Co-culturing a novel Bacillus strain with Clostridium tyrobutyricum ATCC 25755 to produce butyric acid from sucrose
© Dwidar et al.; licensee BioMed Central Ltd. 2013
Received: 1 January 2013
Accepted: 21 February 2013
Published: 4 March 2013
Currently, the most promising microorganism used for the bio-production of butyric acid is Clostridium tyrobutyricum ATCC 25755T; however, it is unable to use sucrose as a sole carbon source. Consequently, a newly isolated strain, Bacillus sp. SGP1, that was found to produce a levansucrase enzyme, which hydrolyzes sucrose into fructose and glucose, was used in a co-culture with this strain, permitting C. tyrobutyricum ATCC 25755T to ferment sucrose to butyric acid.
B. sp. SGP1 alone did not show any butyric acid production and the main metabolite produced was lactic acid. This allowed C. tyrobutyricum ATCC 25755T to utilize the monosaccharides resulting from the activity of levansucrase together with the lactic acid produced by B. sp. SGP1 to generate butyric acid, which was the main fermentative product within the co-culture. Furthermore, the final acetic acid concentration in the co-culture was significantly lower when compared with pure C. tyrobutyricum ATCC 25755T cultures grown on glucose. In fed-batch fermentations, the optimum conditions for the production of butyric acid were around pH 5.50 and a temperature of 37°C. Under these conditions, the final butyrate concentration was 34.2±1.8 g/L with yields of 0.35±0.03 g butyrate/g sucrose and maximum productivity of 0.3±0.04 g/L/h.
Using this co-culture, sucrose can be utilized as a carbon source for butyric acid production at a relatively high yield. In addition, this co-culture offers also the benefit of a greater selectivity, with butyric acid constituting 92.8% of the acids when the fermentation was terminated.
KeywordsButyric acid Sucrose Levansucrase Co-culture Clostridium tyrobutyricum ATCC 25755 Bacillus sp. SGP1
Reinforced clostridial medium
High-performance liquid chromatography
Flame ionization detector
Refractive index detector
Gas chromatography–mass spectrometry
Electron impact ionization
Butyric acid has several potential applications in chemical and pharmaceutical industries and can be useful for fuel production . It is also widely used in foodstuffs and beverage industries . Nowadays, butyric acid is produced mainly for industry from crude oil through petrochemical methods . However, butyric acid obtained from microbial fermentation is more promising due to the growing needs for bio-based natural products especially for food additives, and pharmaceuticals .
Currently, the most promising microorganism used for the bio-production of butyric acid is Clostridium tyrobutyricum ATCC 25755T[4–7]. This strain is capable of producing butyric acid with high selectivity and can tolerate high concentrations. However, it can only ferment monosaccharides like glucose, xylose and fructose, and is currently unable to utilize disaccharides, such as sucrose and lactose [8–10]. Despite the presence of several butyric acid bacteria which can utilize sucrose as a carbon source, such as Clostridium butyricum[11, 12] and some strains of C. tyrobutyricum, such as C. tyrobutyricum ZJU8235 , C. tyrobutyricum ATCC 25755T offers the highest butyric acid yields and final concentrations and so is the preferred strain.
Since the cost of culture media components has a great influence on the overall cost of butyric acid produced by fermentation [2, 10], we wanted to utilize cheaper substrates for butyric acid production. One such substrate, sucrose, is an important and relatively cheap carbon source and is readily available in nature. It forms more than 90% of the total carbohydrates in sugar canes  and sugar beet in which sugars can account for 12% to 20% of the plant's dry weight. Sucrose also constitutes large percentage of the total sugar content in molasses. Consequently, our lab seeks to use sucrose as a carbon source for butyric acid production.
In previous studies, several research groups evaluated the use of co-culturing to address limitation in substrate utilization by individual strains for the eventual production of different fermentation products. For example, Zhang et al. (2009)  discussed the possibility of using Clostridium thermocellum in a co-culture with Clostridium thermobutyricum for butyric acid production from cellulose, while Chang et al. (2008)  reported that the combination of aerobic Bacillus and anaerobic Clostridium may play the key role in the future of biofuel production from biomass. Their rationale was that strains of Bacillus generally have a high growth rate and secrete many saccharification-related extracellular enzymes into the medium, such as amylase, pectinase, protease, cellulose, and hemicellulase. Trana et al. (2010)  have recently shown the potential of co-culturing Bacillus subtilis with Clostridium butylicum for butanol production from starch. However, the Bacillus species used in the above studies are aerobic and, as such, they cannot be stably used in co-cultures for long periods with Clostridia, which are anaerobic. In contrast, Bacillus strains which are facultative anaerobes such as Bacillus licheniformis, will be more appropriate for long term co-culturing with Clostridium strains under anaerobic conditions.
In this study, we characterized the levansucrase activity of a newly isolated strain of Bacillus which was found to be closely related to B.licheniformis ATCC 14580T and Bacillus sonorensis NRRL B-23154T based upon 16S rRNA gene sequencing. To demonstrate its applicability, we next performed co-culture fermentations with this strain and C. tyrobutyricum ATCC 25755T for butyric acid production using sucrose as a carbon source.
Results and discussion
B. sp. SGP1 Hydrolyzes sucrose through levansucrase
B. sp. SGP1 was isolated from a waste water sludge sample obtained from anaerobic digestion tank. Identifying this strain through 16S rRNA gene sequencing revealed that its closest relative was Bacillus licheniformis ATCC 14580T (99.4%, 16S rRNA gene similarity) followed by Bacillus sonorensis NRRL B-23154T (99.3%) and Bacillus aerius 24KT (99.3%). The phylogenetic analysis (Additional file 1: Figure S1) showed that the strain SGP1 formed a monophyletic group with B. licheniformis ATCC 14580T, B. sonorensis NRRL B-23154T and B. aerius 24KT.
Many groups reported previously that several Bacillus strains, and specifically B. licheniformis, are able to produce levansucrase. This enzyme can both hydrolyze sucrose into its constituent glucose and fructose subunits and form the levan polymer from the fructose subunits [17, 18]. Previous studies demonstrated that levansucrase also degrades levan to individual fructose monomers especially when the reducing sugar concentration is low [18–20]. Consequently, we tested for the production of levansucrase by B. sp. SGP1. For colonies grown on agar plates containing sucrose, the production of a polysaccharide was obvious as they showed a slimy mucoid growth. In contrast, when grown on agar plates containing glucose, none of the colonies were mucoid. This polysaccharide was consequently precipitated from liquid cultures of B. sp. SGP1 grown on sucrose and hydrolyzed to its constituent monosaccharides. Analysis of these sugars by HPLC (High-performance liquid chromatography) found fructose monomers, implying that the polysaccharide was levan and that the extracellular enzyme is levansucrase.
Co-culturing B. sp. SGP1 with C. tyrobutyricum ATCC 25755T
Interestingly, it was found that the amount of butyric acid produced by the co-culture in serum bottles using sucrose is actually higher than that produced from C. tyrobutyricum ATCC 25755T when grown alone on the same amount of glucose (Figure 4A). In contrast, the final acetate concentration in the co-culture was significantly lower than that seen from C. tyrobutyricum ATCC 25755T cultured solely on glucose (Figure 4B). These two findings are likely related as the final total acid concentrations within the cultures were similar, 14.6 g/L for the co-culture and 14.8 g/L for C. tyrobutyricum, indicating that the co-culture is capable of utilizing more of the acetate within the media and producing a subsequently higher amount of butyric acid. One possible explanation that can account for the lower acetate levels found in the co-culture is the glucose concentration, which was undetectable (< 0.1 g/L) during the exponential growth phase. Consequently, it would appear that it was consumed immediately by the culture as soon as it was produced from sucrose. Under similar glucose-limited conditions, other groups have found that C. tyrobutyricum ATCC 25755T shifts more carbon to the butyrate pathway instead of acetate production [3, 21].
When lactic acid was added along with glucose, C. tyrobutyricum ATCC 25755T showed an enhanced growth and butyric acid production, which can be in part attributed to the pH buffering effect described above. In confirmation of Figure 4A, the acetate concentration did not increase, as was seen when glucose was added alone, while the lactic acid concentration decreased, suggesting that these two acids were once more being converted into butyric acid.
Lactic acid is converted only into butyric acid not acetic acid
Optimizing pH conditions in the co-culture for butyric acid production
The optimum pH for culturing B. sp. SGP1 was around 7.0 while more acidic conditions led to a decreased growth (Additional file 3: Figure S3). However, for C. tyrobutyricum ATCC 25755T, the optimum pH for growth and butyric acid production is known to be 5.9 to 6.0 [4–6, 23], a value that we used previously to produce more than 50 g/L butyric acid . For this co-culture, we tried fed-batch fermentations at different pH values (5.9, 5.7, 5.5, and 5.3). All fermentations were performed at 37°C as this temperature was found to be the best for the co-culture based upon butyric acid final concentration (Additional file 4: Figure S4).
On the other hand, at the higher pH values tested (5.7, and 5.9), the final butyric acid concentrations were both lower; 26.3 g/L and 16.6 g/L for pHs 5.7 and 5.9 respectively while the acetic acid final concentrations were 1.0 g/L and 1.5 g/L respectively. It is thought that this may be attributed to the enhanced growth of B. sp. SGP1 at these higher pH values, leading to competition for the sugars produced. On the other hand, the final butyrate and acetic acid concentrations with the lowest pH tested, i.e., 5.3 were 20.8 and 10.0 g/L respectively. It was also observed that the glucose concentration was lower than 0.1 g/L throughout the course of fermentation when the pH was 5.5, 5.7, and 5.9 while, at pH 5.3, it was increasing constantly during the course of fermentation to reach 10.0 g/L near the end of the fermentation (Data not shown). It is also worthy to note that, with an optimum pH of 5.5, the sucrose hydrolysis was higher during the early stages of the fermentation than during the latter (>100 h), suggesting that the Bacillus strain was limited in its growth or activity.
The ability of the co-culture to grow in RCM medium without previous purging with an inert gas
Although C. tyrobutyricum ATCC 25755T alone is unable to utilize sucrose, using a co-culture of C. tyrobutyricum ATCC 25755T and B. sp. SGP1, this study demonstrates that sucrose can be utilized as a carbon source for butyric acid production. In addition, this co-culture offers also the benefit of a greater selectivity for butyric acid. Another advantage of this co-culture is its ability to produce butyric acid at a relatively high yield, which is similar to the yield obtained by pure cultures of C. tyrobutyricum ATCC 25755T grown on glucose. One drawback for this proposed co-culture is the low productivity during fed-batch fermentations. This can be improved in future studies by changing the medium composition or the fermentation mode to optimize the growth kinetics of the two strains. Likewise, future work may also include using this novel B. sp. SGP1 strain or other Bacillus strains in similar co-cultures with C. tyrobutyricum ATCC 25755T for production of butyric acid but from other carbon sources, such as starch or cellulose.
Materials and methods
All of the chemicals and reagents used were purchased from the Sigma-Aldrich Co. USA, unless mentioned otherwise. The reagents, when available, were molecular biology grade. All solutions were prepared using these reagents and sterile distilled water.
Bacterial strains and culturing methods
B. sp. SGP1 was isolated from a waste water sludge sample obtained from the Hyum Ki So Hwa Jo anaerobic digestion tank, Seoul, South Korea. This strain was co-cultured with C. tyrobutyricum ATCC 25755T in reinforced clostridial media (RCM). The media was made fresh using the individual components and supplemented with 40.0 g/L of either sucrose for the co-culture tests or B. sp. SGP1 monocultures or glucose for C. tyrobutyricum ATCC 25755T monocultures. The composition of the RCM medium was as follows: beef extract (Difco, USA) 10.0 g/L, peptone (Difco, USA) 5.0 g/L, yeast extract (Difco, USA) 3.0 g/L, tryptone (Difco, USA) 5.0 g/L, starch 1.0 g/L, sodium acetate 3.0 g/L, and NaCl 5.0 g/L. The medium was purged with argon gas and aliquoted into 60 ml serum bottles (20 ml in each), after which it was sterilized by autoclaving at 121°C for 20 minutes. For some experiments, P2 medium was used instead of RCM. The composition of the P2 medium was as follows: yeast extract (Difco, USA) 1.0 g/L, K2HPO4 0.5 g/L, KH2PO4 0.5 g/L, ammonium acetate 2.2 g/L, p-aminobenzoate 0.001 g/L, thiamin 0.001 g/L, biotin 0.00001 g/L, MgSO4.7H2O 0.2 g/L, MnSO4.H2O 0.01 g/L, FeSO4.7H2O 0.01 g/L, and NaCl 0.01 g/L. Unless else is mentioned, incubation and growth of the cultures was done in a shaking incubator at 37°C and 200 rpm.
Quantitative analysis of the volatile fatty acids (acetic and butyric acids) was done using GC (gas chromatography) (6890N, Agilent Technology, USA). Helium was used as the carrier gas. The oven temperature was programmed to increase from 50 to 240°C at a rate of 10°C/min. The temperature of the injection port and FID (Flame Ionization Detector) were 250°C. The concentration of the acids was determined according to a standard calibration curve. For analysis of the sugars and lactic acid, an HPLC (1200 Series, Agilent Technology, USA) with an RID (Refractive Index Detector) was used under the following conditions: mobile phase, 0.01M sulfuric acid; column, Aminex HPX-87H (dimensions - 300*7.8, Biorad, USA); flow rate, 0.5 ml/min; temperature, 40°C. The optical densities of the samples were measured using spectrophotometer (UV-mini 1240, Shimadzu, Japan) at 600 nm. To prove that [1, 2, 3 13C3] lactic acid is converted to butyric acid and not acetic acid by C. tyrobutyricum ATCC 25755T, spent culture broth was analyzed using GC equipped with a Pegasus 2D TOF-MS (LECO, St. Joseph, MI, USA) for detecting the presence of 13C in acetic and butyric acids. The column was HP-INNOWAX Polyethylene glycol (PEG) capillary column (17 m × 0.32 mm internal diameter, 0.25 μm film thickness, Agilent Technology, USA). The oven temperature was programmed to increase from 50°C to 150°C at a rate of 10°C/min. Helium was used as the carrier gas with a flow rate of 3 ml/min. The ionization mode employed was EI (electron impact ionization) with a scan rate of 100 spectra/sec and mass range of 40~500 m/z. The ion source temperature was 230°C .
Identification of the isolated Bacillus strain
The 16S rRNA gene sequence of the strain SGP1 was used as a query against the EzTaxon-e database [25, 26] and the 16S rRNA gene sequence of relatives were retrieved. The construction and evaluation of the phylogenetic trees were performed as described previously . The alignment and phylogenetic analysis were performed using MEGA5 . The 16s rRNA gene sequence of the strain SGP1 was submitted to GenBank (http://www.ncbi.nlm.nih.gov) and is registered under accession number HQ188291.1.
Fed-batch fermentations were done using a 3 L capacity bioreactor (Fermentec Co. Ltd., Korea) with an initial working volume of 1 L. The medium was RCM supplemented with 60.0 g/L sucrose (Sigma-Aldrich Co. USA). After autoclaving, the medium was purged with filtered oxygen-free argon gas. Overnight pre-cultures of the two strains were inoculated into the reactor at a concentration of 5% (v/v) for each strain. All fermentations were performed using agitation (200 rpm) and the pH was adjusted to the required values (5.3 – 5.9) by addition of a 10% ammonia solution, prepared by diluting an ammonium hydroxide solution (Cat # 320145, Sigma-Aldrich Co. USA) using sterile distilled water.
Observing levansucrase enzyme on an SDS-PAGE gel
A one-day (24 h) aerobic culture of B. sp. SGP1 was centrifuged and the supernatant was taken as a source for levansucrase enzyme. Concentration of the supernatant proteins was done through addition of acetone, centrifugation and then re-dissolving the precipitated proteins in phosphate buffer (pH 7.0). SDS-PAGE gel was then carried out as described by Laemmli with a 7% (w/v) acrylamide gel. After migration, the gel was cut into two halves. One half (containing the protein markers) was stained with Coomassie blue. The other half (containing the samples) was stained using periodic acid-Schiff reagent method [29–32] as follows: first, the gel was washed in a washing buffer containing 10 mM phosphate buffer (pH 7), CaCl2 (2 mM), and Tween 80 (10.0 g/L). Then the gel was incubated in the same solution but with 50 g/L sucrose, for 24 h at 37°C. After that, the formed levan was fixed on the gel through immersing in 70% ethanol, then the oxidizing solution (periodic acid o.7% (w/v), acetic acid 5% (w/v)) was added, followed by sodium bisulfate washing, and finally immersing in Schiff’s reagent (Sigma Aldrich, USA, Cat #; 3952016).
Sucrase and levan forming activity assay
The sucrase and levan forming activities at different temperatures and different pH values were determined by adding 50 μl of spent media, i.e. without cells, from a 20 h old B. sp. SGP1 culture to 950 μl of the reaction mixture (0.5 M sucrose solution in phosphate buffer) . To study the effect of the temperature, the pH was first adjusted to 6.6 before incubation at the test temperatures. Likewise, a temperature of 37°C was used when studying the effect of the pH. The incubation was performed for 10 h, after which, the sucrase and the levan forming activities were measured. Sucrase activity was determined by measuring the glucose liberated using glucose kits (Merck, Germany, Cat # 1.16720.0001), while levan formed was measured using the following procedure: 500 μl of the sample was added to 750 μl ethanol (to precipitate levan), and then the mixture was centrifuged. The precipitate was then washed with ethanol again, dried and then boiled with 0.1 N HCl for 30 min. Finally, the liberated fructose was analyzed using HPLC [14, 18].
The authors would like to thank Mr Taewoo Yum (Clean Energy Center, KIST, South Korea) for his help in the isotope analyses. This work was supported by the Advanced Biomass R&D Center (ABC) of the Global Frontier Project and by the National Research Foundation of Korea (Grant no. NRF-2009-C1AAA001-2009-0093499) both of which are funded by the Ministry of Education, Science and Technology. The authors are grateful for their support.
- Zigova J, Sturdik E: Advances in biotechnological production of butyric acid. J Ind Microbiol Biot 2000, 24:153–160.View Article
- Dwidar M, Park JY, Mitchell RJ, Sang BI: The future of butyric acid in industry. TheScientificWorldJournal 2012, 2012:471417.View Article
- Zhang CH, Yang H, Yang FX, Ma YJ: Current Progress on Butyric Acid Production by Fermentation. Curr Microbiol 2009, 59:656–663.View Article
- Liu XG, Zhu Y, Yang ST: Butyric acid and hydrogen production by Clostridium tyrobutyricum ATCC 25755 and mutants. Enzyme Microb Tech 2006, 38:521–528.View Article
- Zhu Y, Yang ST: Effect of pH on metabolic pathway shift in fermentation of xylose by Clostridium tyrobutyricum . J Biotechnol 2004, 110:143–157.View Article
- Wu ZT, Yang ST: Extractive fermentation for butyric acid production from glucose by Clostridium tyrobutyricum . Biotechnol Bioeng 2003, 82:93–102.View Article
- Mitchell RJ, Kim JS, Jeon BS, Sang BI: Continuous hydrogen and butyric acid fermentation by immobilized Clostridium tyrobutyricum ATCC 25755: Effects of the glucose concentration and hydraulic retention time. Bioresource Technol 2009, 100:5352–5355.View Article
- Bryant MP, Burkey LA: The Characteristics of Lactate-Fermenting Sporeforming Anaerobes from Silage. J Bacteriol 1956, 71:43–46.
- Gibson T: Clostridia in Silage. J Appl Bacteriol 1965, 28:56.
- Dwidar M, Lee S, Mitchell RJ: The production of biofuels from carbonated beverages. Appl Energ 2012, 100:47–51.View Article
- He GQ, Kong Q, Chen QH, Ruan H: Batch and fed-batch production of butyric acid by Clostridium butyricum ZJUCB. J Zhejiang University 2005, 6B:5.
- Zigova J, Sturdik E, Vandak D, Schlosser S: Butyric acid production by Clostridium butyricum with integrated extraction and pertraction. Process Biochem 1999, 34:835–843.View Article
- Jiang L, Cai J, Wang JF, Liang SZ, Xu ZN, Yang ST: Phosphoenolpyruvate-dependent phosphorylation of sucrose by Clostridium tyrobutyricum ZJU 8235: Evidence for the phosphotransferase transport system. Bioresource Technol 2010, 101:304–309.
- Viikari L, Gisler R: By-products in the fermentation of sucrose by different Zymomonas strains. Appl Microbiol Biotechnol 1986, 23:5.View Article
- Chang JJ, Chou CH, Ho CY, Chen WE, Lay JJ, Huang CC: Syntrophic co-culture of aerobic Bacillus and anaerobic Clostridium for bio-fuels and bio-hydrogen production. Int J Hydrogen Energ 2008, 33:5137–5146.View Article
- Tran HTM, Cheirsilp B, Hodgson B, Umsakul K: Potential use of Bacillus subtilis in a co-culture with Clostridium butylicum for acetone-butanol-ethanol production from cassava starch. Biochem Eng J 2010, 48:260–267.View Article
- Ghaly AE, Arab F, Mahmoud NS, Higgins J: Production of Levan by Bacillus licheniformis for Use as a Soil Sealant in Earthen Manure Storage Structures. Am J Biotechnol Biochem 2007, 3:8.
- Ssenthikumar V, Gunasekaran P: Influence of fermentation conditions on levan production by Zymomonas mobilis CT2. Indian J Biotechnol 2005, 4:6.
- Hettwer U, Gross M, Rudolph K: Purification and Characterization of an Extracellular Levansucrase from Pseudomonas-Syringae Pv Phaseolicola. J Bacteriol 1995, 177:2834–2839.
- Ben Ammar Y, Matsubara T, Ito K, Izuka M, Limpaseni T, Pongsawasdi P, Minamiura N: Characterization of a thermostable levansucrase from Bacillus sp TH4–2 capable of producing high molecular weight levan at high temperature. J Biotechnol 2002, 99:111–119.View Article
- Fayolle F, Marchal R, Ballerini D: Effect of Controlled Substrate Feeding on Butyric-Acid Production by Clostridium tyrobutyricum . J Ind Microbiol 1990, 6:179–183.View Article
- Kim BH, Gadd GM: Anaerobic fermentation. In Bacterial Physiology and Metabolism. 1st edition. United States of America Cambridge University Press; 2008:265–267.View Article
- Liu XG, Yang ST: Kinetics of butyric acid fermentation of glucose and xylose by Clostridium tyrobutyricum wild type and mutant. Process Biochem 2006, 41:801–808.View Article
- Oshiro M, Hanada K, Tashiro Y, Sonomoto K: Efficient conversion of lactic acid to butanol with pH-stat continuous lactic acid and glucose feeding method by Clostridium saccharoperbutylacetonicum . Appl Microbiol Biot 2010, 87:1177–1185.View Article
- Chun J, Lee JH, Jung Y, Kim M, Kim S, Kim BK, Lim YW: EzTaxon: a web-based tool for the identification of prokaryotes based on 16S ribosomal RNA gene sequences. Int J Syst Evol Microbiol 2007, 57:2259–2261.View Article
- Kim OS, Cho YJ, Lee K, Yoon SH, Kim M, Na H, Park SC, Jeon YS, Lee JH, Yi H: Introducing EzTaxon-e: a prokaryotic 16S rRNA gene sequence database with phylotypes that represent uncultured species. Int J Syst Evol Microbiol 2012, 62:716–721.View Article
- Kim S, Jeong H, Chun J: Clostridium aestuarii sp. nov., from tidal flat sediment. Int J Syst Evol Microbiol 2007, 57:1315–1317.View Article
- Tamura K, Peterson D, Peterson N, Stecher G, Nei M, Kumar S: MEGA5: molecular evolutionary genetics analysis using maximum likelihood, evolutionary distance, and maximum parsimony methods. Mol Biol Evol 2011, 28:2731–2739.View Article
- Kapitany RA, Zebrowsk E: High-Resolution Pas Stain for Polyacrylamide-Gel Electrophoresis. Anal Biochem 1973, 56:361–369.View Article
- Miller AW, Robyt JF: Detection of Dextransucrase and Levansucrase on Polyacrylamide Gels by the Periodic Acid Schiff Stain - Staining Artifacts and Their Prevention. Anal Biochem 1986, 156:357–363.View Article
- Dols M, Remaud-Simeon M, Willemot RM, Vignon M, Monsan P: Characterization of the different dextransucrase activities excreted in glucose, fructose, or sucrose medium by Leuconostoc mesenteroides NRRL B-1299. Appl Environ Microb 1998, 64:1298–1302.
- Munoz-Gutierrez I, Rodriguez-Alegria ME, Munguia AL: Kinetic behaviour and specificity of beta-fructosidases in the hydrolysis of plant and microbial fructans. Process Biochem 2009, 44:891–898.View Article
This article is published under license to BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.