NUR 410

Research Article Mechanism of Antibacterial Activity of Bacillus amyloliquefaciens C-1 Lipopeptide toward Anaerobic Clostridium difficile

Jia Lv,1 Rong Da,2 Yue Cheng,1 Xiaohong Tuo,1 Jie Wei,1 Kaichong Jiang,1

Adediji Omolade Monisayo,1 and Bei Han 1

1School of Public Health, Health Science Center, Xi’an Jiaotong University, Xi’an, China 2Department of Clinical Laboratory, The First Affiliated Hospital of Xi’an Jiaotong University, Xi’an, China

Correspondence should be addressed to Bei Han; [email protected]

Received 11 November 2019; Revised 13 January 2020; Accepted 4 February 2020; Published 4 March 2020

Academic Editor: Frederick D. Quinn

Copyright © 2020 Jia Lv et al. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

Probiotics may offer an attractive alternative for standard antibiotic therapy to treat Clostridium difficile infections (CDI). In this study, the antibacterial mechanism in vitro of newly isolated B. amyloliquefaciens C-1 against C. difficile was investigated. The lipopeptides surfactin, iturin, and fengycin produced by C-1 strongly inhibited C. difficile growth and viability. Systematic research of the bacteriostatic mechanism showed that the C-1 lipopeptides damage the integrity of the C. difficile cell wall and cell membrane. In addition, the lipopeptide binds to C. difficile genomic DNA, leading to cell death. Genome resequencing revealed many important antimicrobial compound-encoding clusters, including six nonribosomal peptides (surfactins (srfABCD), iturins (ituABCD), fengycins (fenABCDE), bacillibactin (bmyABC), teichuronic, and bacilysin) and three polyketides (bacillaene (baeEDLMNJRS), difficidin (difABCDEFGHIJ), and macrolactin (mlnABCDEFGHI)). In addition, there were other beneficial genes, such as phospholipase and seven siderophore biosynthesis gene clusters, which may contribute synergistically to the antibacterial activity of B. amyloliquefaciens C-1. We suggest that proper application of antimicrobial peptides may be effective in C. difficile control.

1. Introduction

Clostridium difficile is an anaerobic, gram-positive, spore- forming bacterium. Clinical signs of C. difficile infection (CDI) range from mild diarrhea to fulminant colitis [1]. The incidence and severity of CDI have increased signifi- cantly, especially by the recently emerged and highly viru- lent epidemic strain BI/NAP1/027 [2]. With increasing antibiotic resistance of C. difficile, there is an urgent need to develop new agents and efficient methods for the treat- ment and control of CDI [1, 3]. Distinct from the traditional antibiotics, many novel antimicrobial agents, such as ramo- planin, surotomycin, and cadazolid, are currently being investigated in clinical trials for the treatment of CDI. Suro- tomycin is an orally dosed lipopeptide antibiotic that acts by disrupting the cell membrane [4].

Bacteriocins are ribosomally synthesized antimicrobial peptides with high activity against other bacteria. Bacterio- cins are secreted by some probiotic microorganisms, such as Lactobacillus species, Saccharomyces boulardii, and bifido- bacteria [5]. Thus, further evaluation should be given to the bacterial resources, antimicrobial mechanisms, and biosafety of probiotics in considering bacteriocins as an alternative or adjunctive therapeutic method for CDI.

Except the ribosomally synthesized antimicrobial pep- tides, Bacillus species could synthetize a mixture of lipopep- tides by nonribosomal peptide synthetases, which mainly include surfactin, iturin, lichenysin, and fengycin families, with broad-spectrum biological activities [6–8]. With the described functional secondary metabolites, many Bacillus spp. strains have been developed as biofertilizers and bio- pesticides, and they are currently regarded as promising

Hindawi BioMed Research International Volume 2020, Article ID 3104613, 12 pages https://doi.org/10.1155/2020/3104613

environmentally friendly means for plant protection and plant growth promotion, and as secondary metabolite facto- ries [9]. However, for B. amyloliquefaciens, there are only a few reports that describe an antimicrobial activity against toxin-producing C. difficile [10].

We isolated B. amyloliquefaciens C-1 from ready-to-eat fruit samples. The bacterial strain is stored in the China Center for Type Culture Collection as a patent strain with the number of CCTCC M2010177. The supernatant of C-1 showed high antioxidant activity and inhibitory activity against foodborne pathogens (Escherichia coli O157:H7, B. cereus, S. aureus, etc.) and human pathogens (C. difficile, Klebsiella pneumoniae, Enterococcus faecium, etc.) [11, 12]; however, no effect against fungi was found.

Comparative genomic analysis showed evolutionary traits for B. amyloliquefaciens strain adaptation to host habi- tats [13]. The C-1 strain exhibited a biosurfactant activity phenotype against pathogens. The molecular bases/mechan- isms of this pathogen-specific activity were unknown. In this study, we investigated the anti-C. difficile mechanisms of the secreted B. amyloliquefaciens C-1 extracellular lipopeptides. We systematically investigated the effects of C-1 lipopeptide on C. difficile cell growth, morphological structure, cell wall and membrane integrity, and genome. Then, we resequenced the entire C-1 genome and identified relevant gene clusters, locations, and potential regulatory sequences, including genes for bacteriocins, ribosomally synthesized antibacterial peptides, phospholipase, siderophores, and genes that pro- vide resistance to toxic compounds.

2. Material and Methods

2.1. Bacterial Strains and Culture. B. amyloliquefaciens C-1, a patent strain (Chinese patent no. ZL201410260574.2), was isolated and stored in our lab. It was inoculated into fermentation medium (12.4 g/l tryptone, 20 g/l glucose, 5 g/l NaCl, 1.5 g/l K2HPO4·3H2O, 0.04 g/l MnSO4·H2O, 1.7 g/l FeSO4·7H2O, and 1.2 g/l MgCl2·6H2O, pH 7.2) and grown with shaking of 200 rpm in flasks for 72 h at 30°C. Clostrid- ium difficileATCC 9689, 700057, and BAA-1870 strains were obtained from the American Type Culture Collection and stored at -80°C. C. difficile strains were cultured in sterile Reinforced ClostridiumMedium (RCM) and incubated over- night at 35°C in an anaerobic chamber (Coy Laboratory Products Inc., Ann Arbor, Michigan) with an atmosphere of 82% N2, 15% CO2, and 3% H2.

2.2. Isolation and Identification of B. amyloliquefaciens C-1 Extracellular Lipopeptide. Lipopeptide isolation was per- formed by acid precipitation according to Zhang et al. [14]. Briefly, the crude cell-free culture was adjusted to pH 2.0 with 6M HCl and placed overnight at 4°C. After centrifugation, the precipitate was extracted twice with methanol. The solution was dried in a vacuum freeze dryer, and the dry residue was dissolved in 50mM Tris-HCl (pH 7.5) and passed through a 0.22μm filter. The extracted lipopeptide sample was analyzed with a UV-VIS spectrophotometer (UV5, Mettler Toledo).

2.3. Thin-Layer Chromatography. The purified lipopeptide was examined by thin-layer chromatography (TLC) on a sil- ica gel G plate [15]. TLC assay was performed on a silica gel G plate (10 × 20 cm, Silica gel 60, Germany). A chloroform- methanol mixture (10 : 1, v/v) was used as the mobile phase. A sample was spotted onto the TLC plate and hydrolyzed with 6M HCl for 2 h in a sealed container. Once dried, the plate was developed in the mobile phase. After development, the plate was sprayed with 0.5% ninhydrin and placed in an oven at 110°C for 10min to detect the peptides as red spots.

2.4. Semipreparative HPLC Analysis. The putative lipopep- tides were identified by HPLC analysis. Briefly, crude extract spots were removed from the TLC plates and dissolved in 10% methanol; the supernatants were analyzed by semipre- parative high-pressure liquid chromatography using an Agi- lent LC 1200 system. The chromatographic separation was performed with a C-18 Column (4:6 × 250mm). The column outlet was coupled to an Agilent MSD Ion Trap XCT mass spectrometer equipped with an ESI ion source. The lipopep- tide fragments were selectively desorbed with methanol gradients from 35% to 65% within 140min. All elution pro- grams used a flow rate of 0.5ml/min at 214 nm and detection occurred using the negative ion mode at m/z ranging from 400 to 2000. The isolated fragments were collected for the fol- lowing experiments.

2.5. Detection of Lipopeptide Synthesis-Related Genes. Lipo- peptide biosynthesis genes (sfr, ituD, and fenB) were identi- fied by PCR (sfr-F: 5′ ATGAAGATTTACGGAATTTA 3′, sfr-R: 5′ TTATAAAAGCTCTTCGTACG 3′; ituD-F: 5′ ATGAACAATCTTGCCTTTTTA 3′, ituD-R: 5′ TTATTT TAAAATCCGCAATT 3′; fenB-F: 5′ CTATAGTTTGT TGACGGCTC 3′, fenB-R: 5′ CAGCACTGGTTCTTGT CGCA 3′) [9]. PCR conditions consisted of an initial dena- turation step at 94°C for 5min followed by 30 cycles of denaturation at 94°C for 1min, 54°C annealing for 45 sec (sfr, ituD) or 1min (fenB), and 72°C extension for 1min followed by a final extension step at 72°C for 7min. The amplified PCR product was purified and sequenced by an automated sequencer (3730 DNA Analyzer). PCR product sequences were identified using GenBank nucleotide data and BLAST from the National Center for Biotechnology Information, Bethesda, MD, USA (http://www.ncbi.nlm .nih.gov/blast/).

2.6. The Inhibitory Activity of Lipopeptide against C. difficile. Antimicrobial activities of the lipopeptides were detected by disc diffusion assay. 0.5McF (106CFU/ml) of C. difficile cells was inoculated onto the surface of blood agar plates. An Oxford cup (6mm diameter), containing 100μl lipopeptide with concentrations of 5, 10, and 15μg/ml dissolved in 10% methanol, was placed on test C. difficile-seeded plates. A cup containing 100μl 10% methanol was used as negative control. Each C. difficile strain was plated in triplicate. The plates were incubated anaerobically overnight at 35°C, and antimicrobial activity was evaluated by measuring inhibition zones against the tested C. difficile cells. The minimal inhibi- tory concentrations (MIC) against C. difficile strains were

2 BioMed Research International

determined by the broth microdilution method in 96-well microplates with a final concentration of 105CFU/ml, and the final concentration of the added lipopeptide ranged from 10μg/ml to 0.0095μg/ml; bacterial growth without lipopeptide was set as control [16]. The MIC was defined as the lowest lipopeptide concentration at which growth was completely inhibited after overnight anaerobic incuba- tion of the plates at 35°C.

2.7. Growth of C. difficile Incubated with Lipopeptides. For time-kill analyses, 0.5McF (106CFU/ml) of C. difficile cells (strains ATCC 9689, ATCC 700057, and ATCC BAA-1870) was prepared and inoculated into fresh RCM containing 0, 0.25, 0.5, and 0.75 MIC of lipopeptide separately, and incu- bated in the anaerobic chamber at 35°C. Cell viability was determined every 2 h for 24 h [17]. Each treatment was performed with three biological replicates.

2.8. Scanning Electron Microscope Analysis of C. difficile Cells Treated with Lipopeptides. An overnight culture of C. difficile was transferred into fresh RCM medium with 0.5McF and 0.25MIC of lipopeptide, and incubated anaerobically at 35°C for 1 h. The C. difficile cells were collected and washed three times with sterile PBS solution. The cells were fixed with 2% glutaraldehyde overnight at 4°C and then dehy- drated by a graded series of ethanol (50%, 70%, 80%, 90%, 95%, and 100%) for 20min at each step. After the critical point of drying and gold coating, the surface structure of treated C. difficile cells was observed with a scanning electron microscope (Hitachi S-2460N, Hitachi, Ltd., Tokyo, Japan) at an acceleration voltage of 20 kV [18].

2.9. Fluorescence Microscope Analysis of C. difficile Cells Treated with Lipopeptides. Propidium iodide (PI) penetrates only damaged cell membranes, whereupon it binds to double-stranded DNA and fluoresces red with 488nm illu- mination. To clearly detect an effect on the plasma mem- brane, 0.5MIC of the purified lipopeptide was incubated with C. difficile cells (107 cells/ml) in RCM liquid medium. The mixture was incubated anaerobically at 35°C for 1 h. Then, 10μl of 100μg/ml PI solution was added to the cell suspension, and the mixture was incubated for 30min in the dark. Finally, the treated C. difficile cells were observed with a Nikon TI-S fluorescence microscope with the filters set at an excitation wavelength of 488nm and an emission wavelength of 538nm. Cells treated with the same amount of sterile water were used as a negative control [19]. All experiments were repeated three times.

2.10. Determination of Extracellular Alkaline Phosphatase Activity of C. difficile Cells Treated with Lipopeptide. An over- night culture of C. difficile was subcultured into fresh RCM liquid medium. Lipopeptide was added separately into the C. difficile culture at final concentrations of 0.25 and 0.5MIC. Bacterial supernatant (0.5ml) was collected every 12 h for the measurement of extracellular alkaline phospha- tase (AKPase) activity using an AKP assay kit (Nanjing Jian- cheng Technology Co., Ltd., Nanjing, China) as described in [20]. The AKPase unit was defined as 1mg of phenol pro- duced by 100ml of bacterial culture supernatant reacted with

the substrate at 37°C for 15 minutes. Cells treated with the same amount of sterile water were used as negative control. Each test was performed in three biological replicates.

2.11. Lipopeptide Binding to C. difficile Genomic DNA. Gel retardation experiment assays were performed to identify the DNA binding activity of the lipopeptide as described in [21]. Briefly, 50ng of C. difficile genomic DNA was mixed with 1μl of 1, 2, and 5μg/ml lipopeptide in 20μl of binding buffer (10mM Tris-HCl, 1mM EDTA buffer, pH 8.0). One μl of sterilized water mixed with C. difficile genomic DNA was used as negative control. Mixtures were incubated at 35°C for 1 h. All samples were subjected to 1.0% agarose gel electrophoresis and stained with ethidium bromide.

2.12. Whole Genome Sequencing of B. amyloliquefaciens C-1. The genomic DNA of strain C-1 was isolated and purified by a kit (Applied Biosystems® 4413021) and sequenced on the Pacific Bioscience (PacBio) RS II system at Genefund, Shang- hai, China. The genome was assembled with SMRT analysis v.2.3 and RS_HGAP_Assembly.3, and the genome assembly was improved by using the software Pilon. Identification of protein-coding open reading frames (ORFs) and annotation of ORFs were performed by using the NCBI Prokaryotic Genome Annotation Pipeline. Genes were functionally annotated by BLAST search in COG (Gene Ontology Con- sortium), Nr (NCBI RefSeq), and Pfam Databases [22, 23]. KEGG (Kyoto Encyclopedia of Genes and Genomes) data- base was used in the analysis of metabolic pathways of lipopeptide-producing Bacillus species. All amino acid sequences derived from the Bacillus genomes were submitted to the KEGG database, and the metabolic functions of these sequences were annotated by KASS. The KO (KEGG Orthol- ogy) term and corresponding KEGG pathway for each ORF were automatically generated. Secondary metabolite clusters present in the genome of the B. amyloliquefaciens collection have been evaluated using antiSMASH 5.0 [24].

The 16S rRNA gene sequences of Bacillus species were extracted from genome sequences and aligned using the CLUSTALX [25]. Phylogenetic trees were constructed using the neighbor-joining method implemented in the software package MEGA version 7.0.26 [26]. Evolutionary distances were calculated using Kimura’s two-parameter model.

The C-1 genome sequence data were deposited into the Sequence Read Archive (SRA) of NCBI and can be accessed via accession number SRP127533.

2.13. Statistical Analysis. All experimental data are expressed as the average with standard deviation of at least three inde- pendent replicates. Statistical analyses were performed using the t-test and analysis of variance (ANOVA), JMP pro (SAS Institute Inc., NC, US), STAMP10, and SPSS V20.0 (IBM Inc., IL, US). Significant differences were considered at P < 0:05.

3. Results and Discussion

3.1. Production, Purification, and Identification of C-1 Lipopeptide. Bacillus spp. produces abundant secondary metabolites, such as proteinase, amylase, bacteriocin, and exopolysaccharide. In this study, we focused on the

3BioMed Research International

lipopeptide-producing B. amyloliquefaciens strain C-1. The growth profile of C-1 is shown in Figure 1. The maximum growth and cell dry weight were reached at 48 h; the maxi- mum lipopeptide production (3:49 ± 0:26mg/ml) was reached at 72 h. Lipopeptide production was conducted in a fermentation medium. The active compound from the cul- ture supernatant was scanned from 190 to 900nm, and the maximum absorption occurred at 213 nm, which is the typi- cal absorption wavelength of peptides. Three spots were observed by thin-layer chromatography, and three compo- nent peaks were detected by RP-HPLC. Mass spectroscopy showed that the molecular masses of the three components at m/z were 1067Da, 1477Da, and 1506Da, which corre- sponded to surfactin, fengycin, and iurin, respectively. PCR products of 675 bp, 1400 bp, and 482 bp corresponded to Srf, FenB, and ItuD genes, and the PCR fragments were sequenced and showed 99% identity with surfactin, fengycin, and iturin biosynthesis gene clusters, individually (S1).

3.2. Antimicrobial Activity of C-1 Lipopeptide against C. difficile. Bacillus spp. lipopeptides have an inhibitory activity against plant pathogenic fungi and pathogenic bacteria and have been developed as biocontrol agents [27]. Although, in our previous report, C-1 did not show any inhibitory activity against fungi, it did have antibacterial activity toward several human pathogens, and this antibacterial activity in the C-1 supernatant was verified to be the contribution of lipopep- tide, not exopolysaccharide [11, 12].

In plate tests, the C-1 lipopeptide displayed antagonistic activities against three C. difficile strains. Inhibition zone diameters ranged between 7.05mm and 22.00mm, and the largest inhibition zone was from 15μg/ml lipopeptide against strain C. difficile ATCC 9689 (Table 1). The MICs against C. difficile strains ATCC 9689, ATCC 700057, and ATCC BAA- 1870 were 0.75, 2.5, and 2.5μg/ml, separately. Within a cer- tain concentration range (0.0095μg/ml-10μg/ml), the inhib- itory effect was positively correlated with the concentration of the C-1 lipopeptide. To analyze the inhibitory effect, we determined the growth curves of the three C. difficile strains. At 24 h of continuous measurement, the maximum OD600 of C. difficileATCC 9689, ATCC 700057, and ATCC BAA-1870 treated with 1/4 MIC lipopeptide reached 51.57%, 51.54%, and 56.12% of the control. For the 1/2 MIC treatment, growth was reduced to 43.15%, 46.39%, and 46.12% of the control. For the 3/4 MIC treatment, growth was further reduced to 38.95%, 40.21%, and 41.84% of the control. The C. difficile ATCC9689 was significantly the most sensitive to the treatment (P < 0:01), and the inhibitory effect was stronger with an increased concentration of the C-1 lipopep- tide (Figure 2).

Because C. difficile ATCC9689 was more sensitive to the C-1 lipopeptide, and it was also a tcdA and tcdB positive strain, we assessed the following antibacterial mechanism toward this strain.

3.3. Effect of C-1 Lipopeptide on C. difficile Morphology, Cell Wall Permeability. Scanning electron microscopy showed that the surface of C. difficile ATCC 9689 was damaged after treatment with the C-1 lipopeptide. Exudates surrounded the

bacteria, and the cell wall and cell membrane were inter- rupted and indistinct, whereas untreated cells were smooth and uninterrupted (Figures 3(a) and 3(b)). In addition, with increasing concentrations of lipopeptide, the bacteria were surrounded by exudate that may have been cytoplasm extruded from the cells. Propidium iodide (PI) penetrates only a damaged cell membrane, after which it binds to double-stranded DNA and fluoresces red with illumination at 488nm. We stained the C-1 lipopeptide-treated and C-1 lipopeptide-untreated C. difficile cells with PI and observed the cells with a fluorescence microscope. C. difficile ATCC 9689 cells treated with lipopeptide were stained with PI as shown by red fluorescence, which indicated a damaged cell membrane (Figures 3(d) and 3(f)), whereas the untreated cells did not show any fluorescence (Figures 3(c) and 3(e)).

The destroyed cell membrane increased the permeability of C. difficile ATCC 9689 after C-1 lipopeptide treatment. To verify the effect of lipopeptide on cell wall permeability, we measured alkaline phosphatase (AKPase) activity. A dam- aged cell wall and cell membrane increased cell permeability and caused an increase in extracellular AKPase. After lipo- peptide treatment, extracellular AKPase activity increased continuously and was significantly higher than that of the untreated cells (Fig S2). With 36 h of incubation, the extracel- lular AKPase content of C. difficile ATCC 9689 treated with 1/4 MIC and 1/2 MIC of lipopeptide increased 4.7-fold and 7.7-fold.

The antibacterial activity of Bacillus spp. lipopeptide was observed with other pathogens, such as S. aureus [28], Vibrio anguillarum [18], and E. clocae [8]. However, our report is the first to document the effects of lipopeptide on toxin- producing C. difficile. The inhibitory mechanism of the C-1 lipopeptide against C. difficile could be explained by destroy- ing the bacterial cell by forming ion-conducting channels in the cell membrane as described by Etchegaray et al. [29]. This mode of action drastically reduces the chance of the

72 0.0 0

1

2

3

4

0.5

1.0

1.5

C ell

d ry

w ei

gh t (

m g)

Li po

pe pt

id e (

m g/

m l)

2.0

2.5

3.0

3.5

6048 Time (h)

3624120 0.0

0.5

1.0

1.5

O D

60 0

2.0

2.5

Lipopeptied OD600

Cell dry weight

Figure 1: The fermentation and lipopeptide production in B. amyloliquefaciens C-1 for 72 hours. The solid line indicates the growth curve of OD600, the dotted line indicates the growth curve of cell dry weight, and the shaded columns represent the lipopeptide production.

4 BioMed Research International

development of resistance by microbes, offering a promising alternative for the treatment of CDI.

3.4. Genome Sequencing of B. amyloliquefaciens C-1. The cir- cular chromosome of C-1 contains 3,934,216 bp, 46.5% GC content, 27 rRNA and 86 tRNA genes (Table 2,

Figure 4(a)). Genome annotation at the RAST server showed that the C-1 genome encodes 4013 proteins, and the corre- sponding functional categorization by COG annotation is in Figure 4(b). The sequence data of the B. amyloliquefaciens C-1 genome were deposited into NCBI and can be accessed via accession number SRP127533.

Table 1: The inhibition of the C-1 lipopeptide against Clostridium difficile in a plate test (the inhibition diameter showed in mm).

Clostridium difficile Concentration (μg/ml)

0 5 10 15

ATCC 9689 0.00 7:05 ± 0:71a,b 14:50 ± 0:71a,b 22:00 ± 1:41a,b

ATCC 700057 0.00 0.00 8:40 ± 0:71a 11:50 ± 0:71a

ATCC BAA-1870 0.00 0.00 8:50 ± 0:71a 10:50 ± 0:71a aSignificant difference of C-1 lipopeptide treatments vs. negative control (P < 0:01). bSignificant difference of lipopeptide treatments between C. difficile strains ATCC 9689 and ATCC 70057, ATCC 9689, and ATCC BAA-1870 (P < 0:01).

0 0.0 0.1 0.2 0.3 0.4 0.5 0.6

O D

60 0

0.7 0.8 0.9 1.0 1.1

2 4 6 8 10 12 14 16 Time (h)

18 20 22 24 26

0 0.25 MIC

0.5 MIC 0.75 MIC

(a)

0 0.0 0.1 0.2 0.3 0.4 0.5 0.6

O D

60 0

0.7 0.8 0.9 1.0 1.1

2 4 6 8 10 12 14 16 Time (h)

18 20 22 24 26

0 0.25 MIC

0.5 MIC 0.75 MIC

(b)

0 0.0 0.1 0.2 0.3 0.4 0.5 0.6

O D

60 0

0.7 0.8 0.9 1.0 1.1

2 4 6 8 10 12 14 16 Time (h)

18 20 22 24 26

0 0.25 MIC

0.5 MIC 0.75 MIC

(c)

Figure 2: The growth curve of C. difficile ATCC 9689 (a), ATCC 700057 (b), and BAA-1870 (c) treated with different concentrations of the C-1 lipopeptide.

5BioMed Research International

(a) (b)

(c) (d)

(e) (f)

Figure 3: Morphological changes of C. difficile ATCC 9689 treated with 0.5 MIC of the C-1 lipopeptide. Observed with a scanning electron microscope (×25000; (a) untreated cells and (b) treated cells). Observed with a light microscope (×400; (c) untreated cells and (e) treated cells). Observed with a fluorescent microscope (PI staining, ×400; (d) untreated cells and (f) treated cells).

Table 2: Genome project information summary of B. amyloliquefaciens C-1.

Property/attributes C-1 Property/attributes C-1

Finishing quality High-quality draft Total predicted CDS 3805

Sequencing platform PacBio Sequel rRNA operons 27

Total bases (Mb) 757.4 tRNA 86

NCBI taxonomy ID 1386 tmRNA 1

BioProject ID PRJNA427474 Noncoding RNA 81

Genome size (bp) 3934216 Miscellaneous RNA 81

GC content (%) 46.5

6 BioMed Research International

XJTU

BASys

2800 kbp

2400 kbp

2200 kbp 2000 kbp1800 kbp

1600 kbp

1400 kbp

3800 kbp 3600 kbp

3400 kbp

3200 kbp 800 kbp

1000 kbp

1200 kbp

3000 kbp

2600 kbp

200 kbp

400 kbp

600 kbp

Genes encoding proteins

Genes encoding functional RNA

COG functional categories

Cellularprocesses

Metabolism

Fonvard strand

Fonvard strand Reverse strand

Information storage and processing

Reverse strand

Translation, ribosomal structure and biogenesis

DNA replication, recombination and repair

Cell division and chromosome partitioning Posttranslational modification, protein turnover, chaperones Cell envelope biogenesis, outer membrane Cell motility and secretion Inorganic ion transport and metabolism

Energy production and conversion Carbohydrate transport and metabolism Amino acid transport and metabolism Nucleotide transport and metabolism Coenzyme metabolism Lipid metabolism Secondary metabolites biosynthesis, transport and catabolism

General function prediction only Function unknown

Signal transduction mechanisms

Transcription

Poorly characterized

Length: 3,934,217 bp; Genes; 4,223BASYS: Sunday March 05 01:31:32 2017

(a)

Figure 4: Continued.

7BioMed Research International

Blast searches of the 16S rRNA gene sequence of C-1 showed that it was most similar to other B. amyloliquefaciens isolates. B. amyloliquefaciens isolates appear to group into two clades indicated by phylogenetic tree analysis (Fig S3). Although C-1 is in the same clade with known strains such as DSM7, ATCC19217, and ATCC 14580, it appears to be phylogenetically distant from most other isolates.

3.5. Secondary Metabolites from B. amyloliquefaciens Strains. As much as 8.5% of the B. amyloliquefaciens C-1 genome CDS was assigned to categories related to the secondary metabolites responsible for the control of pathogens (“Motility and Chemotaxis” [85 CDS], “Membrane Trans- port” [71 CDS], “Virulence, Disease and Defense” [70 CDS], “Secondary Metabolism” [6 CDS], and “Stress Responses” [108 CDS]). For the functional categories of genes, a possible role in bacteria inhibition may be impor- tant. In the carbohydrate transport and metabolism cate-

gory, 437 genes (10.9% of total genes) were predicted in the C-1 genome. This finding suggests that C-1 possesses a broad battery of genes coding for enzymes required to release a variety of environmental carbon sources.

As a Bacillus. spp., B. amyloliquefaciens possesses an enormous potential to synthesize bioactive secondary metab- olites, especially nonribosomal-synthesized peptides and polyketides. For the nonribosomal peptide synthetases (NRPSs) and polyketide synthases (PKS), we used anti- SMASH to identify related giant gene clusters (Table 3, Table S1). C-1 was found to harbor genes encoding six nonribosomal peptides (surfactins (srfABCD), iturins (ituABCD), fengycins (fenABCDE), bacillibactin (bmyABC), teichuronic, and bacilysin) and three polyketides (bacillaene (baeEDLMNJRS), difficidin (difABCDEFGHIJ), and macrolactin (mlnABCDEFGHI)). Compared with other sequenced B. amyloliquefaciens strains, fengycin, difficidin, bacillibactin, bacilysin, macrolactin, and bacillanen showed

Environmental information processing(535)

Unclassified (417) Genetic information processing (509)

Carbohydrate metabolism (258) Amino acid metabolism (235) Cellular processes (214) Metabolism of cofactors and vitamins (159) Enzyme families (120) Energy metabolism (116)

Lipid metabolism (90) Nucleotide metabolism(85) Organismal systems (30) Xenobiotics biodegradation and metabolism(33) Biosynthesis of other secondary metabolites(41) Metabolism of other amino acids (45) Glycan biosynthesis and metabolism(66) Metabolism of terpenoids and polyketides (66) Human diseases (84)

(b)

Figure 4: The genome map (a) of B. amyloliquefaciens C-1 and overview of the subsystem category coverage of the C-1 genome based on RAST serve (b). The red circle is the CDS of the forward strand, and the blue circle is the CDS of the reverse strand. The outer circle represents the categorization of predicted protein coding sequences in the C-1 genome in COG annotation, and the inner circle represents the genes encoding function RNA.

8 BioMed Research International

100% identity, whereas surfactins had 82% identity (Table S1). Surfactin can inhibit awide range ofmicroorganismsdue to its ability to insert into the cell wall and create ion pores. Bacillomycin D, iturin, and fengycin have antifungal properties primarily based on their ability to disrupt the cell wall [30, 31]. Macrolactins, important 24-membered macrolactones produced by Bacillus spp., exhibit antimicrobial activities, where macrolactin A and E and succinyl macrolactin are the representative compounds. And it is assembled by a modular PKS system like macrolides, which could inhibit the H+-transporting ATPase of the bacterial cells [32]. Polyketide compounds inhibit a wide range of microorganisms by preventing protein synthesis [31]. The identity of surfactin among different strains varied from 82% to 96%, including subunit genes of SrfAA, SrfAB, SrfAC, and SrfAD (Fig S4). C-1 had all the subunit genes, SrfAA, SrfAB, SrfAC, and SrfAD, and the mutation of subunit genes in other strains may indicate loss of the ability to synthesize secondary metabolites [33].

PCR experiments detected C-1 surfactin, iturin, and fengycin genes, an intact Bac operon that included Bacilysin biosynthesis proteins BacA, BacB, BacC, BacD, and BacE (ORF3909-3913), and the oligopeptide permease operon (ORF1362-1366). Bac proteins, nonribosomally synthesized dipeptides active against a range of bacteria and some fungi, are involved in the biosynthesis of bacilysin. The proteolysis of this dipeptide releases the nonproteinogenic amino acid L-anticapsin, which functions as a competitive inhibitor of glucosamine synthase and can cause lysis of fungal cells [34]. Because there was 100% sequence identity of bacilysin,

whereas there was no antifungal activity of the C-1 lipopep- tide and exopolysaccharide, the regulation and expression of the encoded Bac operon, and modification of produced bacilysin, deserve more analysis.

Other antimicrobial gene clusters were predicted by anti- SMASH, such as the lantibiotic amylolysin [35], the bacteri- ocin amylocyclicin [36], and the aminoglycoside antibiotic butirosin [37]. These antibacterials have not been detected by chemical analysis from C-1 supernatants, possibly because of the fermentation medium or condition that had been used [38]. Potential gene clusters may explain the broad activity of C-1 against pathogens.

We also detected in the C-1 genome other beneficial genes, such as phospholipase (ORF767) and siderophore production genes. There were seven gene clusters responsible for siderophore production and iron acquisition, including an ABC-type Fe3+-siderophore transport system (ORF1-2, ORF3417-3418, 4007-4008), an Fe-bacillibactin uptake sys- tem (ORF413-415), an iron compound ABC uptake trans- porter (ORF624-627), and a siderophore biosynthesis protein (ORF1221-1222, 3262-3267). These proteins enable bacteria to sequester iron complexes produced by other path- ogens and antagonize certain pathogens [38, 39].

Moreover, we also detected genes for amphiphilic membrane-active biosurfactants and peptide antibiotics that have powerful antibacterial and mosquito larvicidal activity. The giant gene clusters add to the capacity of the C-1 bac- terium to contribute to the antimicrobial activity against C. difficile. And we also checked the biosafety of the C-1 strain by using Galleria mellonella and intestinal epithelial

Table 3: Identification of gene clusters potentially involved in the synthesis of secondary metabolites by B. amyloliquefaciens C-1.

Clustera Typeb Fromc Toc Secondary metabolited

1 Saccharide 165858 190731 Unknown

2 NRPS 556597 622004 Surfactin

3 Fatty acid 785309 810347 Unknown

4 NRPS 938299 967890 Iturins

5 Other KS 1158436 1199680 Butirosin

6 Fatty acid 1227747 1248724 Unknown

7 Terpene 1281720 1302460 Unknown

8 Fatty acid 1314519 1339344 Citrulline

9 Putative 1378928 1396630 Molybdenum cofactor

10 Lantipeptide 1406687 1451837 Unknown

11 Transatpks 1624433 1706630 Macrolactin

12 Transatpks-NRPS 1932737 2035409 Bacillaene

13 Transatpks-NRPS 2100037 2237835 Fengycin

14 Terpene 2263057 2284940 Unknown

15 T3PKS 2348257 2389357 Unknown

16 Transatpks 2504342 2604794 Difficidin

17 Bacteriocin-NRPS 3235204 3301995 Bacillibactin

18 Saccharide 3504667 3530078 Unknown

19 Saccharide 3624018 3678829 Teichuronic

20 Saccharide 3823278 3895655 Bacilysin aClusters identified using default settings of antiSMASH 5.0. bClass of gene cluster according to antiSMASH 5.0. cLocation of gene clusters in the B. amyloliquefaciens C-1 genome. dSecondary metabolites potentially produced based on the gene clusters.

9BioMed Research International

cells, which all indicated the safety of B. amyloliquefaciens C-1. A future study of how these gene clusters are expressed and regulated will help explain the synthesis of antimicrobial lipopeptides and augment our knowledge for the control of CDI.

4. Conclusions

B. amyloliquefaciens C-1 fermentation supernatant contains a mixture of lipopeptides, namely, surfactin and fengycin, which had a strong inhibitory effect on C. difficile growth and viability. Systematic research of the antibacterial mecha- nism showed that the C-1 lipopeptide damages the integrity and permeability barrier of the cell wall and cell membrane, then leads to C. difficile cell death. The ~3.93Mbp genome of C-1 reveals the genetic basis of its antimicrobial activity, and the antimicrobial compound-encoding gene clusters provide better understanding of the antibacterial mecha- nisms of this strain. Furthermore, the genome analysis will facilitate the production of effective probiotics that inhibit multidrug resistant pathogens in the host intestinal ecosys- tem, especially the phospholipase- and siderophore- producing clusters. Until now, the anti-C. difficile activities of the bacteriocins were known predominantly from in vitro studies; thus, the in vivo efficacies of the majority of these bacteriocins deserve further investigation.

Data Availability

The data used to support the findings of this study are available from the corresponding author upon request.

Conflicts of Interest

The authors declare no conflict of interest.

Authors’ Contributions

Jia Lv and Rong Da contributed equally to this work.

Acknowledgments

This work was financially supported by the National Natural Science Foundation of China (81673199) and the National Science Basic Research Plan in Shaanxi Province of China (2018JM7054).

Supplementary Materials

Fig S1: the isolation, purification, and verification of the B. amyloliquefaciens C-1 lipopeptide. (A) The acid precipitated crude lipopeptide; (B) the UV-VIS spectrophotometer scan- ning analysis of crude lipopeptide; (C) the purified lipopep- tide isolated by a TLC plate; (D) the PCR detection of lipopeptide synthesis-related genes fenB, srf, and ituD in the C-1 genome; phylogenetic tree of PCR fragments fenB (E), srf (F), and ituD (G). Fig S2: effect of the C-1 lipopeptide on AKPase in C. difficile ATCC 9689 (∗P < 0:05, ∗∗P < 0:01 indicated statistically significant differences of C-1 lipopep- tide treatments vs. negative control). Fig S3: neighbor-

joining phylogenetic tree based on 16S rRNA gene sequences of B. amyloliquefaciens strains. 16S rRNA gene sequences were from 16S ribosomal RNA gene partial sequence or directed from the genome annotation in NCBI with accession numbers of C-1 (JX028840), LL3 (CP002634.1), XH7 (CP002927.1), ATCC 13952 (CP009748.1), DSM7 (NC_ 014551), SRCM101267 (CP021505.1), Y2 (HE774679.1), UMAF6614 (NZ_CP006960), 19217 (CP009749.1), TA208 (CP002627.1), ATCC 14580 (CP000002.3), RD7-7 (CP016913.1), S499 (CP014700.1), LFB112 (NC_023073), CC178 (CP006845.1), L-H15 (CP010556.1), L-S60 (CP011278.1), B15 (KT923051.1), KHG19 (NZ_CP007242), ATCC 7050 (DQ297928.1), Y14 (NZ_CP017953), 168 (NC_ 000964), IT45 (NC_020272), ATCC 14581 (JQ579621.1), WS-8 (CP018200.1), LM2303 (MN640968.1), UMAF6639 (GCA_001593765), and DSM 319 (KM051080.1). Fig S4: phylogenetic tree of genome-sequenced B. amyloliquefaciens strains based on the amino acid sequences of surfactin synthe- tases SrfAA, SrfAB, SrfAC, and SrfAD, and the comparison of the gene loci strain C-1 (SRP127533), KHG19 (NZ_ CP007242), UMAF6639 (GCA_001593765), UMAF6614 (NZ_CP006960), LFB112 (NC_023073), IT45 (NC_ 020272), DSM7 (NC_014551), Y14 (NZ_CP017953), and 168 (NC_000964). Table S1: comparison of gene clusters potentially involved in the synthesis of secondary metabolites in B. amyloliquefaciens-sequenced strains. (Supplementary Materials)

References

[1] Z. Peng, L. Ling, C.W. Stratton et al., “Advances in the diagno- sis and treatment ofClostridium difficileinfections,” Emerging Microbes & Infections, vol. 7, no. 1, pp. 1–13, 2018.

[2] A. C. Clements, R. J. Magalhães, A. J. Tatem, D. L. Paterson, and T. V. Riley, “Clostridium difficile PCR ribotype 027: asses- sing the risks of further worldwide spread,” Lancet Infectious Diseases, vol. 10, no. 6, pp. 395–404, 2010.

[3] D. W. Hecht, M. A. Galang, S. P. Sambol, J. R. Osmolski, S. Johnson, and D. N. Gerding, “In vitro activities of 15 antimi- crobial agents against 110 toxigenic Clostridium difficile clini- cal isolates collected from 1983 to 2004,” Antimicrobial Agents and Chemotherapy, vol. 51, no. 8, pp. 2716–2719, 2007.

[4] C. F. Manthey, L. Eckmann, and V. Fuhrmann, “Therapy forClostridium difficileinfection – any news beyond metroni- dazole and vancomycin?,” Expert Review of Clinical Pharma- cology, vol. 10, no. 11, pp. 1239–1250, 2017.

[5] N. Roshan, K. A. Hammer, and T. V. Riley, “Non-conven- tional antimicrobial and alternative therapies for the treatment of Clostridium difficile infection,” Anaerobe, vol. 49, pp. 103– 111, 2018.

[6] S. Torsten, “Bacillus subtilis antibiotics: structures, syntheses and specific functions,” Molecular Microbiology, vol. 56, no. 4, pp. 845–857, 2005.

[7] I. Dimkić, S. Stanković, M. Nišavić et al., “The profile and antimicrobial activity of Bacillus lipopeptide extracts of five potential biocontrol strains,” Frontiers in Microbiology, vol. 8, 2017.

[8] B. C. S. Farias, D. C. Hissa, C. T. M. do Nascimento et al., “Cyclic lipopeptide signature as fingerprinting for the screen- ing of halotolerant Bacillus strains towards microbial

10 BioMed Research International

enhanced oil recovery,” Applied Microbiology and Biotechnol- ogy, vol. 102, no. 3, pp. 1179–1190, 2018.

[9] D. B. Abdallah, O. Frikha-Gargouri, and S. Tounsi, “Bacillus amyloliquefaciens strain 32a as a source of lipopeptides for bio- control of Agrobacterium tumefaciens strains,” Journal of Applied Microbiology, vol. 119, no. 1, pp. 196–207, 2015.

[10] S. Geeraerts, R. Ducatelle, F. Haesebrouck, and F. Van Immer- seel, “Bacillus amyloliquefaciens as prophylactic treatment for Clostridium difficile-associated disease in a mouse model,” Journal of Gastroenterology and Hepatology, vol. 30, no. 8, pp. 1275–1280, 2015.

[11] J. Lv, X. K. Xu, S. K. Hu, R. J. Zhang, and B. Han, “Isolation and identification of antibacterial lipopeptide produced by Bacillus amyloliquefaciens C-1,” Natural Product Research and Devel- opment, vol. 27, no. 2, pp. 199–204, 2015.

[12] W. J. Zhou, J. Yang, Y. Wan, R. J. Zhang, and B. Han, “Antimi- crobial activity on foodborne pathogens of extracellular fer- mented products from Bacillus amyloliquefaciens C-1,” Natural Product Research and Development, vol. 26, no. 1, pp. 123–127, 2014.

[13] L. Belbahri, A. Chenari Bouket, I. Rekik et al., “Comparative genomics of Bacillus amyloliquefaciens strains reveals a core genome with traits for habitat adaptation and a secondary metabolites rich accessory genome,” Frontiers in Microbiology, vol. 8, no. 8, 2017.

[14] X. Zhang, B. Li, Y. Wang et al., “Lipopeptides, a novel protein, and volatile compounds contribute to the antifungal activity of the biocontrol agent Bacillus atrophaeus CAB-1,” Applied Microbiology and Biotechnology, vol. 97, no. 21, pp. 9525– 9534, 2013.

[15] Y. Han, B. Zhang, Q. Shen et al., “Purification and identifica- tion of two antifungal cyclic peptides produced by Bacillus amyloliquefaciens L-H15,” Applied Biochemistry and Biotech- nology, vol. 176, no. 8, pp. 2202–2212, 2015.

[16] I. Wiegand, K. Hilpert, and R. E. Hancock, “Agar and broth dilution methods to determine the minimal inhibitory concen- tration (MIC) of antimicrobial substances,” Nature Protocols, vol. 3, no. 2, pp. 163–175, 2008.

[17] J. S. Eom and H. S. Choi, “Inhibition of Bacillus cereus growth and toxin production by Bacillus amyloliquefaciens RD7-7 in fermented soybean products,” Journal of Microbiology and Biotechnology, vol. 26, no. 1, pp. 44–55, 2016.

[18] D. R. Snydman, N. V. Jacobus, and L. A. McDermott, “Activity of a novel cyclic lipopeptide, CB-183,315, against resistant Clostridium difficile and other Gram-positive aerobic and anaerobic intestinal pathogens,” Antimicrobial Agents and Chemotherapy, vol. 56, no. 6, pp. 3448–3452, 2012.

[19] H. M. Xu, Y. J. Rong, M. X. Zhao, B. Song, and Z. M. Chi, “Antibacterial activity of the lipopeptides produced by Bacillus amyloliquefaciens M1 against multidrug-resistant Vibrio spp. isolated from diseased marine animals,” Applied Microbiology and Biotechnology, vol. 98, no. 1, pp. 127–136, 2014.

[20] C. Ma, N. He, Y. Zhao, D. Xia, J. Wei, and W. Kang, “Antimi- crobial mechanism of hydroquinone,” Applied Biochemistry and Biotechnology, vol. 189, no. 4, pp. 1291–1303, 2019.

[21] B. Zhang, C. Dong, Q. Shang, Y. Han, and P. Li, “New insights into membrane-active action in plasma membrane of fungal hyphae by the lipopeptide antibiotic bacillomycin L,” Bio- chimica et Biophysica Acta, vol. 1828, no. 9, pp. 2230–2237, 2013.

[22] R. D. Finn, J. Clements, and S. R. Eddy, “HMMER web server: interactive sequence similarity searching,” Nucleic Acids Research, vol. 39, suppl, pp. W29–W37, 2011.

[23] C. Camacho, G. Coulouris, V. Avagyan et al., “BLAST+: archi- tecture and applications,” BMC Bioinformatics, vol. 10, no. 1, p. 421, 2009.

[24] K. Blin, M. H. Medema, R. Kottmann, S. Y. Lee, and T. Weber, “The antiSMASH database, a comprehensive database of microbial secondary metabolite biosynthetic gene clusters,” Nucleic Acids Research, vol. 45, no. D1, pp. D555–D559, 2017.

[25] J. Cheetham, F. Dehne, S. Pitre, A. Rau-Chaplin, and P. J. Tail- lon, “Parallel CLUSTALW for PC Clusters,” in Computational Science and Its Applications — ICCSA 2003, pp. 300–309, Springer, Berlin, Heidelberg, 2003.

[26] K. Tamura, D. Peterson, N. Peterson, G. Stecher, M. Nei, and S. Kumar, “MEGA5: molecular evolutionary genetics analysis using maximum likelihood, evolutionary distance, and maxi- mum parsimony methods,” Molecular Biology and Evolution, vol. 28, no. 10, pp. 2731–2739, 2011.

[27] P. Jin, H. Wang, W. Liu, and W. Miao, “Characterization of lpaH2 gene corresponding to lipopeptide synthesis in Bacillus amyloliquefaciens HAB-2,” BMC Microbiology, vol. 17, no. 1, p. 227, 2017.

[28] X. Li, C. He, L. Song et al., “Antimicrobial activity and mecha- nism of Larch bark procyanidins against Staphylococcus aureus,” Acta Biochimica et Biophysica Sinica, vol. 49, no. 12, pp. 1058–1066, 2017.

[29] A. Etchegaray, C. de Castro Bueno, I. S. de Melo et al., “Effect of a highly concentrated lipopeptide extract of Bacillus subtilis on fungal and bacterial cells,” Archives of Microbiology, vol. 190, no. 6, pp. 611–622, 2008.

[30] B. J. Schofield, A. Skarshewski, N. Lachner et al., “Near com- plete genome sequence of the animal feed probiotic, Bacillus amyloliquefaciens H57,” Standards in Genomic Sciences, vol. 11, no. 1, p. 60, 2016.

[31] D. H. Kim, H. K. Kim, K. M. Kim et al., “Antibacterial activities of macrolactin A and 7-O-succinyl macrolactin A from Bacil- lus polyfermenticus KJS-2 against vancomycin-resistant enterococci and methicillin-resistant Staphylococcus aureus,” Archives of Pharmacal Research, vol. 34, no. 1, pp. 147–152, 2011.

[32] X. Yan, Y.-X. Zhou, X.-X. Tang et al., “Macrolactins from marine-derived Bacillus subtilis B5 bacteria as inhibitors of inducible nitric oxide and cytokines expression,” Marine Drugs, vol. 14, no. 11, p. 195, 2016.

[33] B. Fan, J. Blom, H. P. Klenk, and R. Borriss, “Bacillus amy- loliquefaciens, Bacillus velezensis, and Bacillus siamensis form an “Operational Group B. amyloliquefaciens” within the B. subtilis species complex,” Frontiers in Microbiology, vol. 8, 2017.

[34] I. Martínez-Raudales, Y. de la Cruz-Rodríguez, A. Alvarado- Gutiérrez et al., “Draft genome sequence of Bacillus velezensis 2A-2B strain: a rhizospheric inhabitant of Sporobolus airoides (Torr.) Torr., with antifungal activity against root rot causing phytopathogens,” Standards in Genomic Sciences, vol. 12, no. 1, 2017.

[35] A. A. Arias, M. Ongena, B. Devreese, M. Terrak, B. Joris, and P. Fickers, “Characterization of amylolysin, a novel lantibiotic from Bacillus amyloliquefaciensGA1,” PLoS One, vol. 8, no. 12, 2013.

11BioMed Research International

[36] R. Scholz, J. Vater, A. Budiharjo et al., “Amylocyclicin, a novel circular bacteriocin produced by Bacillus amyloliquefaciens FZB42,” Journal of Bacteriology, vol. 196, no. 10, pp. 1842– 11852, 2014.

[37] N. M. Llewellyn, Y. Li, and J. B. Spencer, “Biosynthesis of butirosin: transfer and deprotection of the unique amino acid side chain,” Chemistry & Biology, vol. 14, no. 4, pp. 379–386, 2007.

[38] S. B. Khedher, H. Boukedi, O. Kilani-Feki et al., “Bacillus amyloliquefaciens AG1 biosurfactant: Putative receptor diversity and histopathological effects on Tuta absoluta midgut,” Journal of Invertebrate Pathology, vol. 132, pp. 42–47, 2015.

[39] D. Wang, Y. Zhan, D. Cai, X. Li, Q. Wang, and S. Chen, “Reg- ulation of the synthesis and secretion of the iron chelator cyclodipeptide pulcherriminic acid in Bacillus licheniformis,” Applied and Environmental Microbiology, vol. 84, no. 13, 2018.

12 BioMed Research International

Copyright of BioMed Research International is the property of Hindawi Limited and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use.