Discussion

Stilbenes, such as resveratrol and pterostilbene, represent biologically active secondary metabolites of plants known to have diverse pharmacological properties, which include cancer-chemopreventive activity. They are also considered to be phytoalexins due to their potent antifungal activities [14-16]. To investigate the molecular pathways affected by pterostilbene in eukaryotes, we conducted DNA microarray analysis using the model yeast S. cerevisiae, and identified large-scale transcriptome alterations associated with pterostilbene exposure in this organism. The responsive genes are associated with a diversity of biological, molecular and cellular processes, suggesting that pterostilbene exerts wide-ranging effects on the physiological and biochemical processes of eukaryotic cells (Tables (Tables22 and and33).
Pterostilbene affected the expression of genes involved in sulfur metabolism and/or sulfur assimilation, associated with the highest z-score in this study (Tables (Tables33 and and6).6). Within this category, six genes (ECM17, MET3, MET14, MET16, MET10, MET6) encoding enzymes involved in methionine biosynthesis were repressed (Figure (Figure2,2, Table Table6).6). Methionine metabolism plays an important role in many aspects of cellular physiology. Its derivative S-adenosyl methionine (SAM) serves as a methyl donor in many biosynthetic processes [40]. Metabolism of sulfur-containing amino acids like methionine has been linked to cell cycle progression, and perturbations of these processes lead to diverse cellular anomalies [41,42]. The down-regulation of genes involved in methionine biosynthesis by pterostilbene may cause cellular stress by not only decreasing methionine levels, but also by compromising the supply of donor methyl groups required for methylation reactions in various biosynthetic pathways. Our results suggest that one of the molecular effects of pterostilbene might involve the disruption of methionine biosynthesis, an observation that has not been previously reported for this compound.
The observed down-regulation of genes involved in methionine metabolism in response to pterostilbene treatment (Figure (Figure2,2, Table Table6)6) is perhaps surprising given that several genes in this pathway are actually up-regulated during conditions of oxidative stress and amino acid starvation [43,44]. However, a transcriptional profiling study conducted in yeast cells exposed to the herbicide sulfometuron methyl (SM), which inhibits branched-chain amino acid biosynthesis, indicated that exposure to SM resulted in the down-regulation of several genes involved in methionine metabolism, including MET3, MET6, MET14, SAM1, and SAM2 [45]. One mechanism proposed for this down-regulation was the potentially reduced levels of ATP in SM-treated cells, given that methionine biosynthesis and the production of SAM are ATP-requiring processes. Interestingly, in the present work pterostilbene treatment altered the expression of more than 100 mitochondrial genes (Additional file 3: op. cit.), suggesting large-scale perturbations in mitochondrial function which would eventually lead to ATP deficiency. A second potential mechanism emerged from queries using MET1, MET3, MET6, MET10, MET13, MET14, and MET16 against the Serial Pattern of Expression Levels Locator (SPELL) database [46], which revealed that all of the corresponding transcripts are down-regulated in response to osmotic stress [47]. Given the significant effects pterostilbene exposure is likely to have on lipid metabolism (Table (Table4),4), it is possible that membrane integrity could be compromised leading to an osmotic imbalance in yeast cells. Consistent with this notion, genes involved in osmotic stress response regulation such as GRE1, GRE2, SSK1, PPZ1, and STE11 were induced in pterostilbene-treated cells (Additional file 3: op. cit.).
Of further significance, the present results show that pterostilbene up-regulated OAF1, which encodes a transcription factor that regulates the expression of genes involved in the beta-oxidation of fatty acids in peroxisomes in yeast cells [28]. In addition, genes encoding enzymes required for fatty acid β-oxidation were also up-regulated by pterostilbene. It has been previously shown that pterostilbene lowers lipid/lipoprotein levels in hypercholestrolemic hamsters through activation of the peroxisome proliferator-activated receptor α (PPARα) [10]. PPARα is involved in fatty acid and lipid metabolism, through the activation of genes involved in fatty acid β-oxidation in the liver, heart, kidney and skeletal muscles [10,48,49]. Thus, the up-regulation of genes involved in fatty acid beta-oxidation by pterostilbene in the present report is consistent with previous observations of its effects on mammalian cells. In addition, pterostilbene also up-regulated several genes involved in sterol, phospholipid and sphingolipid metabolism, including genes involved in the regulation of lipid metabolism. Taken together, these results suggest that lipid metabolism is likely to be an important molecular pathway that is affected by pterostilbene.
Transcript levels of a number of genes involved in the pleiotropic/multiple drug resistance response were also found to increase dramatically following pterostilbene exposure (Table (Table5).5). These genes include ABC transporters, multidrug resistance transcription factors, and other drug-responsive genes. The plasma membrane-associated efflux pumps Pdr5p and Snq2p are under the genetic control of the transcription factors Pdr1p and Pdr3p [50]. In yeast cells exposed to pterostilbene, the transcripts of PDR3, PDR5 and SNQ2 increased 9.5-, 6.1- and 8.2-fold, respectively (Table (Table5).5). In addition, a yeast mutant with a deletion in the PDR3 gene showed strong hypersensitivity to pterostilbene, confirming the importance of this gene in conferring resistance to this compound (Figure (Figure4).4). Notably, the ABC transporter YOR1, involved in mediating the export of many different organic anions including oligomycin, was also up-regulated 10.3-fold in pterostilbene-treated cells. Moreover, the expression of a plasma membrane transporter, encoded by AZR1, which is involved in the resistance to azole drugs, dramatically increased by 73.4-fold following exposure. Taken together, the strong up-regulation of this set of transporters suggests that drug efflux is one of the major mechanisms involved in the cellular detoxification of pterostilbene in yeast cells.
As mentioned above, an additional outcome from the present work is the observation of the up-regulation of a large number of genes involved in diverse mitochondrial functions, including mitochondrial respiration, mitochondrial protein synthesis and mitochondrial protein targeting (associated with a z-score of 6.1 – Tables Tables3,3, and Additional file 3: op. cit.). Furthermore, RTG1 and RTG3, genes encoding transcription factors that play an important role in regulating the communication between the mitochondria and the nucleus in yeast cells, were up-regulated by pterostilbene. Mitochondria are vital for energy production, and their disruption in humans has been implicated in aging, diabetes, heart disease, and various neurodegenerative disorders [51]. Recently, one of the mechanisms by which pterostilbene induces apoptosis in human gastric carcinoma cells was found to involve the activation of the caspase cascade via the mitochondrial pathway [52]. Thus, the induction of mitochondrial genes by pterostilbene in the present study is apparently consistent with its effects on this organelle in human cancer cell lines.
Of the eleven mutants tested carrying deletions in selected genes from pterostilbene-responsive pathways, only the pdr3Δ mutant showed hypersensitivity to pterostilbene (Figure (Figure4).4). Lack of correlation between transcriptional profiling and deletion mutant analyses has been reported in the literature by investigators using both selected individual mutants as well as whole-genome mutant populations [53-57]. Such discrepancies could be attributable to factors such as post-transcriptional regulation of target genes [56], functional redundancy [53], as well as the specific nature of the inhibitor tested (e.g., genes required for survival in the presence of the inhibitor could be distinct from those which are transcriptionally responsive; [55]). Nevertheless, the hypersensitivity to pterostilbene exhibited by the pdr3Δ mutant in the present study, coupled with elevated levels of PDR3 as well as other multidrug resistance-related transcripts (Figure (Figure4,4, Table Table5),5), confirms the importance of efflux-mediated detoxification pathways in the cellular response to this compound. Further mechanistic insights could potentially be gained by extending these analyses to whole-genome mutant collections of S. cerevisiae generated by The Saccharomyces Genome Deletion Project [57], which have been successfully used by others for the identification of molecular targets of various clinically relevant drugs, as well as for the investigation of interactions between the cellular pathways affected [58-61]. Such studies, coupled with transcriptional profiling experiments, could provide further corroborative evidence indicating the relevance of specific genes to the mechanism of action of pterostilbene.
The present study provides an overview of the major transcriptional responses to pterostilbene using the model eukaryote S. cerevisiae. While transcriptional profiling was employed as the principal tool to investigate the molecular effects of pterostilbene, an examination of the response of the yeast proteome would further refine our understanding of pterostilbene-induced gene expression changes, given that transcript and protein levels do not correlate in all cases [62-65]. Additionally, the transcriptional profiling results presented in the current work rely on a single drug concentration (IC50) and time point (one doubling), thus it is likely that many of the transcriptional responses observed represent indirect consequences of drug exposure. An expansion of the current data set to include multiple time points and drug concentrations could therefore prove invaluable in distinguishing between the primary and secondary effects of pterostilbene, such as mechanism-related vs. adaptation-related cellular responses. Nevertheless, the present work yields important clues and provides a foundation for further studies directed towards elucidating the precise mechanism of action of pterostilbene.

Conclusion

We have identified the molecular pathways affected by pterostilbene, and our results show that pterostilbene affects the expression of a diverse group of genes in yeast cells. Using Gene Ontology-based analysis, the most significant effects were observed in genes involved in "methionine metabolism," "response to drug," "transcription factor activity," and "mitochondrion" functions. Additional analyses indicated that many genes involved in lipid metabolism were also affected. The observed response of lipid metabolism genes is in agreement with the known hypolipidemic properties of pterostilbene mediated through the activation of PPARα. The induction of a large number of mitochondrial genes by pterostilbene is consistent with its previously-demonstrated role in apoptosis in human cancer cells. Our data also show that pterostilbene has a significant effect on methionine metabolism, perhaps resulting in the depletion of methionine by the inhibition of methionine biosynthesis. The effect of pterostilbene on methionine metabolism has not been previously observed and merits further investigation.

Competing interests

The author(s) declare that they have no competing interests.

Authors' contributions

ZP conceived and coordinated the study, performed the data analysis and drafted the manuscript. AKA participated in the design of the study, contributed to the data analysis, and helped to draft the manuscript. TX carried out the yeast mutant sensitivity assays and RT-PCR primer design. SRB performed the real-time quantitative RT-PCR experiments. AMR synthesized the compound for this study. QF performed the IC50 determination and the DNA microarray experiments. SOD provided intellectual input and guidance, and assisted with manuscript preparation. All authors read and approved the final manuscript.

Acknowledgements

The authors are grateful to Jamie Lamberton and Melanie Mask for excellent technical assistance. AKA, TX, and QF were supported in part by a grant from the Public Health Service, National Institute of Allergy and Infectious Diseases, Grant No. R01 AI27094, and the USDA Agricultural Research Service Specific Cooperative Agreement No. 58-6408-2-0009.

References