Vitamin E amides, a new class of vitamin E analogues with enhanced proapoptotic activity


  • Adisa Tomic-Vatic,

    1. School of Medical Science, Griffith University, Southport, Queensland, Australia
    Search for more papers by this author
  • John EyTina,

    1. College of Pharmacy, University of South Carolina, Columbia, SC, USA
    Search for more papers by this author
  • James Chapman,

    1. College of Pharmacy, University of South Carolina, Columbia, SC, USA
    Search for more papers by this author
  • Elahe Mahdavian,

    1. Department of Chemistry and Physics, Louisiana State University, Shreveport, LA, USA
    Current affiliation:
    1. Department of Chemistry and Physics, Louisiana State University, Shreveport, LA
    Search for more papers by this author
  • Jiri Neuzil,

    Corresponding author
    1. School of Medical Science, Griffith University, Southport, Queensland, Australia
    2. Laboratory of Cell Signalling and Apoptosis, Institute of Molecular Genetics, Czech Academy of Science, Prague, Czech Republic
    • Apoptosis Research Group, Heart Foundation Research Centre, School of Medical Sciences, Griffith University, Gold Coast Campus, Southport, 9726 Queensland, Australia. Phone: +61-7-555-29109
    Search for more papers by this author
    • Fax: +61-7-555-28804

  • Brian A. Salvatore

    Corresponding author
    1. Department of Chemistry and Physics, Louisiana State University, Shreveport, LA, USA
    Current affiliation:
    1. Department of Chemistry and Physics, Louisiana State University, Shreveport, LA
    • Department of Chemistry and Physics, Louisiana State University, One University Place, Shreveport, LA 71115, USA. Phone: +1-318-7975224
    Search for more papers by this author


Vitamin E (VE) analogues, epitomized by α-tocopheryl succinate (α-TOS), are proapoptotic agents with selective antineoplastic activity. The molecule of α-TOS comprises several structurally and functionally distinct moieties that can be modified in order to yield analogues with higher activity. In order to find analogues with higher apoptogenic efficacy, we prepared novel compounds where the ester bond was replaced by an amide bond. All of these analogues were significantly more proapoptotic than their ester counterparts, with α-tocopheryl maleyl amide being the most effective. Importantly, methylation of the free carboxylic group completely obliterated apoptogenic activity of the compounds. Similarly as shown for the ester analogues, the amides induced apoptosis by mitochondrial destabilization. Superiority of amides over the ester analogues may be due to their higher partitioning into the lipid phase, as suggested by the log p-values that were lower for the amides than the corresponding esters. In conclusion, we present evidence that modification of the ester bond of agents such as α-TOS can be used as a basis for generating novel analogues with higher efficacy of killing malignant cells, an activity that suggests anticancer effect of the agents. © 2005 Wiley-Liss, Inc.

The search for novel antineoplastic agents has been focused on compounds that induce apoptotic cell death selectively in malignant cells. Many established treatments involve deleterious side effects that lead to serious secondary complications. An ideal anticancer candidate agent should induce effective apoptosis in cancer cells with high selectivity, i.e., low or no toxicity toward normal cells and tissues, should cooperate with agents featuring different modes of action and should suppress cancer growth in a preclinical model.1 A novel group of inducers of apoptosis and anticancer agents, vitamin E (VE) analogues, meets this premise.2 Thus, these agents, epitomized by α-tocopheryl succinate (α-TOS), are potent apoptogens3, 4 with high selectivity for cancer cells.5

The mechanism by which α-TOS induces apoptosis has been elucidated to some extent.6 It has been shown that central to α-TOS-triggered apoptosis are mitochondria,7 a novel target for cancer treatment.8, 9 Accordingly, exposure of multiple cells to α-TOS induces mitochondrial destabilization with ensuing relocalization of apoptotic mediators, including cytochrome c, Smac/Diablo and the apoptosis-inducing factor,10 a process modulated by the mitochondrial Bcl-2 family proteins.7, 10 The membrane-destabilizing effect of α-TOS is likely based on its detergent-like activity. In this respect, the VE analogue has been shown to cause also lysosomal destabilization that is likely to amplify the mitochondrial pathway by the release of lysosomal proteases, since cathepsin D-deficient cells were less susceptible to the inducer.11 Detergent-like activities have been reported as a principle for apoptosis induction by several reagents.12, 13 Recent data suggest that mitochondrial destabilization involves a decrease in the inner membrane potential, formation of the permeability transition pores, i.e., channels made up of the mitochondrial proteins bax and/or bid, as well as of the lipid second messengers, such as ceramide.14, 15, 16, 17

We have investigated the structure-function activity of VE analogues as inducers of apoptosis.18, 19 Our studies revealed that of the functional moieties, the agent absolutely required both the aliphatic tail, the tocopheryl head group and a dicarboxylate esterified to the phenolic hydroxyl.19 This is based on experiments in which analogues lacking either of the 3 essential moieties failed to induce apoptosis.19 Thus, the amphiphilic structure of VE analogues is essential for their proapoptotic activity, and their modulatory effect on signaling pathways20, 21, 22 may be secondary to the mitochondria-destabilizing function.1 In the course of synthesis of novel analogues of VE with the goal of increasing their proapoptotic and antineoplastic efficacy, we prepared a group of highly potent agents in which the ester bond was replaced by an amide bond that further enhanced activity of the compounds. This modification can be used as the basis for the synthesis of VE analogs with even higher proapoptotic activity that may translate into a potent anticancer effect.

Material and methods


Structures of the compounds used in the study are shown in Figure 1(a). Methylene chloride and pyridine were distilled over calcium hydride under a nitrogen atmosphere. The diazomethane solution in ether was generated from Diazald using a distillation apparatus with Clear-Seal glass joints (both items were purchased from Aldrich).

Figure 1.

(a) Structures of compounds used in this study. The numbers refer to the following compounds: 1, α-TOS; 2, δ-TOS; 3, α-TOM; 4, δ-TOM; 5, α-TAS; 6, δ-TAS; 7, α-TAM; 8, δ-TAM; 9, α-TASM. (b) Generic structures of vitamin E analogues indicating the individual domains, including the functional, signaling and hydrophobic domain.

General procedure for the preparation of tocopheryl amide carboxylic acids from tocopheryl amines.

α-tocopheryl amine23 (251 mg, 0.584 mmol) was dissolved in anhydrous dichloromethane (1 ml) in a 5 ml round-bottom flask that had been purged with nitrogen gas and equipped with a magnetic stir bar. Freshly distilled pyridine (200 μl, 2.5 mmol) was added via a syringe. The solution was cooled in an ice water bath. Succinic anhydride (146 mg, 1.46 mmol) was added in one portion, with brisk stirring. The cold-water bath was removed and the reaction mixture was allowed to stir overnight at room temperature under a nitrogen atmosphere. Thin-layer chromatography (TLC) analysis revealed that the reaction had gone to completion. The solution was concentrated via rotary evaporation, followed by placement of the crude residue on a high vacuum manifold (0.1 torr) for several hours to remove residual pyridine. The crude residue was then dissolved in a minimal amount of dichloromethane and loaded onto a flash silica gel column (mobile phase: 3:2 hexane/ethyl acetate with 0.5% acetic acid). This yielded 246 mg (79.5%) of the desired product as a clear viscous oil.

Preparation of methyl ester of tocopheryl amide succinic acid (α-TASM).

α-tocopheryl amide succinic acid (α-TAS; 85 mg, 0.161 mmol) and dichloromethane (1 ml) were added to a scratch-free 25 ml Erlenmeyer flask. An ice-cold solution of diazomethane in ether (0.7 ml, ∼ 0.3 mol) was added to the flask drop-wise with a flame-polished Pasteur pipette. Nitrogen gas evolved for several minutes, and the solution acquired a yellow color when the reaction neared completion. The progress of the reaction was monitored by TLC; it was generally complete within 5 min. Upon completion, a dilute solution of acetic acid in methanol was added drop-wise, until the yellow color disappeared and the solution became colorless. The solvent was removed by rotary evaporation, followed by placement under high vacuum (0.1 torr). This reaction generally provides quantitative yields of the desired methyl ester (pure by TLC). Flash column chromatography of the product is optional. This can be performed in hexane/ethyl acetate (6:1) to provide a sample for analytical characterization.

Cell culture

The human Jurkat T lymphoma and U937 leukemic cells were maintained in RPMI-1640, while the human Meso-2 malignant mesothelioma (MM) cells (sarcomatoid phenotype)24 and human foreskin fibroblasts (line AG01518; Coriell Institute, Camden, NJ) were grown in DMEM, both supplemented with 2 mM L-glutamine, 100 units/ml penicillin, 100 μg/ml streptomycin and 10% FCS at 37°C in a humidified atmosphere of 5% CO2. The suspension cells were used in experiments at 0.5 × 106 per ml, while the adherent cells at 50–60% confluency.

Cytotoxicity analysis

Meso-2 cells were plated in 96-well flat-bottom tissue culture plate at 2.5 × 103 per well. The cells were allowed to attach overnight, then incubated for 24 hr with individual VE analogues at ≤ 100 μM, dissolved in ethanol and diluted in complete DMEM to the final concentration and added to cells at 0.1% (v/v) of ethanol. Cell viability was determined using the MTT assay.25 Briefly, following exposure of cells, 10 μl of MTT (5 mg/ml in PBS) was added, and after incubation for 4 hr at 37°C, the medium was removed and combined with 200 μl of 1% SDS. Absorbance was read at 550 nm using an ELISA plate reader and background absorbance was considered as 100%. Survival curves were generated and the IC50 values determined.

Assessment of apoptotic markers

Apoptosis was quantified using the annexin V-FITC method, which detects phosphatidyl serine (PS) externalized in the early phases of apoptosis.26 Briefly, treated cells were collected, washed, resuspended in 0.1 ml binding buffer (10 mM HEPES, 140 mM NaCl, 5 mM CaCl2, pH 7.4), incubated for 20 min at room temperature with 2 μl annexin V-FITC (PharMingen), supplemented with 10 μl of propidium iodide (PI; 10 μg/ml) and analyzed by flow cytometry (FACSCalibur; Becton Dickinson, San Jose, CA) using channel 1 for annexin V-FITC binding and channel 2 for PI staining.

Mitochondrial destabilization was assessed by flow cytometric evaluation of the mitochondrial inner transmembrane potential (ΔΨm) using the polychromatic probe 5,5′,6,6′-tetrachloro-1,1′,3,3′-tetraethylbenzimidoazolyl-carbocyanino iodide (JC-1; Molecular Probes, Eugene, OR). JC-1 accumulates in highly energized mitochondrial with ΔΨm ≥ −110 mV, giving strong red fluorescence. Upon dissipation of ΔΨm, JC-1 leaks into the cytosol, where it yields green fluorescence. In brief, treated cells were collected, washed, incubated with 10 μM JC-1 in DMSO (30 min, 37°C) and assessed for increase of green fluorescence in a flow cytometer.

Flow cytometric assessment of protein expression

Cells were treated as indicated, harvested, fixed with 3.7% formalin in PBS and permeabilized with 0.2% saponin in PBS containing 2% FBS (both at room temperature for 1 hr). The cells were then exposed to anti-Bcl-2 (Boehringer, Mannheim, Germany), anti-Bcl-xL, anti-Mcl-1 and anti-Bax IgG (all Santa Cruz Biotechnology, Santa Cruz, CA), followed by an FITC-conjugated secondary antibody. Finally, the level of protein expression was estimated by analyzing the cells with a flow cytometer.

Transfection of cells

The plasmids (pcDNA3.1) were used in which the Bcl-2 or Bcl-xL genes, as well as their N-terminal deletion mutants, were fused to enhanced green fluorescence protein (EGFP).27 Transfections were carried out basically as reported earlier.7 In brief, Meso-2 cells were seeded in 6-well plates and allowed to reach 50% confluency. The cells were then washed and incubated for 3–4 hr with pcDNA3.1-Bcl-2-EGFP, pcDNA3.1-Bcl-xL-EGFP, pcDNA3.1-ΔBcl-2-EGFP, or pcDNA3.1-ΔBcl-xL-EGFP, 3 μg each, preincubated at room temperature for 10 min with 10 μl LipofectAmine-2000 (Invitrogen, Mount Waverley, VIC, Australia) and combined with 250 μl OptiMEM (Life Sciences). The cells were then washed, incubated for 24 hr in complete DMEM, after which they were maintained in DMEM supplemented with 10 μg/ml of the selection antibiotic G418 (Sigma, St. Louis, MO). After 4–5 passages, the cells were assessed for apoptotic efficacy on the basis of expression of EGFP using a fluorescence microscopy. In all cases, > 90% of selected cells showed green fluorescence.

Calculation of log p-values

The log p-values for individual compounds were calculated using the CLogP program provided by Daylight Chemical Information Systems (Los Altos, CA).

Results and Discussion

Previous results suggested that modification of several moieties of the prototypic α-TOS modulates its proapoptotic activity. Thus, we have shown that changes in the functional, hydrophobic and signaling domains (Fig. 1b) all contribute to the overall efficacy of the analogues to induce programmed cell death.19 A yet unexplored possibility was to change the ester bond such as to replace it with an amide bond, since it is anticipated that the amides analogues of VE esters may have longer half-life in vivo than their ester counterparts. Before using the amide in preclinical models of cancer, it was essential to investigate their apoptogenic activity in vitro, since induction of apoptosis is a major determinant of a potentially important anticancer agent. This notion is supported by results revealing that the superiority of α-TOS in suppression of cancer in preclinical models over α-TOH (the natural form of VE) is directly related to its capacity to induce apoptosis.28, 29

We thus investigated the apoptogenic propensity of VE amides in several cell lines, including the human T-lymphoma Jurkat cells, the leukemic U937 cells and the malignant mesothelioma Meso-2 cells. Figure 2 clearly demonstrates that all amides were more effective than their ester counterparts, with α-TAM being the most potent of the agents tested. In fact, substantial apoptosis was observed with 50 μM α-TAM in Jurkat cells as early as after 1 hr of treatment (not shown). The order of efficacy of the amides was similar to that found for the esters.19 Importantly, too, apoptogenic activity was completely abrogated by esterification of the free carboxyl group on the functional group of the VE amides, as shown here for α-tocopheryl succinyl amide (α-TAS). These data confirm our previous findings that the functional domain requires a free carboxyl for esters of VE to be apoptogenic.19 Moreover, it is consistent with the premise that agents of VE induce apoptosis due to their amphiphilic nature, i.e., by possession of a long hydrophobic tail (the hydrophobic moiety) and a charged group on the opposite side of the molecule (the functional group).11 Consistent with the apoptosis data are the IC50 values of VE amides (Table I). Again, α-TAM was the most efficient amide with IC50 of ∼2 μM, while the least efficient δ-TAS showed ∼ 10-fold higher IC50 value. Thus, VE amides are apparently very efficient inducers of apoptosis for a variety of malignant cell lines, including lymphoma, leukemia and mesothelioma cell lines (this report) as well as breast cancer cells.30

Figure 2.

Vitamin E amides are superior in apoptosis induction to their ester counterparts for malignant cells. Jurkat (ad), U937 (eh) and Meso-2 cells (il) were seeded and exposed to α-TOS, α-TAS and α-TOSM (a, e, i), δ-TOS and δ-TAS (b, f, j), α-TOM and α-TAM (c, g, k) and δ-TOM and δ-TAM (d, h, l), all at 25 and 50 μM except for δ-TASM (50 μM only), for 6, 12 and 24 hr. The cells were harvested and assessed for apoptosis by flow cytometry using the annexin V-FITC binding method.

Table I. Amides of VE Analogues Feature Lower IC50 Compared to The Corresponding Esters
AnalogueIC50 ± SD (μM)
  1. Meso-2 cells were plated in 96-well plates at 2.5 × 103 per well. The cells were allowed to attach overnight, then incubated for 24 hr with individual VE analogues, dissolved in EtOH, at ≤ 100 μM. Cell viability was determined using the MTT assay,24 as detailed in text. In brief, following exposure of cells, 10 μl of MTT was added, and after incubation for 4 h at 37°C, absorbance was read at 550 nm using an ELISA plate reader and absorbance of control untreated cells was considered as 100%. Survival curves were generated and the IC50 values determined.

α-TOS42.8 ± 6.6
δ-TOS65.4 ± 8.9
α-TOM21.9 ± 5.2
δ-TOM48.7 ± 7.8
α-TAS12.5 ± 3.2
δ-TAS19.8 ± 4.2
α-TAM2.1 ± 0.8
δ-TAM8.7 ± 1.9
α-TASM> 100

It has been shown previously that esters of VE with apoptogenic activity are selective for malignant cells, one of the important feature of these potential anticancer agents. We therefore assessed also their amide counterparts for toxicity toward a cell type that epitomizes normal cells, human fibroblasts. As shown in Figure 3, there was minimal apoptosis caused by the VE amides. Although α-TAM at 50 μM did cause some apoptosis in fibroblasts at 24 hr, it was negligible compared to that caused by the analogue in malignant cells (cf Fig. 2). Thus, amides of VE appear to be selective for malignant cells, as are VE esters.

Figure 3.

Fibroblasts are largely resistant to vitamin E amides. The AG01518 human foreskin fibroblasts were plated and exposed to α-TAS, δ-TAS, α-TAM and δ-TAM, each at 50 μM, for 6, 12 and 24 hr. The cells were harvested and assessed for apoptosis by flow cytometry using the annexin V-FITC binding method.

Our results and those of others strongly implicate mitochondria in apoptosis induction by VE esters.7, 10, 14, 31 We were interested whether the amide analogues follow a similar path. To do this, we investigated 2 aspects of mitochondrial function. First, we studied the effect of VE amides on the mitochondrial inner transmembrane potential (ΔΨm), since ΔΨm as a rule dissipates relatively early during apoptosis, in which the mitochondrial route is the major pathway. Figure 4 reveals that ΔΨm declined when Jurkat cells were treated with VE amides, and this event preceded the onset of apoptosis (cf Fig. 2). The decline of ΔΨm was most pronounced in the case of α-TAM.

Figure 4.

Vitamin E amides cause dissipation of the mitochondrial inner transmembrane potential. Jurkat cells were treated with 50 μM α-TAS, δ-TAS, α-TAM and δ-TAM for 6, 12 and 24 hr, harvested and the percentage of cells with low ΔΨm estimated by flow cytometry following cell staining with the polychromatic probe JC-1.

We next studied the role of mitochondria-stabilizing proteins in apoptosis induced by VE amides. It has been documented earlier that overexpression of the antiapoptotic proteins Bcl-xL and Bcl-2 protects cancer cells from α-TOS.7 To investigate whether similar protection is operational in cells exposed to VE amides, we stably transfected Meso-2 cells with plasmids harboring the Bcl-xL or Bcl-2 genes fused to the gene coding for EGFP. We also transfected the cells with plasmids carrying a gene in which the mitochondrial-docking N-terminus of Bcl-xL and Bcl-2 was deleted, since mitochondrial association of Bcl-xL and, to certain extent, also that of Bcl-2 is a prerequisite for their antiapoptotic effect. Figure 5 documents that overexpression of both antiapoptotic proteins suppressed susceptibility of the mesothelioma cells to VE amides. This was not observed when ΔBcl-xL or ΔBcl-2 was used. These data further document the important role of mitochondria in apoptosis induced by VE amides. To see whether VE amides regulate expression of Bcl-2 family proteins, we assessed the level of expression of Bcl-2, Bcl-xL, Mcl-1 and Bax in Jurkat cells exposed to the VE amides for 6 hr, i.e., time at which lower extent of apoptosis was detected for most of the agents used (cf Fig. 2). Expression of none of the proteins assessed for was significantly altered by the treatment (data not shown), suggesting that VE amides do not induce apoptosis by changing the ratio of the antiapoptotic vs. the proapoptotic Bcl-2 family proteins. These data together with those showing dissipation of ΔΨm strongly suggest mitochondrial involvement in apoptosis signaling of VE amides. This is analogous to the signaling pathways of the corresponding esters, as shown in our previous reports,7, 19 although the effects of the amides are stronger at equimolar concentrations compared to their ester counterparts, consistent with their stronger apoptogenic activity.

Figure 5.

Overexpression of mitochondrial antiapoptotic proteins confers resistance to vitamin E amides. Meso-2 cells were stably transfected with plasmids harboring the Bcl-2-EGFP, Bcl-xL-EGFP, ΔBcl-2-EGFP, or ΔBcl-xL-EGFP genes, or mock-transfected with the empty vector (pcDNA3.1). The cells were seeded and exposed to α-TAS, δ-TAS (both 24 hr), α-TAM (6 hr), or δ-TAM (12 hr), all at 50 μM except for α-TAM (25 μM), and assessed for apoptosis by flow cytometry using the annexin V-FITC binding method.

Our findings clearly show that, in all cases, VE amides were more apoptogenic compared to their ester counterparts. One possibility may be that amides are more stable against hydrolysis than esters. However, this option can be ruled out, since cancer cells have low propensity to hydrolyze esters of VE, as found for both Jurkat and Meso-2 cells (data not shown). Moreover, we observed high level of apoptosis with 50 μM α-TAM in Jurkat cells as early as within 1 hr, which appears too fast for the amide to be hydrolyzed.

As VE analogues need the hydrophobic moiety to induce apoptosis,19 strongly suggesting their association with membranous structures, we reasoned that more avid membrane partitioning may explain the superiority of the amides over their ester counterparts. It has been suggested that partitioning of compounds into the lipid phase may be modulated by modification of their structure,32 and this may determine their biologic effect.33 To learn about lipophilicity of VE analogues, we calculated the log p-values that are indicative of affinity of compounds for lipidic structures. Table II reveals that the log p-values were lower for all amides than those of the corresponding esters, indicating that better partitioning into the lipid phase may be a reason for superiority of the amide analogues of VE in apoptosis induction. In fact, even small differences in log p-values can result in major changes in binding to proteins of mitochondrial origin. Thus, the differences cited in our paper clearly could result in similar differences in receptor binding, wherever they may be within the mitochondrial membrane. The differences between the values for the amides and the corresponding esters (about 0.25 to 1.8) signify differences in partitioning of up to almost 50-fold due to the log scale and therefore may be quite significant. Also, as stated elsewhere,34 optimum log p-values for the water solubility and intestinal membrane partitioning required for oral absorption are typically below 5, although many other factors also come into play. Optimum values for penetration through the lipid stratum corneum of skin are 2.5–6.0 and are therefore similar to those for oral absorption. In reducing the log p-values in the amide isosteres relative to the esters, we have brought these molecules closer to the aforementioned optimum values (although the log p-values are still quite high) for membrane permeation, which may indicate increased cellular penetration. In more general terms, the log p-values reflect not only partitioning into the lipidic phase per se, but also the efficacy of dissolution, oral absorption, penetration into the target tissue and, to some extent, the metabolism of individual compounds,35 and thereby suggest that the VE amides may be also more efficient in vivo.

Table II. Vitamin E Amides Have Lower Log P-Values Than Their Ester Counterparts
  • 1

    The letter “X” stands for O or A, referring to the ester or amide analogues, respectively. The log p-values were calculated using the CLogP program provided by Daylight Chemical Information Systems.


In pharmaceutical chemistry, the replacement of esters with amides is a common bioisosteric substitution that does not drastically change the hydrophilicity of the molecule.36 Our rationale for introducing an amide bond in place of the ester was based on the well-established fact that anilinic amides are much less prone to hydrolysis than the corresponding phenolic esters. Enhancing the stability of these tocopheryl ester derivatives would protect these molecules in vivo, allowing them to stay intact longer, thus increasing their bioavailability. We also reasoned that isosteric replacement of the esters by amides should make them less prone to enzymatic hydrolysis as well. There are several nonspecific esterases in the intestinal mucosal cells and in the blood. In contrast, peptidases exhibit a much narrower specificity. For example, prodrugs with an amino acid in amide linkage are more stable in the intestine and blood than the corresponding ester analogues.37

VE analogues present an intriguing group of agents with proapoptotic and anticancer activity in several preclinical models. These include solid tumors, such as colon,28 breast38 and melanoma cancers,39, 40 as well as colon cancer metastasis.29 Our recent data reveal that α-TOS also efficiently suppresses mesotheliomas, a thus far untreatable type of neoplasia, in a preclinical setting,41 further fueling the potential of this group of compounds. Therefore, it is warranted to design novel analogues of VE that would have enhanced activity against cancer cells, such as are the novel amide analogues, epitomized by the potent α-tocopheryl maleyl amide. We show in this communication that the novel amide analogues of vitamin E are more potent in their apoptogenic activity when compared to their ester counterparts19 while maintaining the selectivity and similar or identical molecular mechanism of induction of cell death, i.e., signaling via the intrinsic mitochondrial pathway. We believe that this communication will lead to synthesis and testing of even more efficient analogues from the vitamin E group.


The authors thank Dr. R.J. Youle for kindly providing the constructs harboring Bcl-xL and Bcl-2 and their deletion mutants. They also thank Pfizer Pharmaceuticals, Inc., for providing a Summer Research Scholarship to J.E. Supported in part by research grants from the Dust Diseases Board of Australia, the Australian Research Council and the Queensland Cancer Fund (to J.N.).