%	18%
4e	60%	58%	53%
4f	85%	37%	0%
4g	62%	50%	87%
4h	71%	39%	20%
4i	86%	70%	54%
4j	82%	70%	99%

3. Biological Results

All nine novel amide bisphosphonates were subjected to enzymatic (GGDPS and FDPS) as well as cellular assays (immunoblot analysis and ELISA) in order to determine the potency and selectivity of these agents as GGDPS inhibitors. Immunoblot analysis of H-Ras (a representative farnesylated protein) and Rap1a (a representative substrate of GGTase I) enables detection of impaired farnesylation and geranylgeranylation while accumulation of intracellular light chain (as measured by ELISA) is a marker for disrupted Rab geranylgeranylation.⁶ Lovastatin, an HMG-CoA reductase inhibitor, was included as a positive control, as it globally disrupts prenylation.⁶ Table 2 summarizes the results of these assays.

Table 2.

Summary of the bioassay results for the novel bisphosphonate amides.

Compound	GGDPS IC50 (μM)	FDPS IC50 (μM)	Fold-selectivity for GGDPS compared to FDPS	LEC (μM)1
120	0.075 ± 0.013	33.3 ± 5.6	445	0.03
2a	2.4 ± 1.2	28.1 ± 5.9	11.7	10
2b	4.7 ± 1.2	21.1 ± 3.1	4.5	5
2c	37.7 ± 3.2	29.9 ± 6.6	0.8	50
2d	5.4 ± 1.5	27.8 ± 3.7	5.1	50
2e	3.4 ± 0.8	48.5 ± 14.8	14.3	5
2g	3.7 ± 0.6	7.8 ± 1.9	2.1	5
2h	0.8 ± 0.3	14.3 ± 3.7	17.9	1
2i	1.1 ± 0.5	31.7 ± 14.6	28.8	0.5
2j	14.4 ± 4.7	44.4 ± 16.8	3.1	50

¹Cellular LEC (lowest effective concentration) is defined as the lowest concentration for which an unmodified Rap1a band is visible in the immunoblot and a statistically significant increase in intracellular lambda light chain is observed in the ELISA.

These studies demonstrated the impact of both chain length and stereochemistry on GGDPS inhibitor potency. As shown in Figure 2, the mixture of the geranyl and neryl amides (2a) was approximately 5-fold more potent than the bishomoprenyl compound (2j) with respect to disruption of Rap1a geranylgeranylation and accumulation of intracellular light chains. Those results, in conjunction with the lack of effect of the compounds on H-Ras prenylation, and the results from the enzymatic studies, confirmed the specificity of these compounds as GGDPS inhibitors. The preparation of the pure geranyl (2b) and neryl (2c) isomers allowed for evaluation of the effects of olefin stereochemistry on potency. As shown in Figure 3A, the geranyl isomer was approximately 10-fold more potent than the neryl isomer with respect to inhibition of Rap1a geranylgeranylation. ELISA results confirmed potency as determined by the Rap1a immunoblots (data not shown). Interestingly, however, the geranyl isomer also disrupted H-Ras farnesylation. These results conflicted with the enzyme assay results, in which the geranyl isomer was found to be a more selective inhibitor against GGDPS compared to FDPS, while the neryl isomer inhibited both enzymes with similar potency. To explore these findings further, FTase and GGTase I enzyme assays were performed. The geranyl isomer 2b was found to be a weak FTase inhibitor (IC₅₀ 47.5 + 8.6 μM) and to have no inhibitory activity against GGTase I. In contrast, the neryl isomer 2c, displayed less than 30% inhibitory activity at 100 μM against FTase.

Figure 2.

Cellular activity of compounds 2j and 2a are consistent with GGDPS inhibition. RPMI-8226 myeloma cells were incubated for 48 hours in the presence or absence of lovastatin (“Lov”) or varying concentrations of compounds 2j or 2a. A) Immunoblot analysis of H-Ras, Rap1a (antibody detects only unmodified protein) and β-tubulin (as a loading control) was performed. B) Intracellular lambda light chain concentrations were determined via ELISA. Data are expressed as percentage of control (mean + SD, n=3). The * denotes p<0.05 per t-test and compares treated cells to untreated control cells.

Figure 3.

Cellular activity of novel bisphosphonate amides. RPMI-8226 cells were incubated for 48 hours in the presence or absence of lovastatin (“Lov”) or varying concentrations of compounds 2b and 2c (A), 2g (B), 2h and 2i (C) or 2e and 2d (D). Immunoblot analysis of H-Ras, Rap1a (antibody detects only unmodified protein) and β-tubulin (as a loading control) was performed.

Extension of the chain by one carbon to yield the homoneryl amide (2g) resulted in a 10-fold increase in GGDPS inhibitor activity with respect to the neryl compound (2c), as shown in Figure 3B. Extension to the bishomologated length further enhanced cellular potency (Figure 3C). The most potent inhibitor of the series was revealed to be bishomoneryl compound 2i which was slightly more potent than the bishomogeranyl 2h in the cellular assays, although equipotent in the enzyme assays. Finally, evaluation of the citronellyl isomers demonstrated that the R-isomer 2e was 10-fold more potent than the corresponding S-isomer 2d in cellular assays (Figure 3D). None of these compounds (2g, 2h, 2i, 2e, 2d) disrupted H-Ras farnesylation and no inhibitory activity was observed in the FTase and GGTase I assays at concentrations as high as 100 μM (data not shown).

4. Discussion

Here we present a series of novel isoprenoid amide bisphosphonate-based GGDPS inhibitors. These amide-based inhibitors were designed as bioisosteres of some of our previously reported triazole-based inhibitors.^{18–19, 21} As shown in Figure 4, we hypothesized that the activity of the bishomoneryl amide 2i would be similar to that of the homoneryl triazole 1, as the two compounds have a high degree of overlap using three-dimensional rendering in which the molecules were minimized with a universal force field.⁴⁴ As shown in Figure 5, when drawn in an extended conformation, compounds 2h and 2i place the first olefin at nearly the same distance from the bisphosphonate head group. In contrast, the two amides with shorter chains (2b and 2c) place this olefin significantly closer to the polar head group than found in the triazole 1. Both of these amides have higher IC₅₀ values, with the Z-olefin 2~4-fold less potent than the Z-olefin 2i. Placement of the terminal olefin in compound 2i also approximates the position of this functionality in the triazole 1 more closely than any other amide. Thus, while it was perhaps not surprising that the most potent amide in this series did turn out to be compound 2i (LEC of 0.5 μM), we note that it is approximately 17-times less potent than the corresponding triazole 1.¹⁹ Although there is a marked similarity in the orientation of the majority of the atoms within compounds 1 and 2i (Figure 5), it is apparent that the supposedly subtle differences induced by the amide moiety in comparison to the triazole moiety, must be sufficient to lead to more significant differences in interaction with key components of the enzyme’s binding sites.

Figure 4.

Three dimensional rendering of the amide 2i and the triazole 1, as well as an overlay of the two compounds. Both structures were generated in an energy-minimized conformation using Avogadro⁴⁴ and then displayed using Chimera.⁴⁵

Figure 5.

Impact of olefin isomer and chain length on potency of amide bisphosphonate inhibitors of GGDPS compared to the homoneryl triazole bisphosphonate. The IC₅₀ values from the in vitro enzymatic assay are shown.

Our previous structure-function work with the triazole-based GGDPS inhibitors consistently demonstrated that the cis-olefin isomers are more potent than the trans-olefin isomers, regardless of chain length.^{18, 20, 22} In the present series, while differences in potency between the geranyl/neryl (2b/2c) and bishomogeranyl/bishomoneryl (2h/2i) analogues were observed, there was not a consistent pattern, suggesting that the amides interact with the isoprenoid binding sites in a manner that is disparate from the triazoles. Computational modeling studies with 1 and its corresponding trans-olefin isomer indicated that while both isomers can occupy either the farnesyl pyrophosphate (FDP) substrate site or the GGDP product site, the homoneryl isomer preferentially occupies the FDP site while the homogeranyl isomer preferentially occupies the GGDP site.²⁰ Whether the cis- and trans-olefin isomers of the bisphosphonate amides have differences in affinity to the FDP and GGDP sites is not known and future modeling studies and co-crystallization studies would be necessary to further understand the interaction/s between these novel inhibitors and the substrate/product binding sites of GGDPS. Furthermore, whether the potency of the bisphosphonate amides can be enhanced by methylation at the alpha-carbon position, as was demonstrated with the corresponding triazole bisphosphonates,²³ or by other modifications at the alpha-carbon position or distal to the amide moiety remains to be determined.

In conclusion, these studies represent the first description of amide-based isoprenoid bisphosphonate GGDPS inhibitors. While thus far, these inhibitors have been found to be less potent than their triazole counterparts, these compounds may still be useful if they are found to have more favorable biodistribution properties in vivo. Our prior preclinical studies with key triazole bisphosphonate GGDPS inhibitors have revealed a complex structure-function relationship with respect to olefin stereochemistry, presence of an alpha-carbon substituent, pharmacokinetic parameters and biodistribution profiles,^{7, 11} and it is unknown whether replacement of the triazole moiety with the amide will further impact these parameters.

Experimental Procedures and methods

5.1. General Experimental Procedures.

Acetonitrile was distilled from calcium hydride prior to use. All other reagents and solvents were purchased from commercial sources and used without further purification. All reactions in non–aqueous solvents were conducted in oven–dried glassware under a positive pressure of argon or nitrogen and with magnetic stirring. For TLC analyses, pre-coated silica polyester- (200 μm thickness, UV254 indicator) or glass-backed plates were used. Flash column chromatography was carried out using silica gel (60 Å, 40–63 μm). The purity of the final compounds was corroborated by HPLC analysis using an Agilent 1120 infinity LC solvent delivery system with a variable wavelength UV detector, and compounds for bioassay were >95% pure at 254 nm. All NMR spectra were obtained at either 300 or 400 MHz for ¹H, 75 or 100 MHz for ¹³C, and 121 or 162 MHz for ³¹P with internal standards of (CH₃)₄Si (¹H, 0.00 ppm) or CDCl₃ (¹H, 7.27; ¹³C, 77.2 ppm) or CD₃OD (¹H, 3.31; ¹³C, 49.0 ppm) or CD₃C(O)CD₃ (¹H, 2.05; ¹³C, 206.3 ppm) or CD₃CN (¹H, 1.94; ¹³C, 118.3 ppm) for non–aqueous samples or D₂O (¹H, 4.80 ppm) for aqueous samples. The ³¹P chemical shifts are reported in ppm relative to 85% H₃PO₄ (external standard). High-resolution mass spectra were obtained by TOF MS ES+ at the University of Iowa Mass Spectrometry Facility.

5.2. (3S)-3,7-Dimethyloct-6-enoic acid (4d).

In a procedure modified from Kim, et al.³⁷, (S)-citronellal 11d⁴¹ (0.93 g, 6 mmol) was dissolved in a mixture of tert-butyl alcohol (30 mL) and 2-methyl-2-butene (6.4 mL, 60 mmol, 10 eq). The resulting solution was stirred and cooled to 0 °C. A mixture of NaClO₂ (80% mixture with NaCl, 0.95 g, 8.4 mmol, 1.4 eq) and NaH₂PO₄ (0.72 g, 6 mmol, 1 eq) was dissolved in a minimum volume of water and then added to the reaction. The mixture was stirred and monitored via TLC until complete consumption of the aldehyde was observed. The reaction was diluted with ethyl acetate and brine and the layers were separated. The aqueous layer was acidified with NaHSO₄ (aq) to pH 3 and extracted three times with ethyl acetate. Additional acid was added between extractions as necessary to maintain pH 3. The combined organic layers were dried (Na₂SO₄) and concentrated under reduced pressure. Flash chromatography of the residual oil afforded carboxylic acid 4d (5.6 g, 6 mmol, >99% yield) with ¹H NMR data consistent with that in the literature.⁴⁶

5.3. (3R)-3,7-dimethyloct-6-enoic acid (4e).

Following the general procedure for carboxylic acid 4d, Pinnick oxidation of commercial (R)-citronellal (11e, 1.12 mL, 5.6 mmol) afforded carboxylic acid 4e (0.95 g, 5.6 mmol, >99% yield) with ¹H NMR data consistent with that in the literature.⁴⁶

5.4. (3Z)-4,8-Dimethylnona-3,7-dienoic acid (4g).

To a solution of alcohol 6g²¹ (1.18 g, 7 mmol) in dichloromethane (70 mL) was added flame-dried celite (11 g) and pyridinium dichromate (2.26 g, 10.5 mmol, 1.5 eq). The reaction was stirred at room temperature until complete consumption of the alcohol was observed by TLC. The reaction was filtered through a bed of Florisil and the solids were washed 3 times with ether. The combined filtrate was concentrated under reduced pressure to afford crude homoneral 11g which was used without further purification. Following the general procedure for (S)-citronellic acid 4d, homoneral 11g was oxidized to the title compound (1.12 g, 6.1 mmol, 88%). The ¹H and ¹³C NMR data matched that in the literature.⁴⁷

5.5. Diethyl 1-(diethoxyphosphoryl)-3-[3,7-dimethylocta-2,6-dienamido]propyl]phosphonate (10a).

In a procedure modified from Zhang, et al.⁴⁰ commercial carboxylic acid 4a (1 mL, 5 mmol) was stirred with a 3.8 M aqueous solution of sodium hydroxide (1.4 mL, 5.25 mmol, 1.05 equivalents) for 30 minutes, after which water was removed on a lyophilizer. The resulting sodium carboxylate was added to a round-bottom flask. Toluene was added and then removed under reduced pressure three times, and the solids were stirred under vacuum for hours. Anhydrous toluene was added (25 mL) followed by pyridine (0.06 mL, 0.75 mmol, 0.15 eq). The resulting mixture was cooled to 0 °C and a 1 M solution of oxalyl chloride in anhydrous toluene (1.00 mL, 11.5 mmol, 2.3 eq) was added dropwise. The reaction was allowed to stir for 45 minutes, the solids were removed via filtration, and the filtrate was concentrated under reduced pressure to afford the acyl chloride 9a (866 mg, 4.6 mmol, 93%) which was used immediately without further purification.

To a round-bottomed flask fitted with a magnetic stirbar was added amine 3^33–35 (930 mg, 2.8 mmol), THF (280 mL), and anhydrous potassium carbonate (3.1 g, 22.4 mmol, 8 eq). After the reaction mixture was stirred and cooled to −10 °C, the acyl halide 9a (836 mg, 4.5 mmol, 1.6 eq) was added dropwise over 2 hours as a 0.25 M solution in THF. The reaction was allowed to stir with cooling and monitored via TLC until the amine was consumed. The mixture was filtered, the solids were washed with ethyl acetate, and the filtrate was concentrated under reduced pressure. The resulting residue was dissolved in ethyl acetate and extracted with water. The aqueous layer was extracted three times with ethyl acetate, after which the combined organic layers were dried (MgSO₄) and filtered. The filtrate was concentrated under reduced pressure and the residue was purified by flash chromatography to afford the desired amide 10a as a yellow oil (785 mg, 1.68 mmol, 60% yield): ¹H NMR (400 MHz, CDCl₃) δ 6.52 (t, J = 5.8 Hz, 1H minor), 6.45 (t, J = 5.3 Hz, 1H major), 5.58 (s, 1H minor), 5.56 (s, 1H major), 5.16 (t, J = 7.1 Hz, 1H minor), 5.07 (t, J = 6.7 Hz, 1H major), 4.18 (quint d, J = 7.3, 3.4 Hz, 8H), 3.48 (m, 2H), 2.39 (m, 1H), 2.13 (m, 9H), 1.83 (s, 3H minor), 1.68 (s, 3H major), 1.62 (s, 3H minor), 1.61 (s, 3H major), 1.57 (s, 3H minor), 1.35 (t, J = 7.1 Hz, 12H); ¹³C (100.6 MHz, CDCl₃) δ 167.8, 166.8, 153.9, 153.8, 132.0, 131.6, 124.0, 123.2, 118.8, 118.1, 62.7 (t, J_PC = 6.4 Hz, 4C), 40.7, 38.1 (t, J_PC = 7.0 Hz), 34.3 (t, J_PC = 133.7 Hz), 32.9, 32.3 (?), 26.7, 26.1, 25.6, 25.1 (t, J_PC = 4.7 Hz), 24.7, 18.0, 17.5, 16.2 (d, J_PC = 6.2 Hz, 4C); ³¹P (162 MHz, CDCl₃) δ 23.5. HRMS (ESI) m/z calculated for C₂₁H₄₂NO₇P₂ [M + H]⁺ 482.2437, found 482.2440

5.6. Diethyl [1-(diethoxyphosphoryl)-3-[(2E)-3,7-dimethylocta-2,6-dienamido]propyl]phosphonate (10b).

Following the general procedure for formation of acyl chloride 9a, the acid chloride 9b was prepared from geranic acid 4b^36–37 (673 mg, 4 mmol) as a yellow oil (670 mg, 3.6 mmol, 90% yield). Subsequently, following the general procedure for formation of amide 10a, the title compound was obtained from the acid chloride 9b (390 mg, 2.1 mmol) and bisphosphonate amine 3 (500 mg, 1.5 mmol) as a yellow oil (340 mg, 0.70 mmol, 48%): ¹H NMR (400 MHz, CDCl₃) δ 6.46 (t, J = 4.8 Hz, 1H), 5.55 (s, 1H), 5.06 (t, J = 6.0 Hz, 1H), 4.17 (m, 8H), 3.49 (q, J = 5.8 Hz, 2H), 2.39 (tt, J_PH = 24.3 Hz, J = 5.9 Hz, 1H), 2.2–2.0 (m, 9H), 1.67 (s, 3H), 1.59 (s, 3H), 1.34 (t, J = 7.0 Hz, 12H); ¹³C (100.6 MHz, CDCl₃) δ 167.4, 154.2, 132.4, 123.4, 118.2, 62.97 (d, J_PC = 6.9 Hz, 2C), 62.89 (d, J_PC = 6.8 Hz, 2C), 40.9, 38.2 (t, J_PC = 6.9 Hz), 34.6 (t, J_PC = 133.9 Hz), 26.2, 25.7, 25.1 (t, J_PC = 4.8 Hz), 18.2, 17.7, 16.4 (d, J_PC = 6.2 Hz, 4C);³¹P (162 MHz, CDCl₃) δ 23.6. HRMS (ESI) m/z calculated for C₂₁H₄₂NO₇P₂ [M + H]⁺ 482.2437, found 482.2439

5.7. Diethyl [1-(diethoxyphosphoryl)-3-[(2Z)-3,7-dimethylocta-2,6-dienamido]propyl]phosphonate (10c).

Following the general procedure for formation of acyl chloride 9a, the acid chloride 9c was prepared from nerolic acid 4c^36–37 (673 mg, 4 mmol) as a yellow oil (495 mg, 2.65 mmol, 66%). Following the general procedure for formation of amide 10a, the title compound was obtained from the acid chloride 9c (420 mg, 2.25 mmol) and bisphosphonate amine 3 (500 mg, 1.5 mmol) as a yellow oil (251 mg, 0.52 mmol, 35%): ¹H NMR (400 MHz, CDCl₃) δ 6.45 (br, 1H), 5.54 (s, 1H), 5.14 (t, J = 6.7 Hz, 1H), 4.15 (bm, 8H), 3.46 (q, J = 4.8 Hz, 2H), 2.61 (t, J = 7.7 Hz, 2H), 2.36 (tt, J_PH = 24.2, J = 5.6 Hz, 1H), 2.14 (m, 4H), 1.80 (s, 3H), 1.65 (s, 3H), 1.60 (s, 3H), 1.32 (t, J = 6.8 Hz, 12H); ¹³C (100.6 MHz, CDCl₃) δ 166.9, 154.3, 132.0, 124.2, 118.9, 62.9 (t, J_PC = 5.7 Hz, 4C), 38.3 (t, J_PC = 6.6 Hz), 34.6 (t, J_PC = 143.4 Hz), 33.2, 27.0, 25.8, 25.2 (t, J_PC = 4.4 Hz), 25.0, 17.8, 16.5 (d, J = 5.9 Hz, 2C); ³¹P (121.5 MHz, CDCl₃) δ 23.6. HRMS (ESI) m/z calculated for C₂₁H₄₂NO₇P₂ [M + H]⁺ 482.2437, found 482.2439.

5.8. Diethyl [1-(diethoxyphosphoryl)-3-[(3S)-3,7-dimethyloct-6-enamido]propyl]phosphonate (10d).

Following the general procedure for formation of acyl chloride 9a, the acid chloride 9d was prepared from (S)-citronellic acid 4d (400 mg, 2.35 mmol) as a yellow oil (275 mg, 1.5 mmol, 62%). Following the general procedure for formation of amide 10a, the title compound was obtained from the acid chloride 9d (226 mg, 1.2 mmol) and bisphosphonate amine 3 (248 mg, 0.75 mmol) as a yellow oil (221 mg, 0.46 mmol, 61%): ¹H NMR (300 MHz, CDCl₃) δ 6.55 (t, J = 4.6 Hz, 1H), 5.08 (tm, J = 7.3, 1H), 4.17 (d quint, J = 7.2, 2.9 Hz, 8H), 3.46 (q, J = 6.0 Hz, 2H), 2.9–1.8 (m, 9H), 1.66 (s, 3H), 1.58 (s, 3H), 1.34 (t, J = 7.1 Hz, 12H), 1.16 (m, 1H) 0.91 (d, J = 6.3 Hz, 3H); ¹³C (100.6 MHz, CDCl₃) δ 172.8, 131.3, 124.4, 62.8 (t, J_PC = 6.2 Hz, 4C), 44.5, 38.6 (t, J_PC = 7.2 Hz), 37.0, 34.7 (t, J_PC = 133.9 Hz), 30.4, 25.7, 25.5, 25.1 (t, J_PC = 4.6 Hz), 19.5, 17.7, 16.4 (d, J_PC = 5.3 Hz, 4C); ³¹P (121.5 MHz, CDCl₃) δ 23.5. HRMS (ESI) m/z calculated for C₂₁H₄₄NO₇P₂ [M + H]⁺ 484.2593, found 484.2594

5.9. Diethyl [1-(diethoxyphosphoryl)-3-[(3R)-3,7-dimethyloct-6-enamido]propyl]phosphonate (10e).

Following the general procedure for formation of acyl chloride 9a, the acid chloride 9e was prepared from (R)-citronellic acid 4e (460 mg, 2.7 mmol) as a yellow oil (304 mg, 1.6 mmol, 60%). Following the general procedure for formation of amide 10a, the title compound was obtained from the acid chloride 9e (304 mg, 1.6 mmol) and bisphosphonate amine 3 (381 mg, 1.15 mmol) as a yellow oil (320 mg, 0.66 mmol, 58%). Both ¹H and ³¹P NMR data matched that given above for the (S)-enantiomer 10d.

5.10. Diethyl [1-(diethoxyphosphoryl)-3-[(3E)-4,8-dimethylnona-3,7-dienamido]propyl]phosphonate (10f).

Following the general procedure for formation of acyl chloride 9a, the acid chloride 9f was prepared from homogeranic acid 4f (54 mg, 0.3 mmol) as a yellow oil (51 mg, 0.25 mmol, 85%). Following the general procedure for formation of amide 10a, the title compound was obtained from the acid chloride 9f (51 mg, 0.25 mmol) and bisphosphonate amine 3 (53 mg, 0.16 mmol) as a yellow oil (29 mg, 0.06 mmol, 37%): ¹H NMR (400 MHz, CDCl₃) δ 6.42 (t, J = 5.2 Hz, 1H), 5.30 (t, J = 7.3 Hz, 1H), 5.07 (t, J = 6.3, 1H), 4.16 (quint d, J = 7.2, 4.1 Hz, 8H), 3.44 (q, J = 6.3 Hz, 2H), 2.95 (d, J = 7.4 Hz, 2H), 2.32 (tt, J_PH = 24.0 Hz, J = 6.2 Hz, 1H), 2.2–2.0 (m, 6H), 1.68 (s, 3H), 1.63 (s, 3H), 1.60 (s, 3H), 1.34 (t, J = 7.0 Hz, 12H); ¹³C (100.6 MHz, CDCl₃) δ 171.9, 140.9, 132.0, 124.0, 116.8, 62.97 (d, J_PC = 6.7 Hz, 2C), 62.92 (d, J_PC = 7.1 Hz, 2C), 39.8, 38.7 (t, J_PC = 7.4 Hz), 36.2, 34.6 (t, J_PC = 133.8), 26.6, 25.9, 25.4 (t, J_PC = 4.9), 17.9, 16.6 (d, J_PC = 6.6 Hz, 4C), 16.4; ³¹P (162.0 MHz, CDCl₃) δ 23.3. HRMS (ESI) m/z calculated for C₂₂H₄₃NO₇P₂Na [M + Na]⁺ 518.2412, found 518.2409.

5.11. Diethyl [1-(diethoxyphosphoryl)-3-[(3Z)-4,8-dimethylnona-3,7-dienamido]propyl]phosphonate (10g).

Following the general procedure for formation of acyl chloride 9a, the acid chloride 9g was prepared from homonerolic acid 4g (365 mg, 2 mmol) as a yellow oil (250 mg, 1.25 mmol, 62%). Following the general procedure for formation of amide 10a, the title compound was obtained from the acid chloride 9g (250 mg, 1.3 mmol) and bisphosphonate amine 3 (265 mg, 0.8 mmol) as a yellow oil (200 mg, 0.4 mmol, 50%): ¹H NMR (400 MHz, CDCl₃) δ 6.40 (t, J = 5.0 Hz, 1H), 5.30 (t, J = 7.0 Hz, 1H), 5.08 (m, 1H), 4.17 (m, 8H), 3.45 (q, J = 6.4 Hz, 2H), 2.95 (d, J = 6.2 Hz, 2H), 2.33 (tt, J_PH = 24.0 Hz, J = 6.2 Hz, 1H), 2.2–2.0 (m, 6H), 1.76 (s, 3H), 1.67 (s, 3H), 1.60 (s, 3H), 1.35 (t, J = 7.0 Hz, 12H); ¹³C (75.5 MHz, CDCl₃) δ 172.0, 140.8, 132.2, 123.8, 117.4, 62.9 (t, J_PC = 5.8 Hz, 4C), 38.6 (t, J_PC = 7.3 Hz), 35.9, 34.6 (t, J_PC = 133.7 Hz), 32.1, 26.5, 25.8, 25.4 (m), 23.6, 17.8, 16.5 (d, J_PC = 6.1 Hz, 4C); ³¹P (162 MHz, CDCl₃) δ 23.3. HRMS (ESI) m/z calculated for C₂₂H₄₄NO₇P₂ [M + H]⁺ 496.2593, found 496.2599.

5.12. Diethyl [1-(diethoxyphosphoryl)-3-[(4E)-5,9-dimethyldeca-4,8-dienamido]propyl]phosphonate (10h).

Following the general procedure for formation of acyl chloride 9a, the acid chloride 9h was prepared from bishomogeranic acid 4h²² (589 mg, 3 mmol) as a yellow oil (457 mg, 2.1 mmol, 71%). Following the general procedure for formation of amide 10a, the title compound was obtained from the acid chloride 9h (270 mg, 1.26 mmol) and bisphosphonate amine 3 (300 mg, 0.9 mmol) as a yellow oil (180 mg, 0.312 mmol, 39%): ¹H NMR (300 MHz, CDCl₃) δ 6.52 (t, J = 5.0 Hz, 1H), 5.09 (q, J = 5.3 Hz, 2H), 4.18 (d quint, J = 7.3, 3.1 Hz, 8H), 3.46 (q, J = 6.1 Hz, 2H), 2.5–1.9 (m, 11 H), 1.67 (s, 3H), 1.61 (s, 3H), 1.60 (s, 3H), 1.34 (t, J = 7.1 Hz, 12H); ¹³C (100.6 MHz, CDCl₃) δ 173.1, 136.5, 131.4, 124.2, 122.7, 62.9 (t, J_PC = 6.3 Hz, 4C), 39.7, 38.6 (t, J_PC = 7.0), 36.8, 34.6 (t, J_PC = 133.8), 26.7, 25.7, 25.1 (t, J_PC = 4.7 Hz), 24.3, 17.7, 16.4 (d, J_PC = 6.2 Hz, 4C), 16.0; ³¹P (121.5 MHz, CDCl₃) δ 23.5. HRMS (ESI) m/z calculated for C₂₃H₄₆NO₇P₂ [M + H]⁺ 510.2750, found 510.2749.

5.13. Diethyl [1-(diethoxyphosphoryl)-3-[(4Z)-5,9-dimethyldeca-4,8-dienamido]propyl]phosphonate (10i).

Following the general procedure for formation of acyl chloride 9a, the acid chloride 9i was prepared from bishomonerolic acid 4i²² (589 mg, 3 mmol) as a yellow oil (553 mg, 2.6 mmol, 86%). Following the general procedure for formation of amide 10a, the title compound was obtained from the acid chloride 9i (346 mg, 1.61 mmol) and bisphosphonate amine 3 (381 mg, 1.15 mmol) as a yellow oil (400 mg, 0.81 mmol, 70%): ¹H NMR (300 MHz, CDCl₃) δ 6.54 (t, J = 5.1 Hz, 1H), 5.06 (t, J = 6.9 Hz, 1H), 5.06 (t, J = 6.9 Hz, 1H), 4.14 (d quint, J = 7.3, 3.1 Hz, 8H), 3.41 (q, J = 6.2 Hz, 2H), 2.4–2.0 (m, 11H), 1.64 (s, 6H), 1.57 (s, 3H), 1.31 (t, J = 7.1 Hz, 12 H); ¹³C (75.5 MHz, CDCl₃) δ 173.0, 136.7, 131.7, 124.3, 123.5, 62.9 (br, 4C), 38.6 (t, J_PC = 6.7 Hz), 37.0, 34.7 (t, J_PC = 134 Hz), 32.0, 26.7, 25.8, 25.2, 24.1, 23.5, 17.7, 16.5 (d, J_PC = 5.9 Hz, 4C); ³¹P (121.5 MHz, CDCl₃) δ 23.4. HRMS (ESI) m/z calculated for C₂₃H₄₆NO₇P₂ [M + H]⁺ 510.2750, found 510.2756.

5.14. Diethyl [1-(diethoxyphosphoryl)-3-(5-methylhex-4-enamido)propyl]phosphonate (10j).

Following the general procedure for formation of acyl chloride 9a, bishomoprenoyl chloride 9j was prepared from bishomoprenoic acid 4j²² (360 mg, 2.8 mmol) as a yellow oil (335 mg, 2.30 mmol, 82%). Following the general procedure for formation of amide 10a, the title compound was obtained from bishomoprenoyl chloride 9j (308 mg, 2.1 mmol) and bisphosphonate amine 3 (497 mg, 1.5 mmol) as a yellow oil (466 mg, 1.05 mmol, 70%): ¹H NMR (400 MHz, CDCl₃) δ 6.57 (t, J = 4.8 Hz, 1H), 5.08 (t, J = 7.0 Hz, 1H), 4.18 (quint d, J = 7.3, 4.0 Hz, 8H), 3.45 (q, J = 6.1 Hz, 2H), 2.4–2.2 (m, 3H), 2.2–2.0 (m, 4H), 1.68 (s, 3H), 1.61 (s, 3H), 1.35 (t, J = 7.2 Hz, 12H); ¹³C (100.6 MHz, CDCl₃) δ 173.2, 133.0, 122.9, 62.99 (d, J_PC = 7.1 Hz, 2C), 62.94 (d, J_PC = 6.8 Hz, 2C), 38.6 (t, J_PC = 7.1 Hz), 38.8, 34.5 (t, J_PC = 134.1 Hz), 25.8, 25.1 (t, J_PC = 4.3 Hz), 24.4, 17.8, 16.5 (d, J_PC = 6.7 Hz, 4C); ³¹P (162 MHz, CDCl₃) δ 23.5. HRMS (ESI) m/z calculated for C₁₈H₃₈NO₇P₂ [M + H]⁺ 442.2124, found 442.2124.

5.15. Disodium (1-[bis(sodiooxy)phosphoryl]-3-[3,7-dimethylocta-2,6-dienamido]propyl)phosphonate (2a).

To a round-bottomed flask fitted with a magnetic stirbar was added bisphosphonate 10a (93 mg, 0.2 mmol), collidine (0.19 mL, 1.4 mmol, 7 equivalents), and dichloromethane (7.5 mL, 10 mL per mmol bisphosphonate). The flask was cooled to 0 °C and stirred and trimethylsilyl bromide (0.22 mL, 1.68 mmol, 8 equivalents) was added dropwise. The reaction mixture was allowed to react for 24 hours while warming gradually to room temperature. The volatile materials were removed under reduced pressure. Residual volatiles were removed by diluting the residue with anhydrous toluene and removing the toluene under reduced pressure. This process was repeated three times to ensure complete removal of any residual bromotrimethylsilane. The remaining material was allowed to stand for 5 hours under vacuum. After aqueous sodium hydroxide solution (4 equivalents) was added, the mixture was stirred for at least 15 minutes and then the water was removed via lyophilization. The residual material was dissolved in a minimum volume of warm water, and then diluted carefully with anhydrous acetone. The mixture was cooled to 0 °C for 30 minutes, and the precipitate was collected by decantation and filtration. The precipitate was subjected to repetitive dissolution and precipitation until no collidine was detected in the ¹H NMR spectrum. Final dissolution in water followed by lyophilization afforded the title compound 2a as a white solid (86 mg, 0.19 mmol, 94%): ¹H NMR (300 MHz, D₂O) δ 6.18 (b, 1H major), 5.94 (br, 1H minor), 5.23 (br, 1H), 3.88 (t, J = 6.9 Hz, 2H), 2.47 (br, 1H minor), 2.4–2.1 (m, 6H), 2.1–1.8 (m, 4H), 1.73 (s, 3H), 1.67 (s, 3H); ¹³C (100.6 MHz, D₂O) δ 170.0, 169.7, 154.0, 153.8, 133.97, 133.92, 123.6, 123.4, 118.4, 117.7, 40.1 (t, J_PC = 8.1 Hz), 39.6, 37.9 (t, J_PC = 114.1 Hz), 32.4, 27.6, 26.1 (br), 25.4, 24.9, 24.8, 23.6, 17.7, 17.5, 16.92, 16.90; ³¹P (121.5 MHz, D₂O) δ 19.6 (major), 19.5 (minor). HRMS (ESI) m/z calculated for C₁₃H₂₄NO₇P₂ [M – H]⁻ 368.1028, found 368.1019.

5.16. Disodium (1-[bis(sodiooxy)phosphoryl]-3-[(2E)-3,7-dimethylocta-2,6-dienamido]propyl)phosphonate (2b).

Following the general procedure for preparation of salt 2a, the title compound was obtained from amide 10b (26 mg mg, 0.055 mmol) as a white solid (22 mg, 87%): ¹H NMR (400 MHz, D₂O) δ 5.62 (s, 1H), 5.09 (tt, J = 6.2, 1.4 Hz, 1H), 3.29 (t, J = 7.6 Hz, 2H), 2.1–2.0 (m, 4H), 2.0–1.8 (m, 5H), 1.62 (tt, J_PH = 21.6 Hz, J = 6.3 Hz, 1H), 1.59 (s, 3H), 1.52 (s, 3H); ¹³C (100.6 MHz, D₂O) δ 170.2, 167.9 (Na₂CO₃), 153.7, 134.0, 123.5, 117.8, 40.2 (br), 39.6, 38.2 (t, J_CP = 114.2 Hz), 26.2 (Br), 25.4, 24.8 (Br), 17.7, 16.9; ³¹P (162 MHz, D₂O) δ 19.6. HRMS (ESI) m/z calculated for C₁₃H₂₄NO₇P₂ [M – H]⁻ 368.1028, found 368.1019.

5.17. Disodium (1-[bis(sodiooxy)phosphoryl]-3-[(2Z)-3,7-dimethylocta-2,6-dienamido]propyl)phosphonate (2c).

Following the general procedure for preparation of salt 2a, the title compound was obtained from amide 10c (45 mg, 0.09 mmol) as a white-yellow solid (32.5 mg, 0.071 mmol, 79% yield): ¹H NMR (400 MHz, D₂O) δ 5.71 (s, 1H), 5.17 (t, J = 6.8 Hz, 1H), 3.39 (t, J = 6.3 Hz, 2H), 2.49 (t, J = 7.2 Hz, 2H), 2.15 (m, 2H), 2.1–1.9 (bm, 3H), 1.83 (s, 3H), 1.67 (s, 3H), 1.60 (s, 3H); ¹³C (100.6 MHz, D₂O) δ 170.0, 154.0, 134.1, 123.6, 118.3, 39.2 (t, J_PC = 8.1 Hz), 36.8 (t, J_PC = 116.6 Hz), 32.4, 26.0, 25.1 (br), 24.9, 23.5, 16.9; ³¹P (162.0 MHz, D₂O) δ 19.5. HRMS (ESI) m/z calculated for C₁₃H₂₄NO₇P₂ 368.1028, found 368.1030.

5.18. Disodium (1-[bis(sodiooxy)phosphoryl]-3-[(3S)-3,7-dimethyloct-6-enamido]propyl)phosphonate (2d).

Following the general procedure for preparation of salt 2a, the title compound was obtained from amide 10d (111 mg, 0.23 mmol) as a white solid. Purification by flash chromatography on C18-functionalized silica gel (3:7 H₃CCN:H₂O to 100% H₃CCN) recovered 13.9 mg of pure sodium salt (0.041 mmol, 18%): ¹H and ³¹P NMR data matched that for R-enantiomer 2e.

5.19. Disodium (1-[bis(sodiooxy)phosphoryl]-3-[(3R)-3,7-dimethyloct-6-enamido]propyl)phosphonate (2e).

Following the general procedure for preparation of salt 2a, the title compound was obtained from amide 10e (179 mg, 0.37 mmol) as a white solid. Purification by flash chromatography on C18-functionalized silica gel (3:7 H₃CCN:H₂O to 100% H₃CCN) afforded 65 mg of pure sodium salt 2e (0.20 mmol, 53%): ¹H NMR (300 MHz, CDCl₃) δ 5.25 (t, J = 6.6 Hz, 1H), 3.42 (t, J = 6.2 Hz, 2H), 2.30 (dd, J = 13.4 Hz, J = 6.4 Hz, 1H), 2.2–1.8 (m, 7H), 1.73 (s, 3H), 1.66 (s, 3H), 1.4–1.2 (m, 3H), 0.95 (d, J = 6.5 Hz, 3H); ¹³C (75.5 MHz, D₂O) δ 176.5, 133.5, 124.9, 43.6, 39.7 (t, J_PC = 8.2 Hz), 37.1 (t, J_PC = 115.5 Hz), 36.2, 30.4, 25.3 (br), 25.0, 24.9, 18.8, 17.0; ³¹P (121.5 MHz, D₂O) δ 19.5. HRMS (ESI) m/z calculated for C₁₃H₂₆NO₇P₂ [M – H]⁻ 370.1185, found 370.1179.

5.20. Disodium (1-[bis(sodiooxy)phosphoryl]-3-[(3Z)-4,8-dimethylnona-3,7-dienamido]propyl)phosphonate (2g).

Following the general procedure for preparation of salt 2a, the title compound was obtained from amide 10g (42 mg, 0.085 mmol) as a yellow solid (35 mg, 0.074 mmol, 87% yield): ¹HNMR (300 MHz, D₂O) δ 5.34 (t, J = 7.0 Hz, 1H), 5.21 (m, 1H), 3.40 (t, J = 7.5 Hz, 2H), 3.02 (d, J = 7.4 Hz, 2H), 2.2–1.8 (m, 7H), 1.78 (s, 3H), 1.71 (s, 3H), 1.65 (s, 3H); ¹³C (75.5 MHz, D₂O) 175.5, 141.7, 134.0, 124.1, 116.7, 40.0 (br), 37.2 (t, J_PC = 114.2 Hz), 35.1, 31.3, 25.8, 25.4 (br), 25.1, 22.7, 17.1; ³¹P (121.5 MHz, D₂O) δ 19.43. HRMS (ESI) m/z calculated for C₁₄H₂₆NO₇P₂ [M – H]⁻ 382.1185, found 382.1190.

5.21. Disodium (1-[bis(sodiooxy)phosphoryl]-3-[(4E)-5,9-dimethyldeca-4,8-dienamido]propyl)phosphonate (2h).

Following the general procedure for preparation of salt 2a, the title compound was obtained from amide 10h (102 mg, 0.2 mmol) as a white solid (19.6 mg, 0.040 mmol, 20% yield): ¹H NMR (300 MHz, D₂O) δ 5.11 (m, 2H), 3.28 (t, J = 7.4 Hz, 2H), 2.3–2.1 (m, 4H), 2.1–1.7 (m, 7H), 1.60 (s, 3H), 1.54 (s, 6H); ¹³C (75.5 MHz, D₂O) δ 176.3, 138.0, 133.6, 124.5, 122.5, 40.0 (br), 38.9, 37.3 (br), 36.1, 26.0, 25.3 (br), 24.9, 24.1, 17.1, 15.3; ³¹P (121.5 MHz, D₂O) δ 19.3. HRMS (ESI) m/z calculated for C₁₅H₂₈NO₇P₂ [M – H]⁻ 396.1341, found 396.1338.

5.22. Disodium (1-[bis(sodiooxy)phosphoryl]-3-[(4Z)-5,9-dimethyldeca-4,8-dienamido]propyl)phosphonate (2i).

Following the general procedure for preparation of salt 2a, the title compound was obtained from amide 10i (168 mg, 0.34 mmol) as a white solid (89.8 mg, 0.19 mmol, 54% yield): ¹HNMR (300 MHz, D₂O) δ 5.11 (m, 2H), 3.27 (t, J = 6.8 Hz, 2H), 2.3–2.1 (m, 4H), 2.1–1.7 (m, 7H), 1.60 (s, 3H), 1.59 (s, 3H), 1.53 (s, 3H); ¹³C (125.8 MHz, D₂O) δ 176.4, 138.3, 133.8, 124.4, 123.1, 39.5 (br), 36.2, 31.2, 26.0, 25.1 (br), 24.9, 24.0, 22.6, 17.0; ³¹P (121.5 MHz, D₂O) δ 19.6. HRMS (ESI) m/z calculated for C₁₅H₂₈NO₇P₂ [M – H]⁻ 396.1341, found 396.1339

5.23. Disodium (1-[bis(sodiooxy)phosphoryl]-3-(5-methylhex-4-enamido)propyl)phosphonate (2j).

Following the general procedure for preparation of salt 2a, the title compound was obtained from amide 10j (22 mg, 0.5 mmol) as a white solid (21 mg, 0.5 mmol, 100% yield): ¹H NMR (400 MHz, D₂O) δ 5.15 (t, J = 5.7 Hz, 1H), 3.34 (t, J = 6.9 Hz, 2H), 2.3–2.2 (m, 4H), 2.1–1.8 (m, 3H), 1.69 (s, 3H), 1.61 (s, 3H); ¹³C (75.5 MHz, D₂O) δ 176.3, 134.8, 122.1, 39.8 (t, J_PC = 8.1 Hz), 37.1 (t, J_PC = 113.9 Hz), 35.9, 25.1 (br), 24.8, 24.0, 16.9, ³¹P (121.5 MHz, D₂O) δ 19.4. HRMS (ESI) m/z calculated for C₁₀H₂₀NO₇P₂ [M – H]⁻ 328.0715, found 328.0710.

5.24. FDPS and GGDPS enzyme assays.

Recombinant FDPS was kindly provided by Dr. Raymond Hohl (Penn State Cancer Institute). Recombinant GGDPS was kindly provided by Dr. Edward Snell (Hauptman-Woodward Medical Research Institute). Enzyme assays were performed as previously described.¹⁹ Compounds were tested in duplicate at multiple concentrations and three independent experiments were performed.

5.25. FTase and GGTase I enzyme assays.

FTase and GGTase I activity was determined using the method of Temple et al.,⁴⁸ with modifications as previously described.²² Recombinant FTase and GGTase I were obtained from Jena Biosciences and dansyl-GCVLS and dansyl-GCVLL were obtained from Biosynthesis, Inc. Changes in fluorescence over time were detected using a Molecular Devices Spectramax Gemini EM fluorescence microplate reader. Compounds were tested in duplicate at multiple concentrations and three independent experiments were performed.

5.26. Immunoblot analysis.

RPMI-8226 cells were incubated with drugs for 48 hrs. This time point was chosen to allow for direct comparison of results from prior studies evaluating triazole-based inhibitors.^{18–19, 21, 23} Whole cell lysate was obtained using RIPA buffer (0.15 M NaCl, 1% sodium deoxycholate, 0.1% SDS, 1% Triton (v/v) X-100, 0.05 M Tris HCl) containing protease and phosphatase inhibitors. Protein content was determined using the bicinchoninic acid (BCA) method (Pierce Chemical). Equivalent amounts of cell lysate were resolved by SDS-PAGE, transferred to polyvinylidene difluoride membrane, probed with the appropriate primary antibodies (anti-H-Ras (610001, BD Transduction Laboratories), anti-Rap1a (sc-373968, Santa Cruz Biotechnology), anti-β-tubulin (T-5201, Sigma)), and detected using HRP-linked secondary antibodies and Clarity ECL (for tubulin) or Millipore Immobilon ECL (for Ras and Rap1a) western blotting reagents per manufacturer’s protocols.

5.27. Lambda light chain ELISA.

Lysates generated for the immunoblot studies were utilized for intracellular lambda light chain quantification as determined by ELISA (Bethyl Laboratories).

5.28. Statistical Analysis.

Two-tailed t-testing was used to calculate statistical significance. An α of 0.05 was set as the level of significance. Concentration response curves were analyzed via CompuSyn software (ComboSyn, Inc.,) to determine the IC₅₀ values for the enzyme assays.