Skip to main page content
Access keys NCBI Homepage MyNCBI Homepage Main Content Main Navigation
. 2016 Feb 17;11(2):e0148042.
doi: 10.1371/journal.pone.0148042. eCollection 2016.

Activation of the Nrf2 Cell Defense Pathway by Ancient Foods: Disease Prevention by Important Molecules and Microbes Lost From the Modern Western Diet

Affiliations
Free PMC article

Activation of the Nrf2 Cell Defense Pathway by Ancient Foods: Disease Prevention by Important Molecules and Microbes Lost From the Modern Western Diet

Donald R Senger et al. PLoS One. .
Free PMC article

Abstract

The Nrf2 (NFE2L2) cell defense pathway protects against oxidative stress and disorders including cancer and neurodegeneration. Although activated modestly by oxidative stress alone, robust activation of the Nrf2 defense mechanism requires the additional presence of co-factors that facilitate electron exchange. Various molecules exhibit this co-factor function, including sulforaphane from cruciferous vegetables. However, natural co-factors that are potent and widely available from dietary sources have not been identified previously. The objectives of this study were to investigate support of the Nrf2 cell defense pathway by the alkyl catechols: 4-methylcatechol, 4-vinylcatechol, and 4-ethylcatechol. These small electrochemicals are naturally available from numerous sources but have not received attention. Findings reported here illustrate that these compounds are indeed potent co-factors for activation of the Nrf2 pathway both in vitro and in vivo. Each strongly supports expression of Nrf2 target genes in a variety of human cell types; and, in addition, 4-ethylcatechol is orally active in mice. Furthermore, findings reported here identify important and previously unrecognized sources of these compounds, arising from biotransformation of common plant compounds by lactobacilli that express phenolic acid decarboxylase. Thus, for example, Lactobacillus plantarum, Lactobacillus brevis, and Lactobacillus collinoides, which are consumed from a diet rich in traditionally fermented foods and beverages, convert common phenolic acids found in fruits and vegetables to 4-vinylcatechol and/or 4-ethylcatechol. In addition, all of the alkyl catechols are found in wood smoke that was used widely for food preservation. Thus, the potentially numerous sources of alkyl catechols in traditional foods suggest that these co-factors were common in ancient diets. However, with radical changes in food preservation, alkyl catechols have been lost from modern foods. The absence of alkyl catechols from the modern Western diet suggests serious negative consequences for Nrf2 cell defense, resulting in reduced protection against multiple chronic diseases associated with oxidative stress.

Conflict of interest statement

Competing Interests: The authors have declared that no competing interests exist.

Figures

Fig 1
Fig 1. Chemical structures of Nrf2 pathway activators.
(A) Well known synthetic and natural Nrf2 activators. (B) The alkyl catechol Nrf2 activators that are the focus of this study.
Fig 2
Fig 2. Induction of Nrf2 target gene mRNAs by alkyl catechols and catechol, as measured with RT-PCR.
Y-axis = (mRNA copies)/(106 18S rRNA copies). Nrf2 target genes = heme oxygenase-1 (HO-1), NAD(P)H:quinone oxidoreductase 1 (NQO1), glucose 6-phosphate dehydrogenase (G6PD). Control non-NRF2 target mRNAs relevant to each cell type also were measured: CD31 (PECAM-1), VE-cadherin (cadherin-5), integrin subunit β1, and E-cadherin (cadherin-1). Test compounds were added to a final concentration of 30 μM and cells harvested at the time indicated. Ctrl = vehicle (H2O) control, Cat = catechol, 4MC = 4-methylcatechol, 4VC = 4-vinylcatechol, 4EC = 4-ethylcatechol. (A) Human dermal microvascular endothelial cells at 4 hours; (B) Human dermal microvascular endothelial cells at 24 hours; (C) Human brain astrocytes at 24 hours; (D) Human dermal keratinocytes at 24 hours. For all panels, error bars = ± standard deviation (S.D.); n ≥ 3 for each data point. Summary of data analyses and statistical significance (see Methods): It should be emphasized here that the Nrf2 pathway is activated primarily by stabilization of Nrf2 protein that allows for transcriptional induction of Nrf2 target genes, such as HO-1, NQO1, and G6PD; and therefore, these target gene mRNAs are indicators of Nrf2 pathway activation. Activation of the Nrf2 pathway is not mediated primarily by induction of Nrf2 mRNA, but Nrf2 mRNA induction may contribute modestly as suggested by data shown here (see text for further explanation and references). (Panel A) For HO-1, individual comparisons between vehicle Ctrl and each of the other experimental conditions = all extremely significant (p<0.0001); for NQO1, G6PD, Nrf2, CD31, and VE-cadherin data sets, differences between vehicle Ctrl and each of the other experimental conditions = all not statistically significant. (Panels B, C, D) For HO-1, NQO1, and G6PD, differences between vehicle Ctrl and each of the other conditions = all extremely significant (p<0.0004 to p<0.0001). In contrast, for non-Nrf2 target gene controls (CD31, VE-cadherin, β1 integrin, and E-cadherin), differences between vehicle Ctrl and each of the other experimental conditions = not statistically significant. For Nrf2 in Panel B, small but statistically significant differences were observed between vehicle Ctrl and 4MC (p < 0.026) and vehicle Ctrl and 4EC (p < 0.047); for Nrf2 in Panel C, small but significant differences were observed between vehicle Ctrl and 4MC (p < 0.01) and vehicle Ctrl and 4EC (p < 0.01); for Nrf2 in Panel D, small but significant differences were observed between vehicle Ctrl and 4MC (p < 0.01) and 4VC (p < 0.02).
Fig 3
Fig 3. Induction of HO-1 protein expression by catechol, akyl catechols, and sulforaphane; inhibition by Nrf2 siRNAs.
(A) Western blotting of human microvascular endothelial cells, harvested 24 hours after adding compounds. Ctrl = control, Cat = catechol, 4MC = 4-methylcatechol, 4VC = 4-vinylcatechol, 4EC = 4- ethylcatechol, 4EP = 4-ethylphenol, SF = sulforaphane. Catechol, 4MC, 4VC, 4EC, and 4EP were each added to a final concentration of 30 μM; sulforaphane was added to a final concentration of 20 μM that is the maximum tolerated dose (higher doses cause cell death). CD31 = protein loading control. (B) Western blotting of human umbilical vein endothelial cells, either untransfected (Ctrl) or transfected with control siRNA (ctrl-siRNA) or Nrf2 siRNAs (Nrf2-siRNA#1; Nrf2-siRNA#2)) and harvested 24 hours after addition of 30 μM 4EC (+4EC), where indicated. CD31 = protein loading control.
Fig 4
Fig 4. Immunohistochemical staining of Nrf2 in human endothelial cells.
Cells were incubated with compounds (30 μM final concentration) for 24 hours, fixed, and stained for Nrf2 (green color) and F-actin (red color). Vehicle ctrl = vehicle control, 4EP = 4-ethylphenol, (negative control), 2M4M = 2-methoxy-4-methylphenol (negative control), CFA = caffeic acid (negative control), 4MC = 4-methylcatechol, 4VC = 4-vinylcatechol, 4EC = 4-ethylcatechol. Note bright green staining of Nrf2 in cells stimulated with catechol, 4MC, 4VC, and 4EC, in comparison with controls. See Figs 1 and 5 for all chemical structures. All samples were processed and stained in parallel; green images (Nrf2) were captured at identical exposure; and, similarly, red images (F-actin) were captured at identical exposure. Subsequently, red and green images were merged without any manipulation so that images presented here are valid for direct comparisons.
Fig 5
Fig 5. Compounds with structural similarity to catechols that do not activate the Nrf2 pathway significantly, in comparison with catechol or akyl catechols.
All compounds depicted here were tested in RT-PCR assays and/or western blotting assays with human endothelial cells, as demonstrated in Figs 2 and 3 at a final concentration of 30 μM, with the exception of quercetin and luteolin that were tested at 20 μM (the maximum tolerated dose). None of these compounds induced Nrf2 target gene expression significantly in comparison with catechol or the akyl catechols. Consistent with these negative findings, each of these compounds has structural characteristics consistent with inactivity, either due to methylation of hydroxyls (top panel), lack of appropriate hydroxyls on benzene ring (middle panel), or electron-withdrawing or bulky side groups appended to the catechol moiety (bottom panel). For supporting data, see Fig 3A (for 4-ethylphenol) and subsequent figures (for caffeic acid, chlorogenic acid, and 3,4-dihydroxybenzoic acid) and also S3 Fig (for all other compounds).
Fig 6
Fig 6. Dose comparisons of catechols and sulforaphane for induction of Nrf2 target gene expression.
RT-PCR analyses; Y-axis = (mRNA copies)/(106 18S rRNA copies). Error bars = ± S.D.; n ≥ 3 for each data point. (A) Human dermal microvascular endothelial cells, 24 hours after stimulation with either 5, 10, 20 μM 4-ethylcatechol (4EC) or 5, 10, 20 μM sulforaphane (SF). (B) Human brain astrocytes stimulated for 24 hours with 4EC or SF, as in Panel A. (C) Human dermal microvascular endothelial cells, 24 hours after stimulation with either 5, 10, 20 μM catechol (Cat) or 5, 10, 20 μM 4-methylcatechol (4MC). (D) Human dermal microvascular endothelial cells stimulated with 30 μM 4EC for 2, 4, 8, or 24 hours. Nrf2 target genes = heme oxygenase-1 (HO-1), NAD(P)H:quinone oxidoreductase 1 (NQO1), glucose 6-phosphate dehydrogenase (G6PD). Control mRNAs relevant to each cell type also were measured: CD31 (PECAM-1), VE-cadherin (cadherin-5), integrin subunit β1. Summary of data analyses and statistical significance: As for Fig 2, it is important to emphasize that the Nrf2 pathway is activated primarily by stabilization of Nrf2 protein that allows for transcriptional induction of Nrf2 target genes, such as HO-1, NQO1, and G6PD; and therefore, these target gene mRNAs are indicators of Nrf2 pathway activation. Activation of the Nrf2 pathway is not mediated primarily by induction of Nrf2 mRNA, but Nrf2 mRNA induction may contribute modestly as suggested by data shown here (see text). (Panel A) For HO-1, individual comparisons between Ctrl and each of the other conditions indicated differences that are all extremely significant (p<0.0002), with the exception of 5 μM 4EC (not significant). For NQO1, Ctrl versus (vs.) each of the other conditions = all extremely significant (p<0.0001). For G6PD, Ctrl vs. 5 μM 4EC = very significant (p<0.002) and Ctrl vs. each of the other conditions = all extremely significant (p<0.0002). For Nrf2, modest but significant differences were observed for Ctrl vs. 20 μM 4EC (significant, p<0.05), and Ctrl vs. 20 μM SF = very significant (p<0.002); however, Ctrl vs. each of the other conditions = all not significant. For CD31 and VE-cadherin control mRNAs, Ctrl vs. each of the other conditions = all not significant. (Panel B) For HO-1, NQO1, and G6PD, Ctrl vs. the each of other conditions = all extremely significant (p<0.0001) with the exception Ctrl vs. 10 μM 4EC (HO-1 data) = very significant (p<0.01). For Nrf2, Ctrl vs. 10 μM, 20 μM 4EC, and 10 μM SF = all significant (p<0.03); and Ctrl vs. 20 μM SF = extremely significant (p<0.001). For β1 integrin, Ctrl vs. 10 μM 4EC and Ctrl vs. 20 μM 4EC = not significant, and Ctrl vs. 10 μM SF and Ctrl vs. 20 μM SF = significant (p<0.05). (Panel C) For HO-1, NQO1, and G6PD, individual comparisons for Ctrl vs. the other conditions indicated differences that are all extremely significant (p<0.0005 to p<0.0001), with the exception of Ctrl vs. 5 μM 4MC (HO1 data) = very significant (p<0.01). For Nrf2, Ctrl vs. 5 μM Cat and Ctrl vs. 5 μM 4MC = not significant; Ctrl vs. 10 μM Cat, Ctrl vs. 20 μM Cat, Ctrl vs. 10 μM 4MC = all significant (p<0.05); Ctrl vs. 20 μM 4MC = extremely significant (p<0.001). For CD31 and VE-cadherin, Ctrl vs. other experimental conditions = all not significant, with the exception of Ctrl vs. 20 μM 4MC (VE-cadherin data) = significant (p<0.05). (Panel D) For HO-1, NQO1, and G6PD, individual comparisons for Ctrl vs. the other conditions = all extremely significant (p<0.0001), with the exception of Ctrl vs. 2h and Ctrl vs. 4h (NQO1 data) = very significant (p<0.01), Ctrl vs. 4h and Ctrl vs. 8h (G6PD data) = significant, and Ctrl vs. 2h (G6PD data) = not significant. For Nrf2, Ctrl vs. 2h = not significant; Ctrl vs. 4h, 8h, and 24h = all significant (p<0.05). For CD31 and VE-cadherin, Ctrl vs. other experimental conditions = all not significant.
Fig 7
Fig 7. Activation of the Nrf2 pathway by alkyl catechols and catechol is regulated by oxygen.
Human microvascular endothelial cells were cultured in 21% oxygen (room air) or 2% oxygen (hypoxia), as indicated, and stimulated with the specified compounds (20 μM each) for 24 hours. RT-PCR, as above, was used to quantify mRNAs; Y-axis = (mRNA copies)/(106 18S rRNA copies). Cat = catechol, 4MC = 4-methylcatechol, 4EC = 4-ethylcatechol, HQ = hydroquinone, TBHQ = tert-butylhydroquinone (HQ and TBHQ are well known activators of Nrf2, see Fig 1 for chemical structures). Nrf2 target genes = HO-1, NQO1, G6PD; CD31 (PECAM-1) = internal control; GLUT1 (glucose transporter 1) is induced by hypoxia and serves as a positive control for hypoxia-induced gene expression. Error bars = ± S.D.; n ≥ 3 for each data point. Summary of data analyses and statistical significance: Again, it is important to emphasize that the Nrf2 pathway is activated primarily by stabilization of Nrf2 protein that allows for transcriptional induction of Nrf2 target genes, such as HO-1, NQO1, and G6PD; and therefore, these target gene mRNAs are indicators of Nrf2 pathway activation. Activation of the Nrf2 pathway is not mediated primarily by induction of Nrf2 mRNA, but Nrf2 mRNA induction may contribute modestly as suggested by data shown here in the Nrf2 data panel. Thus, for the HO-1 and NQO1 data sets, representing activation of the Nrf2 pathway, individual comparisons between a specific compound (i.e. Cat, 4MC, 4EC, HQ, or TBHQ), used at 21% oxygen vs. 2% oxygen, indicated oxygen-dependent differences that are all extremely significant (p<0.001). For the G6PD data set, also representing activation of the Nrf2 pathway, individual comparisons between a specific compound used at 21% oxygen vs. 2% oxygen indicated oxygen-dependent differences that are all very significant (p<0.01) with the exception of Cat (p<0.05, significant). Also, for HO-1 and NQO1, additional comparisons for each of the compounds vs. corresponding Ctrl = extremely significant (p<0.0001) for both 21% oxygen and 2% oxygen. For the G6PD data panel and for both 21% oxygen and 2% oxygen: Ctrl vs. Cat (p<0.005, very significant); Ctrl vs. 4MC, Ctrl vs. 4EC, Ctrl vs. HQ, Ctrl vs. TBHQ (all p<0.0002, extremely significant). For Nrf2, and for both 21% oxygen and 2% oxygen data sets: Ctrl vs. each of the compounds = significant (p≤0.03), with the exception of Cat (21% oxygen) = not significant. For CD31, and for both 21% oxygen and 2% oxygen data sets: Ctrl vs. each of the compounds = not significant, with the exception of Ctrl vs. 4EC and Ctrl vs. TBHQ (21% oxygen) = very significant (p<0.01). Nonetheless, these differences are relatively small in comparison with the large inductions of Nrf2 target gene expression shown in HO-1 and NQO1 data panels. Finally, for the 2% oxygen GLUT1 panel, representing induction of GLUT1 by hypoxia, individual comparisons between Ctrl and each of the compounds indicated differences that were all very significant (p<0.01 to p<0.001). Also, individual comparisons between each experimental group in 2% oxygen with the corresponding experimental group in 21% oxygen indicated differences that were all extremely significant (p<0.0001).
Fig 8
Fig 8. Induction of Nrf2 target genes in mice by 4-methylcatechol and 4-ethylcatechol.
RT-PCR analyses; Y-axis = (mRNA copies)/(106 18S rRNA copies). Error bars = ± S.D.; n ≥ 4 for each data point. (Panel A, Kidney) Mice received either intraperitoneal (i.p.) injection or oral gavage of vehicle control (Ctrl), 4-methylcatechol (4MC), or 4-ethylcatechol (4EC) at time zero and again at time zero + 2 hours, for at total dose of 50 mg/kg. Mice were harvested at time zero + 4.5 hours and kidneys dissected for RNA isolation. (Panel B, Lung) 4MC or 4EC, as indicated, were administered i.p. at time zero and again time zero + 2 hours for a total of 50 mg/kg. Mice were harvested at time zero + 4.5 hours, and lungs dissected for RNA isolation. Summary of data analyses and statistical significance: Again, it is important to emphasize that the Nrf2 pathway is activated primarily by stabilization of Nrf2 protein that allows for transcriptional induction of Nrf2 target genes, such as HO-1 and NQO1; and therefore, that robust induction of HO-1 and NQO1 mRNAs, rather than Nrf2 mRNA, indicate Nrf2 pathway activation (see text). (Panel A) Inductions of HO-1 and NQO1 mRNAs by 4MC and 4EC, administered either i.p. or orally, were all statistically extremely significant (p < 0.001). For Nrf2: Ctrl vs. 4MC (i.p.) = significant (p<0.05); Ctrl vs. 4EC (i.p.) and Ctrl vs. 4EC (oral) = no significant differences. For β-actin: no statistically significant differences within experimental groups. (Panel B) For HO-1, Ctrl vs. 4MC = significant (p = 0.0176), Ctrl vs. 4EC = very significant (p = 0.0025). For NQO1, Ctrl vs. 4MC = significant (p<0.045) and Ctrl vs. 4EC = very significant (p = 0.0031). For Nrf2: Ctrl vs. 4MC and Ctrl vs. 4EC = no significant differences. For β-actin: Ctrl vs. 4MC = very significant (p<0.01) and Ctrl vs. 4EC = significant (p<0.04). Thus, β-actin mRNA was reduced by 4MC and 4EC in lung, in contrast to HO-1 and NQO1 mRNAs that were increased. However, 4MC and 4EC did not reduce β-actin in kidney (Panel A).
Fig 9
Fig 9. Model for bioconversion of inactive dietary precursors to Nrf2 activators by phenolic acid decarboxylase (PAD).
The microbial enzyme, PAD, expressed by Lactobacillus plantarum, Lactobacillus brevis, and other, but not all, lactobacillus strains convert caffeic acid (inactive) to 4-vinylcatechol (Nrf2 activator). Similarly, PAD converts 3,4-dihydroxybenzoic acid (inactive) to catechol (Nrf2 activator). See text for references and subsequent figures for supporting data.
Fig 10
Fig 10. Biotransformation of caffeic acid by Lactobacillus plantarum and Lactobacillus brevis, as demonstrated with RT-PCR.
Y-axis = (mRNA copies)/(106 18S rRNA copies). Human dermal microvascular endothelial cells, 24 hours after addition of test samples: Ctrl = control, CFA = caffeic acid, LP = control supernatant from L. plantarum incubated with PBS-glucose and filter-sterilized, (LP + CFA) = supernatant from L. plantarum incubated with CFA in PBS-glucose and filter-sterilized, LB = control supernatant from L. brevis incubated with PBS-glucose and filter-sterilized, (LB + CFA) = supernatant from L. brevis incubated with CFA in PBS-glucose and filter-sterilized. CFA and lactobacillus-incubations with CFA were added to a final concentration corresponding to 30 μM CFA starting material (see Methods). 4EC = 4-ethylcatechol positive control (30 μM). Nrf2 target genes = HO-1, NQO1, G6PD. Control mRNAs = CD31 and VE-cadherin. Error bars = ± S.D.; n ≥ 3 for each data point. Summary of data analyses and statistical significance: As described in previous figures and in the text, induction of the Nrf2 target genes HO-1, NQO1, and G6PD, rather than induction of Nrf2 mRNA, indicates activation of the Nrf2 pathway. Only LP+CFA, LB+CFA, and 4EC (positive control) demonstrated Nrf2 pathway activation by these criteria. For HO-1, NQO1, and G6PD data panels, individual comparisons between Ctrl (or CFA) vs. LP+CFA, Ctrl (or CFA) vs. LB+CFA, and Ctrl (or CFA) vs. 4EC indicated differences that are all extremely significant (p< 0.0001). In contrast, for the HO-1, NQO1, and G6PD panels, individual comparisons between Ctrl vs. CFA, Ctrl vs. LP, and Ctrl vs. LB indicated no significant differences. For Nrf2, Ctrl vs. CFA = small but significant difference (p<0.05), Ctrl vs. LP = not significant, Ctrl vs. LP+CFA = very significant (p<0.01), Ctrl vs. LB = not significant, Ctrl vs. LB+CFA = significant (p<0.05), Ctrl vs. 4EC = significant (p<0.05). Finally, for CD31 and VE-cadherin data sets, statistical analyses indicated no significant differences.
Fig 11
Fig 11. Biotransformation of caffeic acid by L. plantarum and L. brevis, as demonstrated with HPLC.
Y-axis = absorbance at 254nm (mAU), X-axis = minutes. Top panel: HPLC of caffeic acid and 4-vinylcatechol standards. Middle panel: HPLC of supernatant from caffeic acid + L. plantarum incubation, consistent with conversion of caffeic acid to 4-vinylcatechol. Bottom panel: HPLC of supernatant from caffeic acid + L. brevis incubation, consistent with conversion of caffeic acid to 4-vinylcatechol. Retention times: caffeic acid = 8.1 minutes, 4-vinylcatechol = 10.7 minutes.
Fig 12
Fig 12. Biotransformation of 3,4-dihydroxybenzoic acid by Lactobacillus plantarum, as demonstrated with RT-PCR.
Y-axis = (mRNA copies)/(106 18S rRNA copies). Human dermal microvascular endothelial cells, 24 hours after addition of test samples: Ctrl = control, 3,4-DHBA = 3,4-dihydroxybenzoic acid, LP = control supernatant from L. plantarum incubated with PBS-glucose and filter-sterilized, (LP + 3,4-DHBA) = supernatant from L. plantarum incubated with 3,4-DHBA in PBS-glucose and filter-sterilized. 3,4-DHBA and lactobacillus-incubations with 3,4-DHBA were added to a final concentration corresponding to 30 μM 3,4-DHBA starting material (see Methods). Cat = catechol positive control (30 μM). Nrf2 target genes = HO-1, NQO1, G6PD. Control mRNAs = CD31 and VE-cadherin. Error bars = ± S.D.; n ≥ 3 for each data point. Summary of data analyses and statistical significance: As described in previous figures and in the text, induction of the Nrf2 target genes HO-1, NQO1, and G6PD, rather than induction of Nrf2 mRNA, indicates activation of the Nrf2 pathway. For HO-1 and NQO1 data panels, individual comparisons for Ctrl (or 3,4-DHBA) vs. LP+3,4-DHBA and Ctrl (or 3,4-DHBA) vs. Cat indicated differences that are extremely statistically significant (p< 0.0001). For the G6PD data panel, individual comparisons for Ctrl vs. LP+3,4-DHBA and Ctrl vs. Cat indicated differences that are also extremely significant (p< 0.0005). In contrast, for the HO-1, NQO1, and G6PD panels, individual comparisons for Ctrl vs. 3,4-DHBA and Ctrl vs. LP indicated no significant differences. For Nrf2, only Ctrl vs. Cat was statistically significant (p<0.05). For CD31: no statistically significant differences. For VE-cadherin, only Ctrl vs. 3,4-DHBA was statistically significant (p<0.05).
Fig 13
Fig 13. Biotransformation of 3,4-dihydroxybenzoic acid by L. plantarum, as demonstrated with HPLC.
Y-axis = absorbance at 254nm (mAU), X-axis = minutes. Top panel: HPLC of 3,4-dihydroxybenzoic acid standard. Middle panel: HPLC of catechol standard. Bottom panel: HPLC of supernatant from 3,4-dihydroxybenzoic acid + L. plantarum incubation, consistent with conversion of 3,4-dihydroxybenzoic acid to catechol. Retention times: 3,4-dihydroxybenzoic acid = 5.5 minutes, catechol = 6.0 minutes.
Fig 14
Fig 14. Biotransformation of caffeic acid and 3,4-dihydroxybenzoic acid, as demonstrated with western blotting.
(A) Biotransformation by L. plantarum as demonstrated with western blotting of human dermal microvascular endothelial cells, harvested 24 hours after addition of test samples. Blots were stained for the Nrf2 target gene HO-1, Nrf2 itself, and CD31 as loading control. Key: Ctrl = control, CFA = caffeic acid, LP = control supernatant from L. plantarum incubated with PBS-glucose and filter-sterilized, (LP + CFA) = supernatant from L. plantarum incubated with CFA in PBS-glucose and filter-sterilized, 3,4-DHBA = 3,4-dihydroxybenzoic acid, (LP + 3,4-DHBA) = supernatant from L. plantarum + 3,4-DHBA incubated in PBS-glucose and filter-sterilized. CFA, 3,4-DHBA, and lactobacillus-incubations with each were added to a final concentration corresponding to 30 μM CFA and 30 μM 3,4-DHBA starting material (see Methods). Positive controls = catechol (30 μM) and 4-ethylcatechol (4EC, 30 μM). (B) Biotransformation by L. brevis (LB), with experimental conditions otherwise identical to panel (A), above. Also, for experiment shown in panel (B), 4EC = 4-ethylcatechol positive control was added to a final concentration of 15 μM instead of 30 μM.
Fig 15
Fig 15. Model for multi-step bioconversion of inactive dietary precursors to Nrf2 activators by Lactobacillus collinoides.
Microbial cinnamoyl esterase from L. collinoides converts chlorogenic acid (inactive) to caffeic acid (inactive), thereby providing substrate for phenolic acid decarboxylase (PAD)-mediated generation of 4-vinyl catechol (Nrf2 activator). Finally, microbial phenolic acid reductase, also expressed by L. collinoides, reduces 4-vinylcatechol to 4-ethylcatechol (Nrf2 activator). See text for supporting references and subsequent figures for supporting data.
Fig 16
Fig 16. Biotransformation of chlorogenic acid and caffeic acid by Lactobacillus collinoides, as measured with RT-PCR.
Y-axis = (mRNA copies)/(106 18S rRNA copies). RNA was isolated from human dermal microvascular endothelial cells, 24 hours after addition of test samples: Ctrl = control, CA = chlorogenic acid, CFA = caffeic acid, LC = control supernatant from L. collinoides incubated with PBS-glucose and filter-sterilized, (LC + CA) = supernatant from L. collinoides incubated with CA in PBS-glucose and filter-sterilized, (LC + CFA) = supernatant from L. collinoides incubated with CFA in PBS-glucose and filter-sterilized. CA, CFA, and L. collinoides incubations with each were added to a final concentration corresponding to 30 μM CA and 30 μM CFA starting material (see Methods). (4EC) = 4-ethylcatechol positive control (30 μM). Nrf2 target genes = HO-1, NQO1, G6PD; control mRNAs = CD31 and VE-cadherin. Error bars = ± S.D.; n ≥ 3 for each data point. Summary of data analyses and statistical significance: As discussed previously, the Nrf2 pathway is activated primarily by stabilization of Nrf2 protein that allows for transcriptional induction of Nrf2 target genes, such as HO-1, NQO1, and G6PD; and therefore, these target gene mRNAs are indicators of Nrf2 pathway activation. Activation of the Nrf2 pathway is not mediated primarily by induction of Nrf2 mRNA, but Nrf2 mRNA induction may contribute modestly as suggested by data shown here (see text for further explanation and references). Thus, for HO-1, NQO1 and G6PD data panels, individual comparisons for Ctrl vs. LC+CA, Ctrl vs. LC+CFA, and Ctrl vs. 4EC indicated differences that are all extremely statistically significant (p< 0.0001). In contrast, individual comparisons for Ctrl vs. CA, Ctrl vs. CFA, and Ctrl vs. LC indicated no significant differences. For Nrf2, Ctrl vs. CA, Ctrl vs. CFA, and Ctr vs. LC = no significant differences; Ctrl vs. LC+CA, Ctrl vs. LC+CFA, and Ctrl vs. 4EC = all very significant differences (p<0.01). For CD31 and VE-cadherin: no statistically significant differences.
Fig 17
Fig 17. Biotransformation of chlorogenic acid by L. collinoides, as demonstrated with HPLC.
Y-axis = absorbance at 254nm (mAU), X-axis = minutes. Top panel: HPLC of chlorogenic acid standard. Middle panel: HPLC of 4-vinylcatechol and 4-ethylcatechol standards. Bottom panel: HPLC of supernatant from chlorogenic acid + L. collinoides incubation, consistent with conversion of chlorogenic acid to 4-ethylcatechol. Retention times: chlorogenic acid = 7.1 minutes, 4-vinylcatechol = 10.7 minutes, 4-ethylcatechol = 11.6 minutes.
Fig 18
Fig 18. Biotransformation of caffeic acid by L. collinoides, as demonstrated with HPLC.
Y-axis = absorbance at 254nm (mAU), X-axis = minutes. Top panel: HPLC of caffeic acid standard. Middle panel: HPLC of 4-vinylcatechol and 4-ethylcatechol standards. Bottom panel: HPLC of supernatant from caffeic acid + L. collinoides incubation, consistent with bioconversion of caffeic acid to 4-ethylcatechol. Retention times: caffeic acid = 8.1 minutes, 4-vinylcatechol = 10.7 minutes, 4-ethylcatchol = 11.6 minutes.

Similar articles

See all similar articles

Cited by 12 articles

See all "Cited by" articles

References

    1. Hallowell B. Free Radicals in Biology and Medicine, 4th edition Oxford, UK: Oxford University Press; 2007.
    1. Roberts RA, Laskin DL, Smith CV, Robertson FM, Allen EM, Doorn JA, et al. Nitrative and oxidative stress in toxicology and disease. Toxicol Sci. 2009;112(1):4–16. 10.1093/toxsci/kfp179 - DOI - PMC - PubMed
    1. Dalleau S, Baradat M, Gueraud F, Huc L. Cell death and diseases related to oxidative stress: 4-hydroxynonenal (HNE) in the balance. Cell Death Differ. 2013;20(12):1615–30. 10.1038/cdd.2013.138 - DOI - PMC - PubMed
    1. Halliwell B. Oxidative stress and cancer: have we moved forward? Biochem J. 2007;401(1):1–11. 10.1042/BJ20061131 . - DOI - PubMed
    1. Wells PG, McCallum GP, Chen CS, Henderson JT, Lee CJ, Perstin J, et al. Oxidative stress in developmental origins of disease: teratogenesis, neurodevelopmental deficits, and cancer. Toxicol Sci. 2009;108(1):4–18. 10.1093/toxsci/kfn263 . - DOI - PubMed

Publication types

MeSH terms

Feedback