Journal of Chemical Ecology

, Volume 38, Issue 12, pp 1521–1527

Production and Antimicrobial Activity of 3-Hydroxypropionaldehyde from Bacillus subtilis Strain CU12


  • C. Wise
    • Department of ChemistryCarleton University
  • L. Novitsky
    • Department of ChemistryCarleton University
  • A. Tsopmo
    • Department of ChemistryCarleton University
    • Department of ChemistryCarleton University

DOI: 10.1007/s10886-012-0219-2

Cite this article as:
Wise, C., Novitsky, L., Tsopmo, A. et al. J Chem Ecol (2012) 38: 1521. doi:10.1007/s10886-012-0219-2


Bacillus subtilis strains are known to produce a vast array of antimicrobial compounds. However, some compounds remain to be identified. Disk assays performed in vitro with Bacillus subtilis CU12 showed a significant reduction in mycelial growth of Alternaria solani, Botrytis cinerea, Fusarium sambucinum, and Pythium sulcatum. Crude B. subtilis culture filtrates were subsequently extracted with ethyl acetate and butanol. A bioassay guided purification procedure revealed the presence of one major antifungal compound in the butanol extract. Purification of the compound was performed using a reverse-phase C18 solid phase extraction (SPE) cartridge and flash column chromatography. NMR data showed that the main antimicrobial compound was a cyclic dimer of 3-hydroxypropionaldehyde (HPA). This study demonstrated the antimicrobial activity of B. subtilis strain CU12 against phytopathogenic microorganisms is mediated at least in part by the production of HPA. It also suggests that this B. subtilis strain could be effective at controlling pathogens through protection of its ecological niche by antibiosis.


Bacillus subtilis3-hydroxypropionaldehydeAntifungal compoundPlant pathogensEcological niche


The plant phyllosphere and rhizosphere are extremely complex microbial ecosystems containing saprophytes, epiphytes, endophytes, pathogens, and other microorganisms. Natural ecosystems tend to contain microbial communities living in relative harmony where all populations balance out in their quest for food and an ecological niche (Bélanger and Avis, 2002). In artificial systems such as agriculture, modifications to the natural balance can drastically alter the microbial community and lead to ingress or increase in plant pathogens that may have devastating effects on plant health and productivity (Avis et al., 2008). Integration of beneficial microorganisms into agricultural or other disturbed ecosystems has the potential to suppress some of the problems associated with plant pathogens. Among these beneficial microorganisms, some of the most prominent plant pathogen antagonists being studied are members of the Bacillus genus.

Bacillus spp. are gram positive endospore-forming rhizobacteria that are widely studied because they are microbial factories for biologically active compounds. The most commonly studied species of this genus, B. subtilis (Ehrenberg) Cohn has been found to have approximately 4–5 % of its entire genome devoted solely to antibiotic synthesis leading to a vast array of structurally diverse compounds (Stein, 2005). Bacillus subtilis isolates produce peptide (Ongena and Jacques, 2008) and/or non-peptide (Hamdache et al., 2011) antimicrobial compounds depending on the strain. However, all possible antimicrobial compounds from Bacillus spp. have yet to be elucidated.

Recently, a strain of B. subtilis (named strain CU12) was isolated in our laboratory from an agricultural field with a history of low plant disease. However, no work on this strain’s ability to produce biologically active molecules had yet been performed. The objectives of this study were (i) to investigate the antifungal and anti-oomycete ability of B. subtilis strain CU12, and (ii) to identify the compound(s) responsible for its antifungal activity.

Methods and Materials

Microbial Material

Bacillus subtilis strain CU12 was isolated from a disease suppressive soil and maintained on tryptic soy agar (TSA, Becton Dickinson, Sparks, MD, USA) at 23 °C prior to usage. The bacterium was identified as B. subtilis by using nucleotide sequence data from the small subunit (16S) of ribosomal RNA (GenBank Accession number JX489167).

The pathogenic fungi Alternaria solani Sorauer, Botrytis cinerea Pers., and Fusarium sambucinum Fuckel, and the pathogenic oomycete Pythium sulcatum R.G. Pratt & J.E. Mitch were maintained on potato dextrose agar (PDA, Becton Dickinson) at 23 °C. Alternaria solani, B. cinerea, and F. sambucinum were obtained from the “Laboratoire de diagnostic en phytoprotection” (MAPAQ, Québec, Québec, CA). Pythium sulcatum was isolated from an infected carrot root and is available at the Canadian Collection of Fungal Cultures (Agriculture and Agri-Food Canada, Ottawa, Ontario, CA).

Effect of B. subtilis CU12 on Mycelial Growth of Pathogens

Bacillus subtilis CU12 was tested for its ability to inhibit the mycelial growth of the pathogens in vitro. Petri dishes (100 × 15 mm) containing PDA were inoculated with four 1-cm streaks of B. subtilis CU12 at the four cardinal points. In B. cinerea and P. sulcatum trials, the streaked bacteria were located 3.5 cm from the center of the dish. In A. solani and F. sambucinum trials, the streaked bacteria were located 2 cm from the center. A mycelial plug (5-mm diam) of each pathogen was deposited individually in the center of the dish. The control treatments consisted of the pathogen inoculations without streaking of the bacterium. The dishes were incubated at 23 °C. When all control mycelia grew to 3.5 cm (B. cinerea and P. sulcatum) or 2 cm (A. solani and F. sambucinum) from the center of the dish, the mycelial growth of all treatments were measured. Mycelial growth was assessed as the average of two perpendicular diameters of the thallus. The experiments were performed according to a completely randomized design with four repetitions. Experiments were repeated twice.

Extraction of Antifungal Compounds

A total of 600 ml of tryptic soy broth (TSB, Becton Dickinson) was inoculated with B. subtilis CU12 and incubated at 28.5 °C under shaking conditions (150 rpm) for 48 hr. Following incubation, the culture broth containing B. subtilis CU12 was centrifuged at 10,000 rpm for 20 min. The supernatant was recovered, and the bacterial pellet was discarded. The supernatant was extracted twice with 300 ml ethyl acetate, and the fractions were combined. The aqueous phase was then re-extracted with n-butanol (2 × 300 ml).

The ethyl acetate and butanol fractions were dried using a Heidolph 4000 rotary evaporator (Schwabach, Germany), re-suspended in ethyl acetate or butanol, and transferred to screw-cap tubes. Extracts were further concentrated under a stream of nitrogen and re-suspended in a small volume of the appropriate solvent to yield a concentration 200 times (ethyl acetate fraction) or 100 times (butanol fraction) greater than the initial extracts.

Effect of Extracts on Mycelial Growth of Pathogens

Concentrated extracts were assayed against the four pathogens using a disk diffusion assay. Sterile paper disks (6-mm diam) were inoculated with 25 μl of either the ethyl acetate or the butanol extract and air dried. Twenty-five microliter of pure ethyl acetate or butanol served as the controls. The centers of the PDA dishes were individually inoculated with a mycelial plug (5-mm diam) of each pathogen. One disk for each treatment was placed individually in the PDA plate. In B. cinerea and P. sulcatum trials, the disks were located 3.5 cm from the center of the dish. In A. solani and F. sambucinum trials, the disks were located 2 cm from the center. Dishes were incubated at 23 °C. When the control mycelia grew to the disks, mycelial growth of all treatments was measured. Measurements were noted as the radius of the mycelial growth for each pathogen. The experiment was performed according to a completely randomized design with three repetitions.

The extracts were assayed additionally by using a direct bioassay on a TLC plate. Four microliter of the concentrated ethyl acetate and butanol extracts were spotted on a silica-backed TLC plate with fluorescent marker (Sigma, Mississauga, Ontario, CA) and left to dry. Pure ethyl acetate and butanol served as controls. To perform the bioassay, 10 ml of ultrapure (Milli-Q) water were mixed with 0.4 g of PDA and heated to the boiling point. A concentrated conidial suspension of A. solani was prepared by adding 10 ml ultrapure water to a Petri dish covered with actively growing mycelium of the fungus. The surface of the mycelium was gently scraped to dislodge the conidia. Once the PDA had cooled (50 °C), the conidial suspension was added and mixed thoroughly. The resulting suspension was thinly spread over the TLC. The inoculated TLC was placed in a moist chamber (>95 % RH) and incubated at 23 °C for 48 hr. Following incubation, active fractions were revealed as white inhibition zones on the dark background of the fungus.

Detection and Fractionation of Antifungal Compounds on SPE Cartridge

Extracts presenting antifungal activity were separated using a sterile 1 g (5 ml) Bond Elut C18 SPE cartridge (Varian, Palo Alto, CA, USA). The cartridge was conditioned with two volumes (10 ml) of ultrapure water prior to the application of 60 μl of the concentrated extract. Compounds were sequentially eluted with 1 volume (5 ml) of increasing concentrations of methanol:water as follows: 0:100, 20:80, 50:50, 80:20, 100:0 (v/v). The five collected fractions and the crude extract were separated on TLC using chloroform:methanol 90:10 (v/v). Following migration, the TLC was placed under short (254 nm) and long (365 nm) ultraviolet (UV) wavelengths and the visible compounds were noted. All TLC were bioassayed using the direct bioassay with A. solani as described previously.

Purification of Antifungal Compounds

Purification of active compounds from the SPE fraction eluted with methanol:water (80:20) was achieved using flash chromatography with C18 silica gel (Sigma Aldrich, Oakville, Ontario, CA). The silica gel was soaked in methanol:water (1:1), loaded into the column, and equilibrated with 5 column volumes (500 ml) of ultrapure water prior to use. Two milliliters of the concentrated extract then were placed on top of the gel. Compounds were sequentially eluted with 1 volume (100 ml) of increasing concentrations of methanol:water as follows: 0:100, 20:80, 50:50, 80:20, 100:0 (v/v). Fractions (15 ml each) were collected and directly bioassayed on TLC as described above, and those containing active compounds were combined for spectroscopic analysis.

NMR Analysis

1H and 13C-NMR experiments were recorded using a Bruker Avance 300 and a Bruker Avance 400 spectrometer (Billerica, MA, USA). Samples were dissolved in methanol-d4 or chloroform-d. Experiments were performed at room temperature, and the spectra were calibrated based on the residual non-deuterated solvents: δH 3.31 and δC 49.1 for methanol, δH 7.26 and δC 77.0 for chloroform.

Statistical Analysis

In the B. subtilis CU12 and crude extract inhibition trials, ANOVA and/or t-tests were performed with SAS software (SAS Institute, Cary, NC, USA) and reported as significant when P < 0.05.


Effect of B. subtilis CU12 on Mycelial Growth of Pathogens

Antifungal activity of B. subtilis CU12 was assayed in vitro as the inhibition of mycelial growth through the production of diffusible compounds in solid media. ANOVA indicated that there was no significant difference between repetitions of this experiment. Therefore, the data were combined and analyzed as a single experiment. It was revealed that B. subtilis CU12 exhibited significant activity against all tested pathogens, although the degree of inhibition varied considerably (Fig. 1) among mycelial pathogens. Analysis of the mycelial growth demonstrated that B. subtilis CU12 significantly inhibited growth of A. solani, B. cinerea, F. sambucinum, and P. sulcatum by 53, 25, 39, and 20 %, respectively (Fig. 1).
Fig. 1

In vitro effect of Bacillus subtilis strain CU12 on mycelial growth of Alternaria solani, Botrytis cinerea, Fusarium sambucinum, and Pythium sulcatum. Growth was measured as the diameter of the thallus. Error bars are standard errors. For each pathogen, an asterisks (*) above a bacterial treatment denotes significant difference from the control according to a t-test (P < 0.05; N = 8)

Effect of Extracts on Pathogens

The concentrated ethyl acetate and butanol extracts of B. subtilis CU12 liquid cultures were subjected to a disk diffusion assay to determine their effects on mycelial growth of the four pathogens. The ethyl acetate fraction did not significantly affect the mycelial growth of any of the tested pathogens when compared to the control (data not shown). Conversely, the concentrated butanol extract significantly inhibited the mycelial growth of A. solani, B. cinerea, and P. sulcatum by 52, 47, and 34 %, respectively (Fig. 2). Fusarium sambucinum mycelial growth was not significantly inhibited by the butanol extract (P > 0.05).
Fig. 2

Effect of butanol (BU) extract from Bacillus subtilis strain CU12 culture filtrates on mycelial growth of Alternaria solani, Botrytis cinerea, Fusarium sambucinum, and Pythium sulcatum. Growth was measured as the radius of the thallus. Error bars are standard errors. For each pathogen, an asterisks (*) above a bacterial treatment denotes significant different from the control according to a t-test (P < 0.05; N = 3)

The B. subtilis CU12 ethyl acetate and butanol fractions also were evaluated by using a direct bioassay on TLC with A. solani. Results showed that both controls (ethyl acetate and butanol solvents) and the B. subtilis CU12 ethyl acetate extract were not inhibitory to A. solani. However, the B. subtilis CU12 butanol extract showed a lack of A. solani growth as revealed by a strong inhibition zone (Fig. 3).
Fig. 3

Inhibition of Alternaria solani growth using a direct TLC bioassay. EAC = ethyl acetate control; EAX = Bacillus subtilis CU12 ethyl acetate extract; BC = butanol control; BX = Bacillussubtilis CU12 butanol extract

Purification of Antifungal Compounds

Many compounds were revealed by analytical TLC under UV light for the crude butanol extract and for fractions from reverse-phase chromatography. When the TLC containing the crude butanol extract and the five SPE-recovered fractions were bioassayed, a single spot, at a retention factor (Rf) of approximately 0.74, was antimicrobial in the crude extract, and from the SPE fraction eluted with 80:20 (v/v) methanol:water (Fig. 4). Further purification on flash column chromatography afforded the antimicrobial compound, which was not visible under UV light.
Fig. 4

TLC separation and direct bioassay of Bacillus subtilis CU12 crude butanol extract and solid phase extraction cartridge fractions. Direct bioassay with Alternaria solani on the TLC: 1 = crude butanol extract; 2 = 0:100 v/v methanol:water fraction; 3 = 20:80 v/v methanol:water fraction; 4 = 50:50 v/v methanol:water fraction; 5 = 80:20 v/v methanol:water fraction; 6 = 100:0 v/v methanol:water fraction

NMR Analysis

Following pooling of pure active fractions from C18 flash chromatography, the purified antimicrobial compound was identified as 3-hydroxypropionaldehyde (HPA) cyclic dimer based on analysis of its 1- and 2-dimensional NMR spectra and comparison with published data (Sung et al., 2003). Proton (1H) and carbon (13C) data are shown in Fig. 5. The 13C NMR spectrum showed six main signals between δc 34 and 108 ppm. Each of the signals appeared as doublets, and therefore, showed that the compound was present in isomeric forms. The 1H NMR spectrum of HPA showed aliphatic proton signals δH 1.20–1.60 ppm as well as signals of protons linked to oxygenated carbons between δH 3.50–4.40. Signals at δH 4.90 were characteristic of anomeric-type sugar protons. There was no evidence of the presence of HPA monomer.
Fig. 5

Carbon-13 (a) and proton (b) NMR spectra of antimicrobial compound in deuterated methanol


Bacillus subtilis strain CU12 was isolated from a disease suppressive soil and potentially could be used as a beneficial bacterial antagonist toward plant disease pathogens. In order to determine its antimicrobial potential, the bacterium was tested against various plant pathogens followed by purification of a chemical responsible for or contributing to the antimicrobial activity. Results showed that B. subtilis CU12 provided significant inhibition of mycelial growth of the tested plant pathogens resulting in an inhibition zone. Moreover, butanol extract of culture filtrates of the bacterium equally showed inhibition in pathogen mycelial growth indicating the production of one or more extracellular antifungal compounds.

Following purification, an antifungal compound was successfully isolated from the n-butanol extract and identified as HPA. NMR (1H and 13C) spectra of the compound were similar to those published by Sung et al. (2003). HPA previously has been produced by the genera Klebsiella, Citrobacter, Enterobacter, Clostridium, and Lactobacillus (Vollenweider et al., 2003). There is one older report of HPA produced as an acrolein intermediate by Bacillus amaracrylus (Voisenet, 1914). The bacterium has since been reclassified as Bacillus polymyxa (Holt, 1986) and now as Paenibacillus polymyxa (Garrity et al., 2004). To our knowledge, this is the first report of a strain of B. subtilis producing HPA.

In aqueous solution, HPA from Lactobacillus reuteri is found as a multi-compound mixture (named reuterin) of monomeric, cyclic dimer, and hydrated forms of HPA (Talarico and Dobrogosz, 1989; Sung et al., 2002). The nature and proportion of HPA isomers in solution vary depending on the pH. NMR spectra obtained at pH 4.0 showed that β-hydroxypropionaldehyde was the main compound, while at pH 10.5 the mixture was more complex and contained only traces of β-hydroxypropionaldehyde with more oligomers (Sung et al., 2003). In this study, NMR data were obtained in both chloroform-d (data not shown) and methanol-d4 without pH control. NMR spectra obtained are similar to those published by Sung et al. (2003) who reported at pH 10.5 the presence of monomer, cyclic dimer, aldol dimer, hemiacetal and acetal forms of HPA. In a study done by Vollenweider et al. (2003), it was observed that the composition of the HPA system is concentration dependent. At concentration levels of 4.9 M, the predominant form present was the HPA dimer, while lower concentrations (<1.4 M) showed a significant decrease in dimer concentration (Vollenweider et al., 2003). The main form appeared to be the cyclic dimer as indicated by protons at δH 4.90 that were directly linked to oxygenated carbons at δC 106–107 based on correlations from the Heteronuclear Multiple Bond Correlation (HMBC) spectrum. The lack of the 3-HPA monomeric form and the presence of the cyclic dimer in the NMR spectra may indicate an elevated concentration of HPA.

At high concentrations, HPA is active against various microorganisms including some genera that produce them (Vollenweider and Lacroix, 2004). The most intensively studied of the HPA producing bacteria, L. reuteri has shown some tolerance to HPA, however, its viability does decrease at very high HPA concentrations (Vollenweider and Lacroix, 2004). In an antimicrobial activity study done by Chen et al. (2002), they determined that an HPA concentration of 20–35 μg/ml can effectively prevent the growth of B. subtilis ATCC 6633 and a 30–50 μg/ml concentration is capable of causing death in this test organism. In our work, B. subtilis strain CU12 was able to grow in concentrations up to 1,360 μg/ml HPA indicating its high tolerance to its own antimicrobial compound.

HPA and the HPA-based multi-compound reuterin system produced by various microorganisms have been described as broad spectrum antimicrobial agents with antibacterial, antiparasite, antiviral, anticancer, and antifungal properties (Dobrogosz and Lindgren, 1988). Broad spectrum activity may be related to the mode of action in which HPA/reuterin acts although this has been difficult to study due to the complex nature of the HPA/reuterin system. Two mechanisms nonetheless have been proposed. The first relates to the reaction of the aldehyde group with thiol and other primary amines, which inactivates proteins and small molecules that contain these groups (Vollenweider and Lacroix, 2004) leading to growth inhibition (Schaefer et al., 2010). A second mechanism suggests that the structural similarities between the HPA-dimer and ribose cause it to act as a competitive inhibitor blocking the ribonucleotide reductase enzyme (Vollenweider et al., 2010). The ribonucleotide reductase enzyme is required for the generation of deoxynucleotides that are required for DNA synthesis.

From an ecological point of view, B. subtilis strain CU12 would gain an advantage through the production of HPA. Indeed, HPA would provide the antagonist with a means to protect its ecological niche against various microbial competitors including plant pathogens. From a practical point of view, the introduction of B. subtilis into agricultural or other artificial systems may protect the microecosystem against unwanted microorganisms such as plant pathogens and could provide a natural means to protect food crops and other plants.

Overall, the identification of the antifungal HPA may explain at least in part, the effectiveness of B. subtilis CU12 in inhibiting plant pathogens. Moreover, screening of other B. subtilis strains for increased production of HPA could help in selection of strains with superior inhibitory properties. Finally, the discovery of HPA will allow further work in elucidating the molecular, biochemical, and ecological phenomena behind the antagonism by B. subtilis strain CU12 and help with the prediction of its inhibitory efficacy.


The authors thank Yichen Du and Justin Falardeau for technical assistance. This work was supported by a research grant from the Natural Sciences and Engineering Research Council (NSERC) of Canada.

Copyright information

© Springer Science+Business Media New York 2012