Transcriptional Profiles of the Response of Methicillin-Resistant Staphylococcus aureus to Pentacyclic Triterpenoids

Staphylococcus aureus is an important human pathogen in both hospital and the community that has demonstrated resistance to all currently available antibiotics over the last two decades. Multidrug-resistant isolates of methicillin-resistant S. aureus (MRSA) exhibiting decreased susceptibilities to glycopeptides has also emerged, representing a crucial challenge for antimicrobial therapy and infection control. The availability of complete whole-genome nucleotide sequence data of various strains of S. aureus presents an opportunity to explore novel compounds and their targets to address the challenges presented by antimicrobial drug resistance in this organism. Study compounds α-amyrin [3β-hydroxy-urs-12-en-3-ol (AM)], betulinic acid [3β-hydroxy-20(29)-lupaene-28-oic acid (BA)] and betulinaldehyde [3β-hydroxy-20(29)-lupen-28-al (BE)] belong to pentacyclic triterpenoids and were reported to exhibit antimicrobial activities against bacteria and fungi, including S. aureus. The MIC values of these compounds against a reference strain of methicillin-resistant S. aureus (MRSA) (ATCC 43300) ranged from 64 µg/ml to 512 µg/ml. However, the response mechanisms of S. aureus to these compounds are still poorly understood. The transcription profile of reference strain of MRSA treated with sub-inhibitory concentrations of the three compounds was determined using Affymetrix GeneChips. The findings showed that these compounds regulate multiple desirable targets in cell division, two-component system, ABC transporters, fatty acid biosynthesis, peptidoglycan biosynthesis, aminoacyl-tRNA synthetase, ribosome and β-lactam resistance pathways which could be further explored in the development of therapeutic agents for the treatment of S. aureus infections.


Introduction
Antibiotic resistance is a growing global threat to the effective treatment of infectious diseases. Over the past two decades resistance has been reported in many clinically important pathogens and to a broad range of antimicrobials in clinical use today [1,2,3]. This has been the driving force for the discovery of novel agents with antimicrobial activity that may potentially lead to therapeutic agents. The isolation of vancomycin, b-lactam, macrolide, tetracycline, aminoglycoside and quinolone-resistant strains of S. aureus provides a message that the availability of structurally novel antibacterials with multiple modes of action need to be part of the antibacterial armamentarium [4]. Over the past 30 years, only two new classes of antibiotics for the treatment of staphylococcal infections have been introduced: the oxazolidinones represented by linezolid, and the lipopeptide daptomycin [5]. However, reports of resistance to these antibiotics have also been reported [6]. Clearly, new approaches and innovative strategies are required to discover and identify agents that target essential bacterial targets in novel pathways, i.e. not targeted by currently used antibiotics, or novel targets in existing pathways. Promising targets for novel antibacterials against S. aureus include cell division, DNA replication and biosynthesis of fatty acid, peptidoglycan and protein [7].
To capture and exploit the wealth and diversity of genetic information of a variety of microorganisms, development of new technologies such as microarray provides a unique tool to understand the molecular basis of drug activity as it allows simultaneous analysis of the expression of all genes in the organism under any given condition. Transcriptional profiles generated by GeneChip analysis of bacteria have been used to investigate differential gene expression in response to antimicrobial agents [8], such as the response of S. aureus strain 8325-4 to antibiotics that act on cell wall such as oxacillin, D-cycloserine and bacitracin [9], and the effects of vancomycin on gene expression in S. aureus [10].
Plant-derived natural products have been reported to show antistaphylococcal activities with micromolar MICs [11,12,13,14,15,16]. In a preliminary study, the effects of 191 extracts towards different strains of bacteria were evaluated. Extracts from Callicarpa farinosa were shown to exhibit strong antimicrobial activities against MRSA [16]. Three pentacyclic triterpenoids, namely a-amyrin, betulinic acid and betulinaldehyde, have been investigated and found to have antibacterial activity against reference strains of methicillin-sensitive and methicillin-resistant S. aureus. Their activity was potentiated by currently used antistaphylococcal drugs [17]. In this study, the genome-wide transcription induced by the exposure of reference strain of S. aureus to pentacyclic triterpenoids was used to elucidate the mechanism of action of the compounds.

Bacterial Strains and Materials
The S. aureus strain used in this study was reference strain ATCC 43300, representing methicillin-resistant strains. The bacterial strain was maintained on Nutrient Agar (NA) (Difco, USA) slopes. Bacterial cultures used for microarray experiments were grown on cation-adjusted Mueller-Hinton broth (CAMHB II) (Difco, USA). Commercial sources of pentacyclic triterpenoids a-amyrin (AM) and betulinic acid (BA) were obtained from Sigma-Aldrich (USA), while betulinaldehyde (BE) was obtained from Advanced Technology and Industrial Co., Ltd (Hong Kong).

Growth of Staphylococcus aureus in the Presence of AM, BA and BE
In order to obtain good microarray results, the concentrations of the compounds need to be optimized at sub-inhibitory concentrations that do not affect the growth of the organism [18,19,20,21]. The treatment time with AM, BA and BE was determined by exposure of the organism to the study compounds and bacterial growth curve analysis. The bacterial culture was grown in CAMHB and incubated at 37uC with shaking at 200 rpm, to an optical density of 0.3 at 600 nm; i.e. at midlogarithmic phase as culture conditions at this phase were most constant and RNA levels were at the highest due to high metabolic rate. The optical density of the cultures was measured using a spectrophotometer (Nanophotometer, Germany). The cultures were then distributed into 100 ml aliquots. AM, BA and BE were added to the aliquots to obtain concentrations of 1/26 MIC (32 mg/ml for AM and BA; 256 mg/ml for BE) and 16 MIC (64 mg/ml for AM and BA; 512 mg/ml for BE) [17]. The cultures were then incubated, and cell growth was spectrophotometrically monitored at OD 600 nm for 5 hours at time intervals of 30 minutes. Untreated bacterial culture was used as control. The growth curve for each treatment and concentration was carried out in triplicates.

Total RNA Isolation and Purification
Three independent bacterial cultures for each treatment and control were prepared as biological replicates for total RNA isolation. After treatment of bacterial cells with the study compounds at concentrations and incubation times as determined from the growth curve, the cells were immediately stabilized with RNA Protect Bacteria Reagent (Qiagen, USA) to minimize RNA degradation before harvesting. Total RNA was isolated from treated and control bacterial cultures using the RNeasy Midikit (Qiagen) protocol. The cell wall of S. aureus was disrupted with Tris-EDTA (TE) buffer with 50 mg/ml lysozyme (Sigma Aldrich) containing proteinase K (Qiagen). DNA was removed using RNase-free DNase I (Qiagen). After purification, RNA concentration was determined by absorbance at 260 nm on Nanophotometer (1 absorbance unit is equivalent to 40 mg/ml RNA). RNA integrity was determined using Agilent RNA 6000 Nano chips (Agilent Technologies, USA). The procedure in preparing the assay, loading the chip and running the assay was carried out according to the manufacturer's protocol at Malaysia Genome Institute, Bangi, Malaysia.
The total RNA analysis showed that the A 260 /A 280 and A 260 / A 230 readings of all the total RNA samples were in the range of 1.8 to 2.1 and a single peak at 260 nm indicating that the samples were free from contaminants such as proteins, phenol and salts, which could interfere with down-stream applications such as cDNA synthesis and microarray analysis. The RIN values were above 7.0, indicative of non-degraded RNA (supporting information Table S1).
cDNA Synthesis, Fragmentation and Terminal Labeling cDNA was synthesized with a starting material of 10 mg of the total RNA, according to manufacturer's protocol (prokaryotic array, Affymetrix, USA). The procedure was carried out in a Mastercycler Gradient thermal cycler (Eppendorf, Germany). The cDNA obtained was then purified using a MinElute PCR Purification Column Kit (Qiagen) and quantified using Nanophotometer at 260 nm absorbance. A minimum of 1.5 mg cDNA was then fragmented in a reaction mix containing DNase I and DNase running buffer (Promega, USA). The fragmented cDNA was then applied directly to the terminal labeling reaction using GeneChip DNA Labeling Reagent (Affymetrix). The fragmented and labeled cDNA (target) was then used for hybridization.

GeneChip Hybridization
The labeled fragmented cDNA from independent RNA preparations were hybridized onto a Genechip probe array masked with known genome of S. aureus N315 (methicillin-resistant strain), Mu50 (methicillin and vancomycin-resistant strain), COL (methicillin-resistant strain) and NCTC 8325 strains (methicillinsensitive strain) (Affymetrix). The array contains probe sets of over 3300 S. aureus ORFs and over 4800 intergenic regions throughout the S. aureus genome. The labeled cDNA was hybridized, washed and scanned according to manufacturer's protocol (prokaryotic array, Affymetrix). The data obtained was analyzed using the Affymetrix GeneChip Command Console software (AGCC).

Microarray Analysis
For gene expression analysis, data from the Affymetrix platform were processed with GeneSpring version 11 software. Data was normalized to minimize systematic non-biological differences to reveal true biological differences. Significance analysis using ANOVA test with multiple testing correction Benjamini Hochberg to correct false discovery rate and asymptotic p-value computation was performed to determine the change of expression of a particular gene due to the effect of different treatment on S. aureus. In addition, fold-change analysis was also performed and reported as the up-or down-regulated fold change. To select the differentially expressed genes, threshold values of $2 fold-change between pentacyclic triterpenoid-treated samples and control samples and significance level of ,5% were used. From the statistical analysis, the gene ontology (GO) analysis was carried out. The pathways affected by the differentially expressed genes were analyzed using the KEGG database (http://www.genome. jp/kegg/pathway.html). To group genes with similar expression profiles together, hierarchical clustering analysis was used in which an entity tree and a condition tree were generated.

Validation of Gene Expression by Quantitative Real-time Quantitative PCR
To determine the validity of the microarray data, transcript level changes obtained with the microarray analysis were compared with those from quantitative real-time PCR (qPCR). Aliquots of cDNA preparations used in the microarray experiments were used for qPCR procedure. A list of genes and the primer sequences used for the real-time PCR analysis is shown in Table 1. The housekeeping gene adhE and gyrA were used as endogenous reference genes. The procedure was performed using the iCycler iQ5 Real Time PCR Detection System (Bio-Rad Laboratories, USA) with Quantitect SYBR Green PCR kit (Qiagen). For each gene, three biological replicates with three technical replicates each were used. Real-time cycler conditions were carried out according to manufacturer's protocol: initial activation step at 95uC for 15 min, followed by 45 cycles of denaturation at 15 s and 94uC, annealing at 30 s and 53 to 60uC (as optimized for each target gene, data not shown), and extension at 30 s and 72uC. Melting curve analysis was also performed to verify PCR specificity. PCR efficiencies were derived from standard curve slopes in the iCycler software v3.1. Relative quantification based on the relative expression of the target gene compared to reference genes adhE and gyrA in treated and control samples was used to determine transcript level changes [22]. Paired student's T test was performed to assess whether the means of a treated and control group (samples) is significantly different from each other using the SPSS software (PASW Statistic 18).

Growth Profile of Staphylococcus aureus in the Presence of AM, BA and BE
The effect of AM, BA and BE on the growth of reference strain of MRSA is as shown in Figure 1. The growth rate of the organism decreased at concentrations K MIC and MIC after 30 min for AM; and 45 min for BA and BE. After 300 min, the growth of the organism at K MIC was generally higher than the growth at MIC. These effects of the growth indicates that AM, BA and BE are bacteriostatic compounds as reported by Chung et al. [17]. Based on the growth curves, the organism was then treated with the three study compounds at K MIC (32 mg/ml for AM and BA; and 256 mg/ml for BE) and further incubated for 30 min (control and treatment with AM) and 45 min (treatment with BA and BE). Subinhibitory concentrations of the compounds were used in the microarray procedure to avoid secondary responses caused by high concentrations of the compounds.

Overview of Transcriptional Profiles of Staphylococcus aureus Treated with Pentacyclic Triterpenoids
From the data processed using GeneSpring version 11 software, an enormous number of genes were significantly differentially regulated (p,0.05, fold change $2) in response to treatment with AM, BA and BE as compared to control ( Table 2). For example, a total of 1265 genes were significantly affected in the treatment of reference methicillin-resistant strains with BE. There is also a vast number of the genes which are uncharacterized, without specific functional category and pathways, and with unknown function (Tables S2, S3).
Overall, the changes in expression levels of most genes were moderate with ratio in the range of 2-to 3-fold (Tables 3, 4, 5, 6). There are 31 genes coding for ribosome which were highly downregulated (more than 5-fold and up to 21-fold), indicating that ribosome assembly may play a prominent role in the response of S. aureus to BE. Four genes in the fatty acid biosynthesis and three genes in the peptidoglycan biosynthesis were down-regulated more than 3-fold while three genes in the ABC transporters were at least 3-fold up-regulated after treatment with BE. ABC transporters and ribosomes are significantly regulated by all three pentacyclic triterpenoids. Only BE affects cell division, while only BA targets genes in b-lactam resistance and DNA replication pathways.
Differentially expressed genes in promising pathways such as cell division (ftsZ), ABC transporters (opp-1C), biosynthesis of fatty acid (fabZ), peptidoglycan (fmhB, pbp2) and protein (asnC), DNA replication (ccrA) and b-lactam resistance (mecR1) were selected to be validated by real-time PCR. The paired student's T test analysis showed that all these selected genes were extremely statistically significantly different between bacterial cells treated with AM, BA or BE, and untreated cells. The difference in the expression of reference genes adhE and gyrA was considered not significantly different (Table S4). A comparison of genes identified in the treatments of all three study compounds showed that genes rpmG and kdpA, which code for the protein of the large subunit 50S and potassiumtransporting ATPase subunit A, respectively, are down-regulated. AM and BA showed up-regulation of opp-1C in the ABC transporter pathway, while kdpE in the two-component systems are similarly down-regulated in BA and BE. Treatment of the bacterial strain with AM and BE affected the highest number of similarly regulated genes, involving a total of seven genes in the two-component systems, fatty acid biosynthesis, peptidoglycan biosynthesis, aminoacyl-tRNA biosynthesis and ribosome pathways (Table 7). Generally, all three study compounds acted on more than one target/pathway that is unique to the microorganisms. This is a highly desirable property of an antibacterial agent and it could potentially delay the development of bacterial resistance to the compounds [23,24].

Validation of Gene Expression by qPCR
The target genes selected for validation represent significant pathways in S. aureus identified as promising targets in the literature, and two genes selected as reference genes because they were not differentially regulated by AM, BA and BE. The target genes were ftsZ (cell division), opp-1C and oppF (ABC transporters), asnC (aminoacyl-tRNA biosynthesis), fabZ (fatty acid biosynthesis), fmhB and pbp2 (peptidoglycan biosynthesis), ccrA (DNA replication), and mecR1 (b-lactam resistance), while the reference genes were adhE (metabolism of fatty acid and glycolysis) and gyrA (DNA replication). All the target and reference genes validated showed single peaks in melting curves, indicating the amplification of specific amplicons (data not shown). The high specificity of the reactions was further confirmed by the single band and size of the amplicons observed in the agarose gel electrophoresis (Figure 2). The expression of the target genes was quantified using the relative quantification method which measures the changes in the expression of the target genes in response to the different treatment normalized to the expression of reference gene(s). The expression of genes opp-1C, oppF, asnC and mecR1 was observed at greater fold-change by real-time PCR than with microarray analysis. For three other genes fabZ, fmhB and ccrA, the levels of gene expression did not differ markedly between the microarray data and real-time PCR analysis. Genes ftsZ and pbp2 showed higher gene expression level in microarray analysis. The comparison of the gene expression level by qPCR and microarray analysis is shown in Table 8.

Novel Pathways Targeted by AM, BA and BE
Cell division. BE inhibits genes encoding proteins FtsZ and FtsA in the cell division process of the bacterial cell. FtsZ is an essential division protein and forms a ring structure that recruits further cell division proteins. Cell division starts with the polymerization of FtsZ into the Z ring, which is the heart of the division apparatus and serves as a landing pad for the recruitment of other proteins to the division site. Suppression of FtsZ prevents     Z rings at membrane. Thus, suppression of FtsZ and FtsA proteins inhibits growth and could exert bactericidal effect [26]. Two-component system. From the microarray analysis, AM, BA and BE have the potential to target the virulence factors of MRSA, which includes structures that promote adhesion, colonization and dissemination; the components responsible for cell communication and regulatory functions; the factors that facilitate evasion of the immune system; and the machineries for toxin production and excretion [27]. Expression of these factors is regulated by a range of global regulators comprising twocomponent regulatory systems which are sensitive to environmental signals [28,29,30]. The study compounds regulates three of the systems of the 16 two-component systems identified by genomic scans in S. aureus, i.e. Agr, Sae and Kdp.
The expression of gene agrD and inhibition of agrA, as observed in the treatment with BE, leads to the decreased production of capsule, toxins and proteases, thus decreasing the virulence of the organism. AM which inhibits saeR and saeS also decreases virulence as a result of reduction in the production in several exoproteins and cell-bound proteins involved in pathogenicity [31]. With the inhibition of kdpA and kdpE as observed in the treatment with BA and BE, the transcription of genes encoding cell wall proteins and polysaccharides, which are beneficial to colonization, are repressed [32]. As a result, the virulence of the organism is reduced with the exposure to each of the compounds.
ABC transporters. The up-regulation of oligopermease transporter oppC by AM and BA, and oppF by BE may increase the efficient transport of peptides for the salvage of essential amino acids from the environment of the bacterial cell. The uptake of these peptides could be to compensate for the halt in protein synthesis due to the inhibition of the aminoacyl-tRNA synthetase and ribosomes. The up-regulation of gene sirB as observed in the treatment with BE, indicates there could be an uptake of iron (III)staphylobactin to satisfy its requirement for iron, which is an essential micronutrient for bacteria and serves as the catalytic center of enzymes involved in critical cellular processes such as DNA synthesis and electron transport [33]. This may represent an attempt by the organism to acquire iron for successful establishment of infection, as a response to the environmental stress caused by BE.

Novel Targets in Existing Pathways Regulated by AM, BA and BE
Fatty acid biosynthesis. BE down-regulates a group of genes in the type II fatty acid synthesis (FAS) pathway, which include genes encoding for enzymes that are involved in the assembly of important cellular components in the bacteria, such as phospholipids, lipoproteins, lipopolysaccharides, myolic acids and the cell envelope [34]. The inhibition of enzymes encoded by fabF, fabI, fabD, fabG and fabZ may cause modification of the permeability of the phospholipid bilayer and destabilization of the cell membrane. As a result, processes such as passive transport of hydrophobic molecules, active transport of solutes and proteinprotein interactions could be affected [35]. Thus, AM and BE may be able to enter the cells relatively easier via passive transport. However, due to the high partition coefficients of AM and BE at log P values of 11.01 and 9.07, respectively (Chemspider database, http://www.chemspider.com/Search.aspx), the entry of these compounds may still not increase significantly, thus the high observed MIC values at 64 and 512 mg/ml as compared to current anti-staphylococcal agents such as methicillin and vancomycin.
Peptidoglycan biosynthesis. Closely linked to cell division is peptidoglycan synthesis. Coordinated synthesis of the peptidoglycan is crucial for the integrity of the bacterial cell, attachment sites for virulence factors and resistance to external stress conditions [36]. BE inhibits three mur genes (murA, murC and murG), fmhB and pbp2 which are involved in the synthesis initiation phase, modification, and polymerization and cross-linking of lipid II, respectively. The suppression of the synthesis of lipid II which is the substrate for Pbp2 and transpeptidation reactions results in the weakening of the peptidoglycan, leading to cell lysis and death. The majority of approved anti-staphylococcal drugs that target the peptidoglycan biosynthesis act on extracellular steps and resistance to these drugs are on the rise. An attractive alternative is to exploit the intracellular steps, involving the mur and fem genes. There are currently no known antibiotics or synthetic chemicals of therapeutic usefulness that inhibits the Mur enzymes, except for MurA which is inhibited by fosfomycin [37].
Aminoacyl-tRNA biosynthesis. The suppression of six aminoacyl-tRNA synthetase, namely cysS, proS, asnC, thrS, serS and lysS, by AM and BE causes the accumulation of the deacylated form of tRNA. The shortage of corresponding aminoacyl-tRNA of a required amino acid prompts the stringent response with increased levels of guanosine tetraphosphates (ppGpp) and guanosine pentaphosphates (pppGpp). RNA polymerase is inhibited, thus directly reducing the production of RNA. This leads to subsequent inhibition of all key downstream cellular processes such as the biosynthesis of protein and peptidoglycan [38]. By perturbing several metabolically important processes, the inhibition of aminoacyl-tRNA synthetase leads to the cessation of bacterial growth and the attenuation of virulence in vivo [39].
Ribosome. The assembly of the ribosome begins with the transcription of rRNA [40] and proceeds through the synthesis of ribosomal proteins [41]. In comparison to the other study compounds, BE inhibits the expression of the highest number of genes in the ribosome, with 13 and 23 of the 30S and 50S  subunits, respectively. The compounds could bind to the precursor forms of rRNA and ribosomal proteins and stall the normal assembly sequence of the ribosome. As a result, continued translation could be reduced, leading to inhibition of protein synthesis and prevention of growth of the bacteria. The interference with the assembly of the ribosomal subunit from its constituent rRNA and proteins could be a new target for inhibition of antibacterial agents.
b-lactam resistance. MecA, the main determinant for methicillin resistance, is inducible and modulated by a signaltransduction system encoded by genes from the SCCmec [42], which comprises a sensor/signal transducer MecR1, and a transcriptional repressor, methicillin repressor MecI [43]. Treatment with BA causes the expression of gene mecR1 that results in the production of fully functional MecR1 and MecI that appear to repress the production of pbp2a [44,45]. The observation indicates that BA could be modified and developed into a more potent  inhibitor of penicillin-binding protein (PBP) that can target MRSA. In addition, BA has the potential to restore the in vitro activity of b-lactams and glycopeptides against MRSA, as shown in the strong synergistic interactions with methicillin and vancomycin [17].

Conclusions
Microbial genomics has allowed the identification of a greater number of desirable and essential antibacterial targets for further evaluation. From the transcriptional profile, AM, BA and BE regulate multiple desirable targets in essential pathways which  could be further explored in the development of therapeutic agents for the treatment of S. aureus infections. In addition, the structure of the study compounds differ from the currently used antistaphylococcal agents, and thus cross-resistance is unlikely and development of resistance to these novel compounds will be delayed. While the compounds have the ability to affect multiple and essential targets, the relatively high MICs in S. aureus suggests that the permeability of the compounds into the bacterial cell may not be optimum and compound modifications may be required to improve the permeability. A more detailed examination of these gene responses is also needed to further characterize the antimicrobial action of each of the regulated pathways.

Supporting Information
Table S1 Quantification and quality assessment of total RNA isolated from reference strain of methicillin-resistant Staphylococcus aureus treated with a-amyrin, betulinic acid and betulinaldehyde at K6 MIC and control. (DOCX)