Berberine Derivatives as Pseudomonas aeruginosa ...

Author: Helen

May. 06, 2024

Berberine Derivatives as Pseudomonas aeruginosa ...

For more information, please visit our website.

Abstract

The natural alkaloid berberine has been demonstrated to inhibit the Pseudomonas aeruginosa multidrug efflux system MexXY-OprM, which is responsible for tobramycin extrusion by binding the inner membrane transporter MexY. To find a structure with improved inhibitory activity, we compared by molecular dynamics investigations the binding affinity of berberine and three aromatic substituents towards the three polymorphic sequences of MexY found in P. aeruginosa (PAO1, PA7, and PA14). The synergy of the combinations of berberine or berberine derivatives/tobramycin against the same strains was then evaluated by checkerboard and time-kill assays. The in silico analysis evidenced different binding modes depending on both the structure of the berberine derivative and the specific MexY polymorphism. In vitro assays showed an evident MIC reduction (32-fold and 16-fold, respectively) and a 2–3 log greater killing effect after 2 h of exposure to the combinations of 13-(2-methylbenzyl)- and 13-(4-methylbenzyl)-berberine with tobramycin against the tobramycin-resistant strain PA7, a milder synergy (a 4-fold MIC reduction) against PAO1 and PA14, and no synergy against the ΔmexXY strain K1525, confirming the MexY-specific binding and the computational results. These berberine derivatives could thus be considered new hit compounds to select more effective berberine substitutions and their common path of interaction with MexY as the starting point for the rational design of novel MexXY-OprM inhibitors.

Keywords:

efflux pump inhibitors, Pseudomonas aeruginosa, berberine derivatives, molecular modeling, multidrug resistance

1. Introduction

Pseudomonas aeruginosa is an opportunistic pathogen and a frequent cause of life-threatening infections in high-risk patients [1,2,3], including those with a compromised immune system due to underlying diseases such as cancer or cystic fibrosis [4]. P. aeruginosa infections are difficult to eradicate due to the large number of already-known intrinsic and acquired antibiotic resistance mechanisms [5]. In particular, the chromosomally encoded efflux pumps are responsible for the multidrug-resistant (MDR) phenotype and their overexpression largely contributes to drug tolerance and failure of antibiotic treatment [5,6]. MexXY-OprM, belonging to the resistance–nodulation–cell division (RND) family, is considered the main resistance mechanism towards aminoglycosides in P. aeruginosa; mutations in its regulatory gene mexZ [6] can lead to overexpression and hamper antibiotic treatments. This efflux pump also represents the best example of adaptive resistance, showing a transient hyperexpression in the presence of antibiotics and a basal expression when the drug is removed [7,8].

The MexXY-OprM system extrudes hydrophilic compounds, including aminoglycosides such as tobramycin. This substrate specificity is not shared by the better-known MexAB-OprM efflux pump due to the different aminoacidic composition of the substrate binding pockets, which are located in the inner membrane transporters MexY and MexB, respectively [9]. In previous studies, we optimized a full 3D model of the MexY structure using the highly similar models of the membrane channels P. aeruginosa MexB and Escherichia coli AcrB [9,10,11]. From a structural point of view, MexXY-OprM is a tripartite system characterized by a drug extrusion mechanism via a proton gradient as in all RND efflux systems [12]. MexY is the inner membrane protein, organized as a homotrimer presenting a transmembrane (TM) domain, with 12 alpha helices and a periplasmic domain, which is directly involved in the access, binding, and extrusion of substrates. The periplasmic domain further includes a porter domain (PD) and a TolC domain [13]. The PD is characterized by four other subdomains called PC1, PC2, PN1, and PN2, while the TolC domain is split into two subdomains named DC and DN ( ). The substrate extrusion pathway is characterized by a functional rotation and a cycling conformational change in each protomer of the homotrimer [14] in three different states: loose (for the access stage), tight (for the binding stage), and open (for the extrusion stage). The PD of each protomer encloses two substrate binding pockets, the access or proximal (AP) and the binding or distant (DP) one. Both of them are involved in ligands’ translocation through a peristaltic motion as suggested for AcrB [15]. These two binding sites are separated by a G-loop (not shown in ), which regulates the passage between these two pockets [16]. Three different pathways can be described for the access of the substrates to the MexY transporter (channel CH) in analogy to AcrB EP [17]: (i) the entrance above TM7/TM8 helices from the outer leaflet of the inner membrane to the proximal and binding pocket (CH1); (ii) periplasmic, through the cleft via hydrophilic compounds between subdomains PC1 and PC2 (CH2); and (iii) through the vestibule between protomers into the central cavity (CH3) ( ).

Open in a separate window

The MexY inner channel of different P. aeruginosa strains is characterized by extensive sequence polymorphisms, which can affect the binding modes of both substrates and efflux pump inhibitors (EPIs). These polymorphisms can explain the differences in the ability of MexY to extrude EPIs, and the study of the binding affinity in different bacterial strains can help to identify new molecules to be used as efficient EPIs.

Although the MexY inhibitory activity of the natural alkaloid berberine was extensively assessed by in silico modeling [11,12,18], the association between berberine and tobramycin resulted in a strain-dependent variable behavior [11]. Previous studies showed that the introduction of an aromatic substituent in position 13 can improve both the synergism of the combination berberine–fluconazole against Candida albicans [19] ( ) and the antimicrobial activity of the alkaloid itself against multidrug-resistant (MDR) Staphylococcus aureus [20].

Open in a separate window

Starting from this evidence, we decided to assess the synergism of the three aromatic alkaloid derivatives with tobramycin [19,20]. Thus, we evaluated berberine and its derivatives’ potential activity in silico considering both affinity (free binding energy) and binding specificity through molecular docking and molecular dynamic simulations and carrying out in vitro microbiological assays. Moreover, we analyzed the aminoacidic sequence of MexY of the three reference P. aeruginosa strains PAO1, PA7, and PA14 to assess the involvement of specific polymorphisms in the synergistic effect of the combinations of the tested compounds with tobramycin.

2. Materials and Methods

2.1. Synthesis of Berberine Derivatives:

The berberine derivatives ( ) were synthesized according to Kotani et al. [18] with minor changes. Briefly, after dropwise addition of the appropriate benzyl bromide (1.0 mmol) to a dihydroberberine (1 equiv) and KI (2 equiv) CH3CN (40 mL) solution, the reaction mixture was refluxed under stirring for 4 h. After filtration and solvent evaporation, the crude residue was purified by silica gel column chromatography using CHCl3/CH3OH (50:1) as the eluent. The characterization data of the compounds obtained were identical to those given in the literature [18,20].

2.2. Computational Methods

2.2.1. Sequence Alignment and 3D Modeling for the Polymorphic MexY

The MexY aminoacidic sequence alignment was performed using the multiple sequence alignment method (MAFFT) [21] with default parameters. The aminoacidic sequence of each MexY protein was retrieved in FASTA format from the NCBI protein database. The 3D models of the three polymorphic MexY proteins belonging to P. aeruginosa PAO1 (NCBI code {"type":"entrez-protein","attrs":{"text":"BAA34300.1","term_id":"3868984"}}BAA34300.1), PA14 (NCBI code {"type":"entrez-protein","attrs":{"text":"ABR84278.1","term_id":"150962253"}}ABR84278.1), and PA7 (NCBI code {"type":"entrez-protein","attrs":{"text":"QDL65075.1","term_id":"1702237932"}}QDL65075.1) were constructed following the comparative molecular modeling approach [2,11] using a X-ray crystallographic protein structure of the MexAB-OprM inner transporter as the template (pdb code: 2V50), since it shares the structural function with an identity percentage of 49.95% (PA7), 50.05% (PA14), and 47.17% (PAO1). The three models were built using the Swiss Model server [22] and, after generation, each new 3D protein structure was minimized in a homotrimeric system within the membrane of 500 1-palmitoyl-2-oleoyl-sn-glycero-3-phosphocholine lipids (POPC), thus reaching the nearest local minimum conformational energy on the potential energy surface (PES). This was performed using cycles of steepest descent and conjugate gradient algorithms implemented in the GROMACS v.2020.6 software [23,24] until the system converged. For each minimization step, the Force f and energy potential V of the molecular conformation were also computed. The models were validated using the protocol already reported in our previous work [11,25].

2.2.2. Molecular Docking

The molecular docking simulation was performed using Autodock 4.2.1 (AD4) [26]. Each minimized protein was retrieved in its monomeric form, adding hydrogens to the structure, converting each pdb file into a pdbqt file, and coordinating it to include the calculated charges. The same approach was used for the ligands’ preparation, and, for each ligand–protein complex, the calculation of the electrostatic potential grid was performed with the Autogrid tool, setting a grid box with 50 Å3 focusing near the cleft site of the MexY protein, including PC1 and PC2 subdomains and two periplasmic loops of TM7 and TM8 helices. The genetic algorithm (GA) was used for the pose generations of each ligand, and the AMBER force-field-based scoring function was used for energy calculations as implemented in the docking software. The number of independent AD4 GA runs was increased up to 100, and the grid spacing was kept at 0.375.

2.2.3. Molecular Dynamics

The MexY models together with berberine and its derivatives were oriented in the membrane through the OPM server [27]. The CHARMM GUI [28,29,30] was used to build a membrane bilayer system composed of 500 POPC molecules. MexY trimers with the docked ligands were generated properly surrounded by the lipid matrix. All the built models were appropriately solvated with water (about 10,000 molecules) and ions (to reach up to 0.15 M NaCl, adding 397 Na ions and 361 Cl ions to balance the trimer charge). CHARMM36 force field parameters [31] were used for all MD simulations together with the TIP3P [32] model for the solvent as implemented in GROMACS [23,24]. Berberine and its derivatives were parametrized within the CHARMM-GUI Ligand Reader and Modeler tools for the chosen force field [33] in their positive charged state ( ). After the model minimization, six equilibration phases and MD simulations were carried out. The overall time of each MD simulation was settled to 50 ns, with a time-step of 0.002 ps. Periodic boundary conditions (PBCs) were applied in all directions using a neighbor searching grid type and setting at 1.4 nm the cut-off distance for the short-range neighbor list. Electrostatic interactions were taken into account by implementing a fast and smooth Particle-Mesh Ewald algorithm, with a 1.4 nm distance for the Coulomb cut-off [34].

2.3. Microbioloogical Methods

2.3.1. Bacterial Strains, Plasmids, Media, and Chemicals

The P. aeruginosa laboratory strain K767 (PAO1), its ΔmexXY mutant K1525, and the plasmid pYM004, containing a copy of the mexXY operon, were kindly provided by Prof. Keith Poole (Queen’s University, Kingston, ON, Canada); the mutant strain was complemented by transferring the pYM004 plasmid by electroporation and selected on Luria–Bertani (LB) agar plates containing 200 µg/mL carbenicillin. P. aeruginosa PA14 was kindly provided by Prof. Olivier Jousson (Integrated Biology Center, University of Trento, Trento, Italy). The P. aeruginosa PA7 strain belongs to the Microbiology section strain collection of the Department of Biomolecular Sciences, University of Urbino “Carlo Bo”, Urbino, Italy, while P. aeruginosa ATCC 27853 belongs to the Microbiology section strain collection of the Department of Life and Environmental Sciences—Microbiology section, Polytechnic University of Marche, Ancona, Italy.

Carbenicillin and tobramycin were purchased from Sigma-Aldrich SRL (Milano, Italy), and all microbiological media were purchased from Oxoid SpA (Milano, Italy).

2.3.2. Antimicrobial Combination Assays

Antimicrobial Susceptibility and Checkerboard Assays

P. aeruginosa’s antimicrobial susceptibility towards tobramycin and berberine derivatives was assessed by determining the Minimal Inhibitory Concentration (MIC) according to CLSI guidelines (2017) [35]. For the derivatives, a concentration of 320 µg/mL was selected as the highest, in order not to exceed 1.5% DMSO. The solvent itself was tested as well. Checkerboard assays were performed as previously described [11,36] using 2-fold dilutions of tobramycin (from 256 to 0.25 µg/mL when testing P. aeruginosa PA7 and from 16 to 0.01 µg/mL when testing P. aeruginosa PAO1 and PA14) and of berberine derivatives (from 320 to 10 µg/mL). A four-fold reduction of tobramycin’s MIC was considered indicative of synergism [11]. The reduction in tobramycin’s MIC in the presence of 80 µg/mL berberine was used as a reference value.

Time Killing Curves

Time killing assays were performed as previously described [11,36] using tobramycin concentrations corresponding to 1/2× the MIC, 1× the MIC, and 2× the MIC, alone or in combination with the different berberine derivatives at the concentration that was more synergic than berberine’s one in checkerboard assays. The survivor amount was determined by a plate count on LB agar after 0, 2, 4, 6, 8, and 24 h from drug(s) exposure. Each count was the average of the results of two technical duplicates.

3. Results

3.1. Computational Results

3.1.1. MexY Polymorphisms in PAO1, PA7, and PA14

The MexY sequences of P. aeruginosa PAO1, PA7, and PA14 were aligned, and an identity matrix percentage is reported ( ).

Table 1

MexY-PA14MexY-PA7MexY-PAO1MexY-PA14100.0096.4693.21MexY-PA796.46100.0090.62MexY-PAO193.2190.62100.00Open in a separate window

To evaluate the aminoacidic variations between these sequences, a multisequence alignment with the MAFFT method was performed considering the PAO1–MexY complex as the reference sequence ( ). Several aminoacidic substitutions were found in P. aeruginosa PA7 and PA14. Some of them (highlighted in red in ) were similar in both strains, while others were strain-specific (reported in green for P. aeruginosa PA7 and in blue for P. aeruginosa PA14).

Open in a separate window

Proline 862 of MexY–PAO1 is lacking in both MexY–PA7 and MexY–PA14, while residues 800, 858, 1037 (1036 in PA7, PA14), and 1040 (1039 in PA7, PA14) (marked in aquamarine) differed in all the three sequences. To better localize variant sites ( ), the aminoacidic patterns that constitute the PC1–PC2–PN1–PN2 subdomains of the porter domain are indicated in this multisequence alignment (MSA). As mentioned above, this porter domain encloses both the access pocket (AP) and the deep pocket (DP), which are generally involved in the access and binding of the substrate, respectively, in its extrusion pathway. These subdomains in MexY variants were identified on the basis of MexB subdomains [37,38,39] through a multi-alignment sequence ( ).

It is worth noting that the differences in aminoacidic composition, as represented in , mainly affect the CH1 entrance cleft site that is adjacent to the AP.

3.1.2. Molecular Docking of Berberine Derivatives

The different binding affinities of the three selected berberine derivatives towards the MexY proteins were firstly assessed, considering the aminoacidic variations in the MexY sequences in P. aeruginosa PAO1, PA7, and PA14, ( , ). The docking of the potential EPIs was focused on the periplasmic site and the cleft site, which are involved in the extrusion of lipophilic, small-sized compounds and hydrophilic larger compounds, respectively [38].

Open in a separate window

Table 2

CompoundΔGbinding
MexY–PAO1ΔGbinding
MexY–PA7ΔGbinding
MexY–PA14Berberine−7.82 kcal/mol−8.66 kcal/mol−8.11 kcal/molo-CH3−8.54 kcal/mol−10.56 kcal/mol−9.38 kcal/molp-CH3−8.39 kcal/mol−10.48 kcal/mol−9.22 kcal/molp-CF3−9.03 kcal/mol−10.22 kcal/mol−9.28 kcal/molOpen in a separate window

Docking results were analyzed considering the best scoring poses for each ligand within MexY complexes in the three different considered strains and comparing the derivatives’ affinity with that of the parent compound berberine. All the tested compounds show a favorable binding energy in all the three complexes, with the best in silico results obtained for the MexY–PA7/ligand complexes for both berberine and its derivatives ( A).

From the docking focusing on the two periplasmic loops of TM7–TM8 helices, it emerges that in MexY–PA7 complexes all berberine derivatives adopt the same orientation inside the cavity ( A), with the aromatic moiety orientated toward the periplasmic side. On the contrary, inside the other two MexY complexes (PAO1, PA14), the ligands are positioned in different orientations and, comparing their binding affinities ( B, ), none is more favorable than the corresponding one for the MexY–PA7 model.

Besides, in the MexY–PA14 complex, the p-CH3-berberine (green) and p-CF3-berberine (blue) are located with same orientation in the target site, while o-CH3-berberine (cyan) is located with an inverted opposite pose ( C). Analyzing in detail the berberine docking interactions ( ) in these three different complexes, we can point out the presence of the hydrophobic residues ILE38 PHE560 LEU669 LEU561 ALA825 e LEU666 that stabilize all MexY-berberine complexes; in particular, the PHE560 side chain group establishes the π–π interaction with the aromatic alkaloid. The stronger binding affinity found in the MexY-berberine complexes of PA7 and PA14 strains also involves the SER831 and GLN856 residues that correspond to GLN831 and PRO856 in the PAO1 strain.

For 13-(4-trifluoro-methyl-benzyl)-berberine (p-CF3), the best docking score was evaluated in the MexY–PA7 strain complex, even if it has a high affinity for both the other two MexY proteins (PAO1 and PA14). In these stabilizations, the hydrophobic residues (PHE560LEU561ALA559ALA825) and hydrophilic residues (SER830 in the PA7 strain, GLN558 and GLN856 in the PA14 strain, and GLN830 in the PAO1 strain) are involved.

In addition, 13-(2-methylbenzyl)-berberine (o-CH3) complexes are stabilized by hydrophobic residues ( ). In detail, the substituent aromatic ring has a hydrophobic interaction with ALA559, ALA825 in MexY–PA7 and PA14 complexes. In addition, SER831 is involved in a OH–π interaction and GLN831 in the PAO1 strain contributes to the derivative stabilization by a hydrogen bond.

Table 3

PAO1 StrainPA7 StrainPA14 Strain13-(2-trifluoromethylbenzyl)-berberine (p-CF3)PHE560, LEU561
ALA559, GLN558
GLN830, GLN831
VAL857PHE560, LEU561
ALA559, PHE675
ALA825, SER830
GLN856PHE560, LEU561
ALA559, ALA825
GLN856, GLU860
ASP56313-(2-methylbenzyl)-berberine (o-CH3)ALA559, PHE560
LEU561, PHE675
GLN830, GLN831PHE560, ALA559
LEU561, PHE675
ALA825, SER831
GLN856PHE560, ALA559
LEU561, PHE675
ALA825, GLN85613-(4-methylbenzyl)-berberine (p-CH3)ALA559, PHE675
PHE560, VAL857
GLN831PHE560, ALA559
PHE675, ALA825
GLN856, SER831ALA559, ALA825
PHE675, PHE560
GLN558, ARG861BerberineALA559, PHE560
LEU666, LEU669
VAL857, ILE38PHE560, LEU561
LEU666, ALA825
SER831, GLN856PHE560, LEU561
SER673, ALA825
GLN856Open in a separate window

Finally, 13-(4-methylbenzyl)-berberine (p-CH3) complexes are stabilized by hydrophobic residues and polar residues such as GLN856, SER831 in PA7, GLN831 in PAO1, and GLN558 and ARG861 in PA14, which are all involved in H-bond interactions.

Analyzing these interacting residues, a further observation is that mutated residues such as GLN831(PAO1)→SER(PA7-PA14) THR861(PAO1)→ARG(PA7-PA14) PRO856(PAO1)→GLN(PA7-PA14) in PA7 and PA14 strain complexes stabilize the interactions with the three different derivative compounds; the most important aminoacidic substitution is related to residue PRO825 (PAO1), which is an alanine in PA7 and PA14 strains and a glycine in the MexY PAO1 strain. This residue is involved together with ALA559 in the stabilization of the substituent aromatic ring in MexY–PA7 and MexY–PA14 complexes with all three derivatives. From a previous study, it is proposed that the F610A mutation is involved in the best interaction with a doxorubicin ligand, so interactions with phenylalanine are less strong [39].

3.1.3. Molecular Dynamics Results

The docked complexes underwent molecular dynamics (MD) simulations to assess the stability of the ligands’ association inside the binding site. From the analysis of MD trajectories, we can distinguish different results for each MexY–ligand complex of the three considered strains. For berberine, the stabilization is hindered by its fluctuation inside this site, which corresponds to the ligand’s partial exit as observed in the MexY–PA14 strain complex. In the PAO1–MexY complex, a transition of the ligand between the two periplasmic loops toward the inner space is observed, while in the PA7–MexY complex berberine also oscillates inside it throughout the entire simulation time. Generally, the aromatic functionalization of berberine is able to increase its stabilization within the CH1 access site, but the type of functionalizing group is important in order to maximize the interactions with the cleft’s residues.

Analyzing the RMSD curves along the MD trajectories, we can notice a different stabilization pathway (Figure S1—Supplementary Materials). For the PAO1 strain, we observe a better stabilization of the o-CH3-berberine with respect to the p-CF3 and p-CH3 derivatives. Notwithstanding the proper orientation of the ligand within the site, the p-CF3 moiety is responsible for the displacement of the ligand from the fissure due to the attraction of the fluorinated group exerted by GLN558 and GLN830, which destabilize the molecular interactions with other residues inside the slit. For the p-CH3 derivative, even if the same initial pose of o-CH3-berberine was found, the methyl group’s functionalization in the para position, instead of in the ortho position, on the aromatic ring is associated with the ligand’s destabilization during the simulation, with the ring swinging and rotating around the benzylic single bond along the MD trajectory. This movement is made possible since there are no stabilizing interactions involving the p-CH3 that can fix it in a specific orientation. This different situation is ascribed to the presence of the ortho-methyl substituent, which is kept almost fixed around the same position during the whole simulation, thus leading to a higher degree of stabilization ( ).

Open in a separate window

For the PA14 strain, from the MD data analysis, after the stabilization of the complex has been reached, all three derivative molecules are not located in the central position within the slit, showing a lower effect on the stabilization of complexes that is evidenced by a reduced number of specific interactions with the protein aminoacids of the binding cleft as reported from a PLIP analysis [40].

For the PA7 strain, the p-CF3 derivative is less stable than the o-CH3 one during the MD simulation due to the attraction of the fluorinated group by SER831, which induces the ligand to move away from the original CH1 site. Concerning the p-CH3 derivative, the para substitution of the methyl group leads to a slight initial destabilization due to a steric hindrance between the Ala 559 and Ala 825 residues. This steric hindrance makes the ring oscillate frequently during simulations, preventing the total stabilization of the EPI inside the cleft. A different scenario is presented by the o-CH3 derivative, which appears stable and strongly anchored inside the pocket; the substitution in the ortho position on the aromatic ring and the methyl group itself leads to stabilizing interactions with the derivative, which are necessary to avoid the excessive reorientation of the ligand inside the fissure, thus increasing its potential inhibitory activity.

3.2. Microbiological Results

3.2.1. Synergistic Activity of Berberine Derivatives with Tobramycin

Preliminary MIC assays excluded any antimicrobial activity of the compounds in the tested concentration range (10–320 µg/mL) and of their solvent (1.5% DMSO) (data not shown). Then, the synergistic effect of the three berberine derivatives o-CH3, p-CH3, and p-CF3 in combination with tobramycin was tested against the P. aeruginosa strains PAO1, PA7, and PA14. The synergistic action of berberine at the active concentration of 80 µg/mL was determined as a comparison ( ).

Table 4

P. aeruginosa StrainTOB MIC (µg/mL) MIC TOB+Be (µg/mL)PAO10.50.25PA725632PA140.50.5Open in a separate window

Against P. aeruginosa PA7, showing a high degree of tobramycin resistance (MIC, 256 µg/mL), the presence of 80 µg/mL berberine induced an 8-fold decrease (from 256 to 32 µg/mL) in tobramycin’s MIC, and its derivatives o-CH3 at 40 µg/mL and p-CH3 at 320 µg/mL a 32-fold (from 256 to 8 µg/mL) and 16-fold (from 256 to 16 µg/mL) decrease, respectively. The p-CF3 derivative was found to be less synergic than berberine, causing only a four-fold decrease (from 256 to 64 µg/mL) in tobramycin’s MIC and was not further considered.

The two CH3 derivatives were then used in association with tobramycin against P. aeruginosa PAO1 and PA14, always causing a reduction in tobramycin’s MIC. The p-CH3 derivative exerted only a 2-fold decrease in tobramycin’s MIC against P. aeruginosa PAO1 at all tested concentrations, whereas in all other assays the compounds exerted at least a 4-fold reduction starting from the concentrations that were active against P. aeruginosa PA7. All the results of the checkerboard assays are summarized in .

Table 5

Compound Concentration (µg/mL) Compound010204080160320P. aeruginosa Straino-CH325632168888PA7p-CH32561286464323216o-CH30.50.060.060.060.060.060.06PAO1p-CH30.50.250.250.250.250.250.25o-CH30.50.250.250.1250.1250.1250.125PA14p-CH30.50.1250.1250.1250.12501250.125Open in a separate window

Finally, the two most active minimal concentrations of both the o-CH3 (40 µg/mL) and p-CH3 (320 µg/mL) derivatives were tested in association with tobramycin in MIC determination assays against the ΔmexXY strain P. aeruginosa K1525 and the same strain complemented with the mexXY plasmid pYM004. While the mutant strain did not show any modification of its tobramycin’s MIC in the absence/presence of both compounds, the complemented strain exhibited a 2- and 4-fold decrease in the tobramycin’s MIC in the presence of the p-CH3 and o-CH3 derivatives, respectively; a behavior similar to that observed with the wild-type strain P. aeruginosa PAO1.

3.2.2. Enhancement of Tobramycin Killing Activity by the Berberine Derivatives

The ability of the o-CH3 and p-CH3 berberine derivatives to improve the tobramycin killing activity was evaluated by killing curve assays against P. aeruginosa PA7 ( ).

Open in a separate window

In combination with both compounds, all tested tobramycin concentrations were found to be bactericidal starting from 2 h of exposure, with a thousand-fold reduction in the inoculum. Specifically, at this time point, the combination with the o-CH3 derivative reduced the P. aeruginosa’s abundance by two logs, irrespective of the drug concentration ( A), and the combination with the p-CH3 derivative reduced the P. aeruginosa’s abundance by 4 logs compared with tobramycin alone ( B). From 4 h to the end of the experiment, the bacterial count was always ≤10 CFU/mL. Moreover, while all cultures exposed to tobramycin alone showed a CFU increase between 8 and 24 h, no CFU increase was observed for those exposed to the EPI/drug combinations.

4. Discussion

In this work, we evaluated the interaction between berberine and its three aromatic derivatives with the access to the periplasmic site between the periplasmic loop of TM7 and TM8 of the MexY protein, the inner transmembrane channel of the P. aeruginosa MexXY-OprM efflux pump, which is responsible for aminoglycoside extrusion. This site was selected because it is involved in the recruitment of lipophilic and small-sized substrates directly from the periplasmic space and it has been investigated in our previous research [11,12,41]. Indeed, a ligand that can bind strongly at this fissure could hamper the binding of natural substrates by hindering the conformational change in protomers in the MexY homotrimer in a non-competitive mechanism. The stability of the complex EPI–substrate is pivotal in order to hamper antibiotic extrusion. To evaluate the binding stability of the berberine derivatives with the different MexY variants, we performed a molecular docking and molecular dynamics investigation. The berberine derivatives showed a higher affinity than the parent compound, with more stable ligand complexes due to the presence of a mono-substituted aromatic ring, which allowed the ligands to best anchor the binding position during all of the MD simulations.

The MexY aminoacidic variants carried by the three analyzed P. aeruginosa strains PAO1, PA7, and PA14 were shown to influence the protein conformations and the binding stability of the simulated complexes. The docking investigation indicated that the compounds’ ability to bind more strongly depends on their conformation and orientation within the task. In particular, many interacting residues were found to be able to stabilize the berberine derivatives in a more specific way than berberine due to the presence of the aromatic substituted group. Moreover, the aminoacidic variants in these three MexY polymorphic forms can affect the binding strength, which is much higher in the case of the complexes within the P. aeruginosa PA7 MexY due to the presence of ALA559 and ALA825 near the aromatic ring. In P. aeruginosa PAO1 complexes, different residues located in the porter domain, inside the access pocket (AP), and inside the periplasmic access site affect the conformation of these regions. In fact, the differences in the aminoacidic composition mainly affect the cleft site being adjacent to the AP and representing the eligible site where EPIs bind to block the aminoglycoside extrusion. These differences in aminoacidic composition between these three MexY variants may be useful to understand the differences in the binding mode of different EPIs and their overall affinity toward the protein site. The numerous aminoacidic differences found in the sequence of PAO1–MexY compared with those found in the homologue protein carried by PA7 and PA14 could explain the quite different binding poses and the hugely different interactions exhibited by the ligands in P. aeruginosa PAO1, which, however, resulted in the worst score when compared with the MexY–ligand complexes in the other two strains.

The o-CH3 berberine, which gave the best in vitro results, gains a different orientation inside the considered access site depending on the MexY polymorphism, as the protein aminoacidic modifications are mainly located inside and toward this location. In vitro data show an EPI activity greater than that of berberine for this derivative, which exerted a reduction in the tobramycin’s MIC up to 4 times greater than the parent alkaloid. This was particularly evident with the strain PA7, which showed the highest MIC (256 µg/mL), but was also remarkable with P. aeruginosa PAO1 and PA14, whose susceptibility to tobramycin was lightly or even unaffected by berberine [11]. Moreover, the compound’s active concentration was 40 µg/mL (0.070 µM), i.e., 3 times lower than that of berberine. This could be due to the stronger interaction at the binding site evidenced by the computational results. In addition, it is worth noting that within the same MexY conformation, the ligand orientation inside the binding pocket strongly depends on its functionalization (i.e., o-CH3 vs. p-CH3 vs. p-CF3). Accordingly, the docking and MD results clearly show that the type and the position of the functionalizing group directly influence the repositioning of the ligand and then its stability inside, as demonstrated by the experimental data.

Similar results were obtained with the p-CH3 derivative against P. aeruginosa PA7 and PA14, although only when used at a concentration greater (320 µg/mL/0.56 μM) than that of berberine (80 µg/mL/0.22 μM). Considering the molar concentration (the molecular weight of the synthesized compound exceeded that of berberine (570.36 (iodide salt) vs. 371.81 g/mol (chloride salt))), the activity of the p-CH3 derivative can thus be considered not so far away from that of berberine.

Time killing assays showed that the two CH3 derivatives were both synergic with tobramycin, with an evident bactericidal activity (a 3-log decrease in the CFU count) of the drug combination after 2 h with the p-CH3 derivative even when using 1/2×MIC tobramycin. Moreover, the lack of a CFU increase at 24 h when using the drug combinations suggests a role for the tested compounds in preventing the development of adaptive resistance in P. aeruginosa subpopulations [38]. This is pivotal for an effective bacterial clearance and the eradication of recurrent infections. These results suggest that the best ligand orientation shows the aromatic moiety oriented inside the periplasmic loops, with the lipophilic methyl group in the ortho position, thus avoiding the steric hindrance, allowing for a better adaptation of the ligand inside the lipophilic access site, and enhancing the binding affinity.

From our joint in silico and microbiological studies, it arises that the substitution of the natural alkaloid berberine with an aromatic lipophilic group leads to an evident increase in the EPI activity due to the bindings being tight and deep inside this site, with the lipophilic residues located inside. The ligands’ positioning in order to exert an inhibitory activity strongly depends not only on the nature of the functionalization but also on the polymorphism of the aminoacidic sequence of the MexY of different P. aeruginosa strains.

5. Conclusions

Polymorphic variants of MexY must be taken into account in order to rationally design new EPIs for combined antibiotic therapy and to counteract P. aeruginosa’s tobramycin resistance due to the drug efflux. We have shown here that polymorphisms act on ligands’ orientation due to the aminoacidic composition all around the binding site and, thus, influence the binding energy, the complex’s stability, and the dynamical evolution of binding complexes, resulting in ligand extrusion or pump blockage. The observed differences in molecules’ orientations are helpful to distinguish how a substitution in a derivative compound is better in terms of chemical features and position. In particular, for the PA7 strain, in silico models evidence differences due to polarity and the substitution sites that were also reported in vitro. In the other two strains, except for the fluorinated derivative, the other two ligands do not show substantial variations in the EPIs’ stability, which was also confirmed in vitro. The o-CH3 berberine was found to be the most active berberine derivative, improving the alkaloid activity and requiring lower concentrations to exert a synergistic effect; this could represent a leading compound for the design of novel and even more potent EPIs. Further in silico studies are in progress to screen a larger number of clinical P. aeruginosa isolates to obtain more information on MexY polymorphisms and their role in both the efflux pump substrate’s specificity and EPI effectiveness.

Supplementary Materials

The following are available online, Figure S1: Calculated RMSD along the MD trajectories for the three berberine derivatives.

Click here for additional data file.(149K, zip)

Author Contributions

Conceptualization, R.G. and F.B.; methodology, R.G., G.G., G.M. (Giovanna Mobbili), N.C.; software, G.G., E.L.; validation, R.G., F.B., and G.M. (Giovanna Mobbili); formal analysis, G.G., M.C., C.M. and E.L.; investigation, G.G., G.S., G.M. (Giovanna Mobbili) and N.C.; resources, R.G.; data curation, G.G., G.M. (Gianmarco Mangiaterra) and N.C.; writing—original draft preparation, G.G. and R.G.; writing—review and editing, E.F., E.L., M.C., G.M. (Gianmarco Mangiaterra), G.M. (Giovanna Mobbili), and F.B.; supervision, R.G. and F.B.; project administration, R.G.; funding acquisition, R.G. and F.B.. All authors have read and agreed to the published version of the manuscript.

Funding

This research received no external funding.

Conflicts of Interest

The authors declare no conflict of interest.

Footnotes

Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Mechanisms of Berberine (Natural Yellow 18)–Induced ...

Abstract

Berberine [Natural Yellow 18, 5,6-dihydro-9,10-dimethoxybenzo(g)-1,3-benzodioxolo (5,6-a) quinolizinium] is an alkaloid present in plants of the Berberidaceae family and used in traditional Chinese and North American medicine. We have previously demonstrated that berberine causes mitochondrial depolarization and fragmentation, with simultaneous increase in oxidative stress. We also demonstrated that berberine causes an inhibition of mitochondrial respiration and a decrease on calcium loading capacity through induction of the mitochondrial permeability transition (MPT). The objective of the present work is to investigate a common target for both induction of the MPT and inhibition of respiration. The hypothesis is that berberine induces the MPT through interacting with the adenine nucleotide translocator (ANT). By measuring induction of the MPT through increased mitochondrial swelling, membrane depolarization and loss of calcium retention, we observed that the effects of berberine were not inhibited by bongkrekic acid although adenosine diphosphate (ADP)/oligomycin completely prevented the MPT. Also, we observed that berberine increased the depolarization effect of oleic acid on liver mitochondria. The initial depolarization observed when berberine is added to mitochondria was not affected by ANT inhibitors. Taken together, we propose that berberine acts on the ANT, altering the binding of the protein to bongkrekic acid but not to cyclosporin A or ADP. It is also clear that the membrane potential is required for berberine effects, most likely for allowing for its mitochondrial accumulation. Mitochondrial effects of berberine can be relevant not only for its proposed antitumor activity but also for the assessment of its organ toxicity, depending on factors such as tissue accumulation or delivery.

Berberine [Natural Yellow 18, 5,6-dihydro-9,10-dimethoxybenzo(g)-1,3-benzodioxolo (5,6-a) quinolizinium], a benzyl tetra isoquinoline plant alkaloid derived from the Berberidaceae family, has been extensively used for many centuries. In fact, extracts containing berberine have been used in the traditional Chinese and Native American medicine. Berberine chemical structure (Fig. 1) contains a quaternary base and it is commercially available as various salts such as berberine chloride and hemisulfate. Many pharmacological applications have been attributed to berberine, such as antidiarrheic (Taylor and Baird, 1995), antimicrobial (Kaneda et al., 1991), anti-inflammatory (Ckless et al., 1995), antiarrhythmic (Hong et al., 2002), antihypertensive (Hong et al., 2002), antiproliferative/antitumoral (Letasiova et al., 2005; Serafim et al., 2008), and as an antioxidant (Shirwaikar et al., 2006). It has also been reported that berberine reduces glucose blood levels in diabetes (Zhou et al., 2007), alters the processing of Alzheimer's amyloid precursor protein decreasing Aβ secretion (Asai et al., 2007), and shows antiangiogenic properties (Lin et al., 2004), inhibiting HIF-α expression via enhanced proteolysis. Berberine is metabolized in the liver by cytochrome P450, suffering phase I metabolism (Pereira et al., 2007)

Previously, we demonstrated that berberine is selectively accumulated by mitochondria on K1735-M2 melanoma cells, arresting cell proliferation, causing mitochondrial fragmentation and depolarization, oxidative stress and a decrease in ATP levels (Pereira et al., 2007). A decrease in the number of mitochondria-like structures and in mitochondrial DNA copy number was also found (Pereira et al., 2007). The work by Barreto et al. (2003) also confirmed the relevance of positively charges on the selective accumulation of alkaloids in mitochondria. We have identified two distinct types of effects of berberine on isolated mitochondrial fractions. Berberine inhibits state 3 and uncoupled respiration, which is more visible when complex I substrates are used (Pereira et al., 2007). On the other hand, berberine also decreases mitochondrial calcium loading capacity through enhanced induction of the mitochondrial permeability transition (MPT) (Pereira et al., 2007). Through the analyses of several parameters, it is proposed that berberine inhibits the phosphorylative system, although no effects on the ATPase activity were observed (Pereira et al., 2007). The interaction of berberine on both the phosphorylative system and on the MPT suggests that a common target may be involved. One interesting candidate is the adenine nucleotide translocator (ANT).

The ANT is a mitochondrial inner membrane adenosine diphosphate (ADP)-ATP antiporter that imports ADP to the matrix and exports ATP to the cytosol, being suggested to be a structural component of the MPT pore (Armstrong, 2006), although recent data cast doubts regarding this issue (Juhaszova et al., 2008). The ANT alternates between two distinct conformations in which adenine nucleotides are either bound to the cytosolic side (c-state) or to the matrix side (m-state) of the inner mitochondrial membrane. It has also been described that the c-conformation is favorable to the MPT (Armstrong, 2006). Carboxyatractyloside which binds to ANT in the c-state activates the MPT pore, whereas bongkrekic acid which binds to ANT in the m-state inhibits the MPT pore. The two ligands are known to be ANT specific inhibitors (Armstrong, 2006).

The role of the ANT in the MPT has not been fully resolved. Although the MPT is an inner membrane phenomenon, there is substantial evidence of the involvement of the VDAC (voltage-dependent anion channel) in the MPT. Crompton et al. (1998) showed that matrix cyclophilin D (CycP-D), the third putative component of the MPT, binds to complexes of VDAC and ANT in order to form the MPT complex. It has also been reported that CycP-D is more of a regulatory component of the MPT pore than a structural one. Cyclosporin A (Cyc A) was show to block the MPT by inhibiting the binding of CycP-D to the ANT (Szabo and Zoratti, 1991).

MPT activation compromises the normal integrity of the mitochondrial inner membrane resulting into uncoupled oxidative phosphorylation, ATP decay, mitochondrial swelling and release of apoptogenic factors, indicating a key role for the MPT in apoptosis, although this subject is still controversial (Armstrong, 2006).

In the present work, we proposed to investigate the mechanism by which berberine induced the MPT pore in isolated rat liver mitochondria; we hypothesize that berberine induces that phenomenon by interacting with the ANT. Experiments with specific ANT ligands (bongkrekic acid, carboxyatractyloside) were performed in order to identify the ANT as the site for berberine action on isolated liver mitochondria.

MATERIALS AND METHODS

Materials.

Berberine hemisulfate was obtained from Sigma-Aldrich (St Louis, MO) and prepared in dimethyl sulfoxide (DMSO), which is routinely used as a solvent for isolated mitochondrial studies. The total volume of DMSO was always smaller than 0.1%, which had negligible effects in all experiments. Concentrations used in this study were based in previous literature that showed effects in both intact melanoma cells and isolated hepatic mitochondria (Pereira et al., 2007).

Calcium Green-5N, hexapotassium salt, was obtained from Invitrogen (Carlsbad, CA) and Bongkrekic acid was obtained from Calbiochem (San Diego, CA, purity: ≥ 92% by high-performance liquid chromatography). Carboxyatractyloside was obtained from Sigma-Aldrich.

All other compounds were of the highest grade of purity commercially available.

Animals.

Male Wistar rats (6–10 weeks old), housed at 22 ± 2°C under artificial light for a 12-h light/day cycle and with free access to water and food, were used throughout the experiments. The research procedure was carried out in accordance with the European Requirement for Vertebrate Animal Research and within the internal Standards for Animal Research at the CNC.

Isolation of rat liver mitochondria.

Mitochondria were isolated from livers of male Wistar rats by conventional methods with slight modifications (Pereira et al., 2007). Homogenization medium contained 250mM sucrose, 10mM 4-(2-Hydroxyethyl)piperazine-1-ethanesulfonic acid (HEPES), pH 7.4, 1mM EGTA, and 0.1% fat-free bovine serum albumin (BSA). Ethylene-bis(oxyethylenenitrilo)tetraacetic acid (EGTA) and BSA were not present at the final washing medium, adjusted at pH 7.4. The mitochondrial pellet was washed twice, suspended in the washing medium, and immediately used. Protein content was determined by the biuret method (Gornall et al., 1949) calibrated with BSA.

Mitochondrial swelling.

Mitochondrial volume changes were followed by the decrease of absorbance at 540 nm with a Jasco V-560 spectrophotometer (Jasco, Tokyo, Japan). The assays were performed in 2 ml of the reaction medium (200mM sucrose, 10mM Tris-HCl, pH 7.4, 10μM EGTA, 1mM KH2PO4), 0.6μM rotenone plus 5mM succinate for complex II–related assays or 5mM glutamate/malate for complex I assays to which 1.5 mg of mitochondrial protein was added under constant stirring at 30°C.

Calcium (26.7–53.3 nmol/mg protein) was added to the preparation approximately 1 min after the start of the experiment. Cyc A (1μM) was added to the mitochondrial preparation before the addition of calcium to inhibit the MPT. Berberine was preincubated with the mitochondrial preparation for approximately 1–2 min.

The ANT inhibitors carboxyatractyloside (2.5μM) and bongkrekic acid (12.5μM) were also used in some experiments in order to modulate the effect of berberine on the MPT. The concentrations of the two ANT inhibitors were chosen with basis in preliminary experiments, regarding inhibition of ADP phosphorylation and MPT modulation (data not shown). Both carboxyatractyloside and bongkrekic acid were added to the mitochondrial preparation before calcium addition. Different concentrations of berberine were preincubated for 1 min with the mitochondrial preparation.

Mitochondrial calcium loading.

Extramitochondrial free Ca2+ was assayed by using the hexapotassium salt of the fluorescent probe Calcium Green 5-N (1μM).

Liver mitochondria (0.75 mg) were suspended in 2 ml of buffer containing 200mM sucrose, 10mM Tris, 10μM EGTA, and 1mM KH2PO4. Rotenone (0.6μM) and succinate (5mM) were also added to the media.

Fluorescence was continuously recorded in a water-jacketed cuvette holder at 30°C using a PerkinElmer LS-55 fluorescence spectrometer (PerkinElmer Life and Analytical Sciences, Boston, MA) with excitation and emission wavelengths of 506 and 531, respectively. Slits used were 5 nm for both excitation and emission.

Different concentrations of berberine were added to the mitochondrial preparation (0.75 mg), before the addition of calcium (10–13.3 nmol/mg protein) which was made 1 min after the start of the experiment.

Cyc A (1 μM) was added to the mitochondrial preparation before the addition of calcium to inhibit the MPT. In some assays, carboxyatractyloside (2.5μM) and bongkrekic acid (12.5μM) were added before calcium addition.

Measurement of the mitochondrial transmembrane potential.

The mitochondrial transmembrane potential (Δψ) was indirectly estimated by the mitochondrial accumulation of the lipophilic cation—tetraphenylphosphonium cation (TPP+) as detected by using a TPP+ selective electrode in combination with an Ag/AgCl saturated reference electrode. Both the TPP+ electrode and the reference electrode were inserted into an open vessel with magnetic stirring and were connected to a pH meter (model 3305, Jenway, Essex, UK). The signals were fed into a potentiometric recorder (model BD 121,Kipp & Zonen B.V., Delft, The Netherlands). No correction was made for “passive” binding of TPP+ to the mitochondrial membranes because the purpose of the experiments was to show relative changes in potentials rather than absolute values.

Mitochondrial protein (1.5 mg) was suspended under constant stirring in 1 ml of reaction medium composed 125mM sucrose, 65mM KCl, 5mM KH2PO4, 2.5mM MgCl2, and 5mM HEPES, pH 7.4 (30°C), supplemented with 3μM TPP+. Mitochondria were energized with 5mM of succinate plus 3μM rotenone.

Different concentrations of berberine were added before mitochondrial protein. Forty to 53.3 nmol of calcium/mg protein was added to initiate the MPT.

Absolute values for membrane potential (in mV) were determined from the equation originally proposed by (Kamo et al., 1979), assuming Nernst distribution of the ion across the membrane electrode.

Cyc A (1μM) was added to the mitochondrial preparation before the addition of calcium to inhibit the MPT. As described before, carboxyatractyloside (2.5μM), bongkrekic acid (12.5μM), or ADP (500μM) plus oligomycin (20 μg) were also preincubated with mitochondria.

In order to investigate the effects of berberine on oleic acid–induced depolarization, freshly isolated mitochondria (1.5 mg) were energized using 5mM succinate plus 3μM rotenone and were allowed to stabilize at the maximum membrane potential for 60 s.

After ΔΨ stabilization, 0.84 nmol/mg (four additions total = 6.88 nmol/mg protein) of oleic acid was added to induce membrane depolarization. Different concentrations of berberine were also added before initiating the assay.

In selected experiments, carboxyatractyloside (2.5μM) or carbonyl cyanide 4-(trifluoromethoxy)phenylhydrazone (FCCP) (0.1μM) was added to the mitochondrial fractions.

Statistical analysis.

Data are expressed as means ± SEM and evaluated by one-away ANOVA followed by Bonferroni multiple comparison tests.

Differences were considered significant if the p < 0.05.

RESULTS

Berberine Induces the MPT in Isolated Hepatic Mitochondrial Fractions

Calcium-dependent mitochondrial swelling measured by light scattering at 540 nm is a characteristic measure of MPT induction (Hunter and Haworth, 1979). In a first approach we investigated whether berberine was able to induce the MPT with complex I substrates, in a similar fashion to what was previously observed for complex II substrates (Pereira et al., 2007). Figures 2A–D demonstrate that berberine in a dose-dependent fashion was able to induce mitochondrial swelling upon calcium addition, irrespective of the substrate used. Succinate was then used as the substrate for the following experiments regarding MPT induction.

FIG. 2.

Open in new tabDownload slide

Mitochondria were incubated with different concentrations of 13.3, 33.3 and 66.7 nmol berberine/mg protein and with succinate plus rotenone or glutamate/malate as substrates. Calcium (26.7–53.3 nmol/mg protein) was added to the preparation in order to induce the MPT pore. (A and C) Typical recording and amplitude of swelling recorded 400 s after calcium addition, when glutamate-malate was used as substrate. (B and D) Typical recording and amplitude of swelling recorded 400 s after calcium addition, when succinate was used as substrate. (C and D) Bars represent the mean ± SEM of 6/7 different preparations. (*) Statistically different compared with the control (p < 0.05). (A and B) 0, reference line; 1, control; 2, 13.3 nmol berberine/mg protein; 3, 33.3 nmol berberine/mg protein; 4, 66.7 nmol berberine/mg protein.

We followed up the selecting one berberine concentration (133.3 nmol/mg protein) and investigating whether classic ANT inhibitors such as carboxyatractyloside and bongkrekic acid could interfere with the action of berberine on the MPT (Fig. 3). Bongkrekic acid was previously shown to inhibit MPT induction in isolated hepatic (Vergun and Reynolds, 2005) and cardiac mitochondria (Haworth and Hunter, 2000).

FIG. 3.

Open in new tabDownload slide

Effects of ANT inhibitors on berberine-induced swelling. (A) Typical recording. Mitochondria were preincubated with the presence and absence of 133.3 nmol berberine/mg protein and/or carboxyatractyloside (CATR) and bongkrekic acid (BA); Calcium (26.7–53.3 nmol/mg protein) was then added to the preparation in order to induce the MPT pore. Mitochondrial swelling traces are representative of six repetitions, each using different mitochondrial preparations from separate animals. Legend: 0- Reference line; 1, bongkrekic acid (BA); 2, Cyc A; 3, control (Ctrl); 4, CATR; 5, berberine (berb); 6, BA + berb; 7, CATR + berb. (B) Data represent the amplitude of absorbance measured before and 400 s after calcium addition. Bars represent the mean ± SEM of six repetitions each analyzed by one-way ANOVA. Differences were considered significant if p < 0.05. *, Ctrl versus BA, CATR, and berb; +, BA versus BA + berb; #, BA + berb versus Cyc A + BA + berb.

As expected, in the absence of berberine, bongkrekic acid was able to inhibit the MPT, as opposed to carboxyatractyloside which increased calcium-induced mitochondrial swelling. The most relevant result was that bongkrekic acid was not able to inhibit berberine-induced mitochondrial swelling (Fig. 3B). The addition of Cyc A, the classic MPT pore inhibitor (Broekemeier et al., 1989), to mitochondria incubated with bongkrekic acid plus berberine was able to decrease mitochondrial swelling.

In order to verify if berberine effects on MPT were dependent on the mitochondrial transmembrane potential, we performed swelling assays with nonenergized mitochondria, where mitochondrial calcium loading was obtained through a calcium ionophore.

The results described in Figure 4 demonstrate that in the absence of a transmembrane electric potential, berberine does not cause mitochondrial swelling in the presence of calcium.

FIG. 4.

Open in new tabDownload slide

Berberine does not cause mitochondrial swelling in nonenergized mitochondria. Data represent the swelling amplitude before and 400 s after calcium addition. Bars represent the mean ± SEM of three separate experiments.

Berberine Induces Mitochondrial Depolarization Dependent on Calcium

Calcium-induced MPT opening leads to the loss of ΔΨ in a Cyc A–sensitive fashion (Pereira et al., 2007). Figure 5 illustrates a typical experimental of measuring the effect of different concentrations of berberine (6.7, 13.3, 33.3, and 66.7 nmol/mg protein) on calcium-induced mitochondrial depolarization.

FIG. 5.

Open in new tabDownload slide

Berberine stimulates calcium-induced depolarization. Freshly isolated mitochondria were energized with 5mM succinate/rotenone and different concentrations of berberine (berb) were added to the solution. Calcium (40–53.3 nmol of calcium/mg protein) was added to initiate time-dependent transmembrane potential alterations, measured by a TPP+ electrode. (A) Mitochondrial traces are representative of 7 repetitions, each using mitochondrial preparations from separate animals. The values corresponde to the Δψ before (max) and 5 or 10 min after calcium addition. A, 6.6 nmol berb/mg protein; B, 16.7 nmol berb/mg protein; C, 33.3 nmol berb/mg protein; D, 66.7 nmol berb/mg protein. (B) Bars represent the mean ± SEM of seven repetitions each analyzed by one-way ANOVA. #, max versus 5, 10 min (0, 6.7, 13.3, 33.3, and 66.7 nmol berb/mg protein); *, control (Ctrl) (5 min) versus 5 min (6.7, 13.3, 33.3, and 66.7 nmol berb/mg protein); +, Ctrl (10 min) versus 10 min (6.7, 13.3, 33.3, and 66.7 nmol berb/mg protein).

Mitochondria were preincubated with different concentrations of berberine for 60 s after being energized with succinate. Figure 5 shows typical recordings on the effect of berberine on calcium-induced depolarization. A graphic pictured in Figure 5B describes the values of ΔΨ before, 5 and 10 min after calcium addition.

The results show that 13.3, 33.3, and 66.7 nmol berberine/mg protein significantly cause mitochondrial depolarization upon calcium addition. Cyc A was able to prevent the loss of ΔΨ in the presence of calcium. Confirming results obtained with mitochondrial swelling, it was observed that the effect of carboxyatractyloside is increased by the presence of berberine (i.e., increasing calcium-induced depolarization) and that the inhibitory effect of bongkrekic acid on the loss of ΔΨ is not observable in the presence of berberine (Fig. 6). Cyc A was able to completely inhibit calcium-induced depolarization in the presence of berberine and bongkrekic acid.

FIG. 6.

Open in new tabDownload slide

Berberine decreases mitochondrial calcium accumulation independent of bongkrekic acid presence. Freshly isolated mitochondria were energized with 5mM succinate/rotenone and two different concentrations (33.3 and 66.7 nmol berberine/mg protein) were added to the solution in the presence and absence of ANT inhibitors. Berberine induced a decrease in ΔΨ time and concentration-dependent. Bars represent the mean ± SEM of six separate experiments. •, carboxyatractyloside (CATR) (0.00) versus CATR (33.33)/CATR (66.67); *, calcium (Ca) versus Berberine (berb) (33.3/66.7); +, Bongkrekic acid (BA) versus BA (33.3)/BA (66.7); #, BA (66.7) versus Cyc A + BA (66.7), numbers represent berberine concentration.

Berberine Decreases Mitochondrial Calcium Loading Capacity

To obtain further evidence that berberine induces MPT pore opening, mitochondrial calcium loading capacity in the presence of different concentrations of the molecule was evaluated. Increased MPT induction translates into a decrease ability of mitochondria to retain calcium. Extramitochondrial calcium levels were followed by using the fluorescent probe Calcium Green 5N. After an increase in probe fluorescence due to calcium addition, fluorescence decreased in all conditions tested, indicating mitochondrial calcium uptake. However, when berberine was present, calcium accumulation decreased (Figs. 7A and 7B) when compared with the control.

FIG. 7.

Open in new tabDownload slide

Assessment of mitochondrial calcium loading capacity in presence of different concentrations of berberine and ANT inhibitors. The fluorescent probe Calcium Green 5-N (1μM) was used to evaluate the levels of extramitochondrial calcium as described under “Materials and Methods.” (A) Mitochondrial traces are representative of six to seven repetitions, each using mitochondrial preparations from separate animals. (A) 0—control (Ctrl); 1—66.7 nmol berberine (berb)/mg of protein; 2—133.3 nmol berb/mg of protein; 266.6 nmol berb/mg of protein. (B) 0—Cyc A/berb/bongkrekic acid (BA); 1—BA; 2—Ctrl; 3—carboxyatractyloside (CATR); 4—BA + berb; 5—berb; 6—CATR + berb. (B) Values indicate the difference in fluorescence values before and 300 s after calcium (10–13.3 nmol/mg protein) addition. Bars represent the mean ± SEM of six to seven repetitions. *, calcium (Ca) (0) versus Ca (133.3 berb); •, CATR (0) versus CATR (133.3 berb); ▪, BA (0) versus BA (133.3 berb); #, BA (133.3 berb) versus Cyc A + BA (133.3 berb).

As in previous results, the most critical piece of data confirmed the berberine bypassed the inhibitory effect of bongkrekic acid on calcium accumulation, whereas Cyc A was able to prevent such effect. For the present study, only one berberine concentration (133.3 nmol/mg protein) was used in order to investigate the effect of ANT modulators.

Berberine does not Affect the Inhibitory Effect of ADP-Oligomycin on ANT

Because bongkrekic acid did not inhibit the stimulatory effect of berberine on the MPT, a possibility is that the ANT would not be involved in the induction of the phenomenon by berberine. To demonstrate the opposite, the next set of experiments involved the use of ADP plus oligomycin, which also acts on the ANT to inhibit pore opening. In fact, ADP is an ANT ligand and a powerful MPT pore inhibitor (Crompton, 1999; Hagen et al., 2003). Oligomycin was added to avoid mitochondrial ADP conversion into ATP.

In the presence of calcium, 33.3 and 66.7 nmol berberine/mg protein decreases membrane potential in a concentration and time-dependent manner, as it was described before. When ADP-oligomycin was added to the mitochondrial suspension no berberine-induced calcium-induced ΔΨ depolarization occurred (Figs. 8A and 8B). In fact, ΔΨ was maintained for several minutes after calcium addition.

FIG. 8.

Open in new tabDownload slide

ADP/oligomycin inhibits berberine stimulation of the MPT. A TPP+ electrode was used to follow ΔΨ variations that occur during the MPT. (A) Mitochondrial traces are representative of five different experiments. (A) A—control (Ctrl); B—33.3 nmol berberine (berb)/mg protein; C—66.7 nmol berb/mg protein; D—ADP + oligomycin (oligo); E—ADP + oligo + 33.3 nmol berb/mg protein; F—ADP + oligo + 66.7 nmol berb/mg protein. (B) The data describe the value of ΔΨ before and 5 min after calcium addition to isolated mitochondria. Bars represent the mean ± SEM of five repetitions each analyzed by one-way ANOVA. (*) Statistically different compared with the control (p < 0.05). Legend: A—Ctrl; B—33.3 nmol berberine (berb)/mg protein; C—66.7 nmol berb/mg protein; D—ADP + oligo; E—ADP + oligo + 33.3 nmol berb/mg protein; F—ADP + oligo + 66.7 nmol berb/mg protein.

Berberine Affects the Depolarizing Effect of Oleic Acid on Mitochondria

Oleic acid–induced uncoupling of mitochondria is known to be mediated by some anion carriers, including the ANT, as also by uncoupling proteins (UCPs) (Khailova et al., 2006).

We observe from Figure 9A that to each oleic acid addition, a small decrease in ΔΨ occurs. We quantified the average ΔΨ decay in the first four additions of oleic acid.

FIG. 9.

Open in new tabDownload slide

Titration with oleic acid in the presence and absence of different concentrations of berberine. (A) Mitochondrial traces are representative of nine repetitions, each using mitochondrial preparations from separate animals. Each oleic acid addition was equivalent to 0.84 nmol oleic acid/mg protein. (B) Berberine increased oleic acid–induced mitochondrial depolarization. The data indicate the average depolarization of four additions of oleic acid = 6.88 nmol oleic acid/mg protein total. Bars represent the mean ± SEM of nine repetitions. (*) Statistically difference compared with control.

The data (Fig. 9B) indicate that the two of the berberine concentrations used increase the magnitude of oleic acid effects on ΔΨ. Surprisingly, carboxyatractyloside was not able to inhibit the depolarizing effect of oleic acid.

In order to verify if the effect of berberine on oleic acid–induced depolarization was due to the depressive effect on ΔΨ (Fig. 9, and see also Pereira et al., 2007), we used a small amount of FCCP to clamp ΔΨ to a value similar to the one obtained after incubation with 133.3 nmol berberine/mg protein. No alteration on the effect of oleic acid was observed when the ΔΨ was decreased with FCCP.

The Initial Depolarization Effect of Berberine is not Dependent on the ANT

One particular effect observed when berberine is added to energized mitochondria is a sudden mitochondrial depolarization followed by a partial recovery of ΔΨ to a value similar to what is observed when berberine is preincubated with mitochondria (Fig. 10A). By using the same concentrations of bongkrekic acid and carboxyatractyloside used before we confirmed that the immediate effects of berberine on mitochondrial ΔΨ were not related with the ANT. In opposition and as expected, ADP phosphorylation was completely inhibited by both carboxyatractyloside and bongkrekic acid.

FIG. 10.

Open in new tabDownload slide

Immediate depolarization of mitochondria by berberine is not inhibited by ANT inhibitors. (A) Mitochondrial traces are representative of eight repetitions, each using mitochondrial preparations from separate animals. (A) A—control (Ctrl); B—carboxyatractyloside (CATR); C—bongkrekic acid (BA) + ADP; D—Ctrl; E—CATR; F—BA (+ 16.7 nmol berb/mg protein. (B) Berberine induced a fast and immediate mitochondrial depolarization even in the presence of CATR and BA. In a control situation (no berberine present) CATR and BA prevented ADP depolarization of mitochondria, by inhibiting the ANT. Bars represent the mean ± SEM of eight repetitions each analyzed by one-way ANOVA. Differences were considered significant if the p value was lower than 0.05. *, CATR + berb versus ADP/CATR; +, BA + berb versus ADP/BA; #, ADP versus ADP/CATR and ADP/BA.

DISCUSSION

Phytomedicine, which involves the use of natural products from traditional herbal medicines for medical and health-fortifying purposes, is gaining international popularity. However, little is known about the mechanisms of action of their active principles (Tsai and Tsai, 2004) although formulations are widely in over the counter medications.

Previously, it was demonstrated that berberine increases oxidative stress on isolated liver mitochondria and melanoma cells. Increased oxidative stress is attributed to the interaction of berberine at a site downstream of the rotenone binding site (Pereira et al., 2007).

The inhibition of oxygen consumption by berberine is more pronounced when respiratory substrates for complex I are used, as described before (Barreto et al., 2003; Pereira et al., 2007). Despite this, succinate-sustained respiration is also inhibited, suggesting a larger range of targets in the respiratory chain (Pereira et al., 2007). The effects on the ADP/O ratio, ADP depolarization and lag phase (Pereira et al., 2007) suggests that berberine interacts with the mitochondrial phosphorylative system, although no effects on the mitochondrial ATPase were found.

Because it was also observed that berberine induces the calcium-induced MPT (Pereira et al., 2007), it is proposed that berberine may also act on the ANT, a proposed component of the pore complex.

The aim of the present investigation was then to identify a common target responsible for both disruption of mitochondrial bioenergetics and the induction of the MPT. Our work hypothesis is that the ANT is a target for berberine. Several end-points for MPT induction were used, including calcium-induced mitochondrial swelling, calcium-induced calcium release and calcium-induced mitochondrial depolarization. The three end-points are considered classic and sensitive techniques to follow MPT induction induced by xenobiotics (Bernardi and Forte, 2007).

The results of this investigation demonstrate that berberine interacts with ANT on rat liver mitochondria, disturbing the effect of bongkrekic acid as a MPT pore inhibitor. Calcium-dependent mitochondrial swelling (Fig. 3) induced by the test compound is in fact not inhibited by the ANT ligand bongkrekic acid. Also, in some of the experimental protocols used, berberine and carboxyatractyloside appear acting together to induce the MPT.

The results suggest a competition by berberine and bongkrekic acid for a common binding place, or a conformational change on the ANT or surrounding lipid membrane caused by berberine that decreases the affinity for bongkrekic acid.

Interestingly, berberine was not able to induce the permeability transition in nonenergized mitochondria which suggests that a polarized inner membrane is required for the effect of berberine, which is in accordance with previously published data indicating that berberine accumulation in mitochondria requires a transmembrane electrical potential.

Although berberine effects on the MPT were not inhibited by bongkrekic acid, ADP plus oligomycin completely prevented calcium-induced depolarization in the presence of berberine. Because ADP is a potent pore inhibitor due to its binding to the ANT (Crompton, 1999; Hagen et al., 2003), the result confirm that berberine stimulates the MPT through an interaction with the ANT and not to an unrelated mechanism.

Recently, it was reported that oleic acid stimulates mitochondrial uncoupling mediated by anionic transporters, such as the ANT, which causes a membrane potential decrease and an increase in the respiratory rate of mitochondria (Khailova et al., 2006). According to the fatty acid cycle hypothesis, UCP and the ANT participate in an uncoupling process by facilitating the efflux of fatty acid anions through hydrophobic barrier of the inner mitochondrial membrane (Skulachev, 1991). According to this, a titration with oleic acid was performed, in order to understand if berberine could also affect with the depolarizing effect of oleic acid, by means of an interaction with the ANT.

The results showed that berberine increases the effect of oleic acid. Depolarization caused by oleic acid in mitochondria was enhanced when different concentrations of berberine were added to the suspension. The effects of berberine on oleic acid–induced ΔΨ depolarization were not due to the decrease in ΔΨ caused by berberine alone, as a small amount of FCCP did not have the same effect. Interestingly and not as expected, carboxyatractyloside did not inhibit the action of berberine. The lack of inhibition is surprising after the results by Khailova et al. (2006) and O'Brien et al. (2008). Nevertheless, it is also likely that oleic acid effects are also different according to the conformational state of ANT. According to our data, apparently berberine facilitates the shuttling of oleic acid across the mitochondrial membrane.

Finally, the role of the ANT in the sudden depolarization observed upon berberine addition to a mitochondrial suspension was investigated. It must be stressed that the depolarization is transient and ΔΨ partly recovers, although to a value lower than the control (Pereira et al., 2007).

The results show that berberine-induced depolarization was not related with the ANT, once carboxyatractyloside and bongkrekic acid had no effect (Fig. 10). The observation indicates that for the concentrations at which both ANT inhibitors hinder ADP phosphorylation, berberine-induced depolarization was not avoided. This depolarization is not Cyc A sensitive (data not shown) hence it is unlikely to be MPT related. The ΔΨ depolarization is clearly different from the one occurring when berberine and calcium are coincubated with isolated mitochondrial fraction. In this case, it is the MPT pore opening that triggers the loss of ΔΨ.

Interaction of berberine with the ANT may help explaining the decreased ATP/ADP ratio found in melanoma cells (Pereira et al., 2007) and the range of effects on isolated mitochondrial fractions, although it must be stressed out that other ANT-independent targets may exist, as uncoupled respiration was also inhibited by berberine (Pereira et al., 2007).

The antitumor effect of berberine (Letasiova et al., 2006a, b; Pereira et al., 2007) appears to possess a mitochondrial component. Inhibition of the ANT in tumor cells can be an advantage if the process results in cell death through induction of the MPT and if a proper targeting and accumulation of the molecule occurs. In fact, the positive moiety in the molecule may be important for accumulating higher doses of berberine in tumor cells. Accumulation of positively charged molecules into mitochondria is driven by the negatively charged mitochondrial matrix. Interestingly, it was previously demonstrated that tumor development and progression is directly linked to the magnitude of mitochondrial Δψ in cancer cells (Heerdt et al., 2005, 2006). It is reasonable to think that differences in mitochondrial membrane potential between subpopulations of tumor cells or even between normal and tumor cells (Dairkee and Hackett, 1991; Davis et al., 1985; Nadakavukaren et al., 1985; Ross et al., 2006) can be used to drive accumulation of positively charged antitumor molecules. As discussed earlier, berberine has other pharmacological indications although it is not clear at the moment if mitochondrial effects can account for its antimicrobial, antiarrhythmic or other pharmacological properties. Nevertheless, it has been described that inhibition of Complex I by berberine could account for its improvement of whole-body insulin sensitivity (Turner et al., 2008).

The data from the present work appear to show that berberine also presents some degree of toxicity to “nontumor” systems, which should be carefully understood. ANT inhibition in nontumor cells by berberine would be responsible for a decrease in energy production and could also result in MPT induction. To the best of our knowledge, no full toxicity assessment exists for berberine in humans, although its use in several commercially available supplements suggests that the compound may present a relatively wide safety interval. In fact, a study with patients with congestive heart failure treated with 1.2 g/day of oral berberine revealed low toxicity and resulted into an average plasma concentration of 0.11 mg/l which would translate into 0.3μM (Zeng and Zeng, 1999). Repeated cumulative treatments, alternative forms of formulation (e.g., topical application vs. injection) or more importantly, active mitochondrial accumulation due to its positive charge would be expected to increase its concentration in cells into the range of concentrations used in this study.

Empirical data from nontraditional medicines plus the use of extensive clinical assays would allow the use of berberine as a promising antimelanoma agent while maintaining its safety for humans. In radial/vertical forms of melanoma, a possible topical application of berberine would also be possible, thus minimizing side effects on other organs.

In conclusion, the present work identifies the ANT as an important target for berberine, with clear relevance for its proposed antitumor effects.

References

Armstrong

The role of the mitochondrial permeability transition in cell death

Mitochondrion

2006

, vol.

(pg.

225

234

)

, vol.(pg.

Asai

Iwata

Yoshikawa

Aizaki

Ishiura

Saido

Maruyama

Berberine alters the processing of Alzheimer's amyloid precursor protein to decrease Abeta secretion

Biochem. Biophys. Res. Commun.

2007

, vol.

352

(pg.

498

502

)

, vol.(pg.

Barreto

Pinto

Arrabaca

Pavao

Inhibition of mouse liver respiration by Chelidonium majus isoquinoline alkaloids

Toxicol. Lett.

2003

, vol.

146

(pg.

)

, vol.(pg.

Bernardi

Forte

The mitochondrial permeability transition pore

Novartis Found. Symp.

2007

, vol.

287

(pg.

157

164

)

, vol.(pg.

Broekemeier

Dempsey

Pfeiffer

Cyclosporin A is a potent inhibitor of the inner membrane permeability transition in liver mitochondria

J. Biol. Chem.

1989

, vol.

264

(pg.

7826

7830

)

, vol.(pg.

Ckless

Schlottfeldt

Pasqual

Moyna

Henriques

Wajner

Inhibition of in-vitro lymphocyte transformation by the isoquinoline alkaloid berberine

J. Pharm. Pharmacol.

1995

, vol.

(pg.

1029

1031

)

, vol.(pg.

Crompton

The mitochondrial permeability transition pore and its role in cell death

Biochem. J.

1999

, vol.

341

Pt 2

(pg.

233

249

)

, vol.(pg.

Crompton

Virji

Ward

Cyclophilin-D binds strongly to complexes of the voltage-dependent anion channel and the adenine nucleotide translocase to form the permeability transition pore

Eur. J. Biochem.

1998

, vol.

258

(pg.

729

735

)

, vol.(pg.

Dairkee

Hackett

Differential retention of rhodamine 123 by breast carcinoma and normal human mammary tissue

Breast Cancer Res. Treat.

1991

, vol.

(pg.

)

, vol.(pg.

Davis

Weiss

Wong

Lampidis

Chen

Mitochondrial and plasma membrane potentials cause unusual accumulation and retention of rhodamine 123 by human breast adenocarcinoma-derived MCF-7 cells

J. Biol. Chem.

1985

, vol.

260

(pg.

13844

13850

)

, vol.(pg.

Gornall

Bardawill

David

Determination of serum proteins by means of the biuret reaction

J. Biol. Chem.

1949

, vol.

177

(pg.

751

766

)

, vol.(pg.

Hagen

Lagace

Modica-Napolitano

Aprille

Permeability transition in rat liver mitochondria is modulated by the ATP-Mg/Pi carrier

Am. J. Physiol. Gastrointest. Liver Physiol.

2003

, vol.

285

(pg.

G274

G281

)

, vol.(pg.

Haworth

Hunter

Control of the mitochondrial permeability transition pore by high-affinity ADP binding at the ADP/ATP translocase in permeabilized mitochondria

J. Bioenerg. Biomembr.

2000

, vol.

(pg.

)

, vol.(pg.

Heerdt

Houston

Augenlicht

The intrinsic mitochondrial membrane potential of colonic carcinoma cells is linked to the probability of tumor progression

Cancer Res.

2005

, vol.

(pg.

9861

9867

)

, vol.(pg.

Heerdt

Houston

Augenlicht

Growth properties of colonic tumor cells are a function of the intrinsic mitochondrial membrane potential

Cancer Res.

2006

, vol.

(pg.

1591

1596

)

, vol.(pg.

Hong

Hui

Chan

Hou

Effect of berberine on regression of pressure-overload induced cardiac hypertrophy in rats

Am. J. Chin. Med.

2002

, vol.

(pg.

589

599

)

, vol.(pg.

Hunter

Explore more:
Nitrile Gloves vs Vinyl Gloves
Frequently Asked Questions About Supplement Vitamin Soft Capsules
Ivermectin API Solution: A Game-Changer for COVID-19? Get the Facts!
What is the price of 3m coverall?
What is the best form of tocopherol?
What is the ratio for pumpkin powder?
Can I use an ultrasound machine at home?

You will get efficient and thoughtful service from chenlv.

Haworth

The Ca2+-induced membrane transition in mitochondria. I. The protective mechanisms

Arch. Biochem. Biophys.

1979

, vol.

195

(pg.

453

459

)

, vol.(pg.

Juhaszova

Wang

Zorov

Nuss

Gleichmann

Mattson

Sollott

The identity and regulation of the mitochondrial permeability transition pore: Where the known meets the unknown

Ann. N. Y. Acad. Sci.

2008

, vol.

1123

(pg.

197

212

)

, vol.(pg.

Kamo

Muratsugu

Hongoh

Kobatake

Membrane potential of mitochondria measured with an electrode sensitive to tetraphenyl phosphonium and relationship between proton electrochemical potential and phosphorylation potential in steady state

J. Membr. Biol.

1979

, vol.

(pg.

105

121

)

, vol.(pg.

Kaneda

Torii

Tanaka

Aikawa

In vitro effects of berberine sulphate on the growth and structure of Entamoeba histolytica, Giardia lamblia and Trichomonas vaginalis

Ann. Trop. Med. Parasitol.

1991

, vol.

(pg.

417

425

)

, vol.(pg.

Khailova

Prikhodko

Dedukhova

Mokhova

Popov

Skulachev

Participation of ATP/ADP antiporter in oleate- and oleate hydroperoxide-induced uncoupling suppressed by GDP and carboxyatractylate

Biochim. Biophys. Acta

2006

, vol.

1757

(pg.

1324

1329

)

, vol.(pg.

Letasiova

Jantova

Cipak

Muckova

Berberine-antiproliferative activity in vitro and induction of apoptosis/necrosis of the U937 and B16 cells

Cancer Lett.

2006a

, vol.

239

(pg.

254

262

)

, vol.(pg.

Letasiova

Jantova

Miko

Ovadekova

Horvathova

Effect of berberine on proliferation, biosynthesis of macromolecules, cell cycle and induction of intercalation with DNA, dsDNA damage and apoptosis in Ehrlich ascites carcinoma cells

J. Pharm. Pharmacol.

2006b

, vol.

(pg.

263

270

)

, vol.(pg.

Letasiova

Jantova

Muckova

Theiszova

Antiproliferative activity of berberine in vitro and in vivo

Biomed. Pap. Med. Fac. Univ. Palacky Olomouc Czech Repub.

2005

, vol.

149

(pg.

461

463

)

, vol.(pg.

Lin

Tsai

Lee

Wang

Liou

Shyu

Berberine inhibits HIF-1alpha expression via enhanced proteolysis

Mol. Pharmacol.

2004

, vol.

(pg.

612

619

)

, vol.(pg.

Nadakavukaren

Chen

Increased rhodamine 123 uptake by carcinoma cells

Cancer Res.

1985

, vol.

(pg.

6093

6099

)

, vol.(pg.

O'Brien

Oliveira

Wallace

Inhibition of the adenine nucleotide translocator by N-acetyl perfluorooctane sulfonamides in vitro

Toxicol. Appl. Pharmacol.

2008

, vol.

227

(pg.

184

195

)

, vol.(pg.

Pereira

Branco

Matos

Pereira

Parke

Perkins

Serafim

Sardao

Santos

Moreno

, et al.

Mitochondrially targeted effects of berberine [Natural Yellow 18, 5,6-dihydro-9,10-dimethoxybenzo(g)-1,3-benzodioxolo(5,6-a) quinolizinium] on K1735-M2 mouse melanoma cells: Comparison with direct effects on isolated mitochondrial fractions

J. Pharmacol. Exp. Ther.

2007

, vol.

323

(pg.

636

649

)

, vol.(pg.

Ross

Da Ros

Blaikie

Prime

Porteous

SeverinaSkulachev

Kjaergaard

Smith

Murphy

Accumulation of lipophilic dications by mitochondria and cells

Biochem. J.

2006

, vol.

400

(pg.

199

208

)

, vol.(pg.

Serafim

Oliveira

Sardao

Perkins

Parke

Holy

Different concentrations of berberine result in distinct cellular localization patterns and cell cycle effects in a melanoma cell line

Cancer Chemother. Pharmacol.

2008

, vol.

(pg.

1007

1018

)

, vol.(pg.

Shirwaikar

Rajendran

Punitha

In vitro antioxidant studies on the benzyl tetra isoquinoline alkaloid berberine

Biol. Pharm. Bull.

2006

, vol.

(pg.

1906

1910

)

, vol.(pg.

Skulachev

Fatty acid circuit as a physiological mechanism of uncoupling of oxidative phosphorylation

FEBS Lett.

1991

, vol.

294

(pg.

158

162

)

, vol.(pg.

Szabo

Zoratti

The giant channel of the inner mitochondrial membrane is inhibited by cyclosporin A

J. Biol. Chem.

1991

, vol.

266

(pg.

3376

3379

)

, vol.(pg.

Taylor

Baird

Berberine inhibition of electrogenic ion transport in rat colon

Br. J. Pharmacol.

1995

, vol.

116

(pg.

2667

2672

)

, vol.(pg.

Tsai

Hepatobiliary excretion of berberine

Drug Metab. Dispos.

2004

, vol.

(pg.

405

412

)

, vol.(pg.

Turner

Gosby

Cheng

Miyoshi

Taketo

Cooney

Kraegen

James

, et al.

Berberine and its more biologically available derivative, dihydroberberine, inhibit mitochondrial respiratory complex I: A mechanism for the action of berberine to activate AMP-activated protein kinase and improve insulin action

Diabetes

2008

, vol.

(pg.

1414

1418

)

, vol.(pg.

Vergun

Reynolds

Distinct characteristics of Ca(2+)-induced depolarization of isolated brain and liver mitochondria

Biochim. Biophys. Acta