Hyperosmotic response of streptococcus mutans: from microscopic physiology to transcriptomic profile
- Chengcheng Liu†1,
- Yulong Niu†2,
- Xuedong Zhou1,
- Keke Zhang1,
- Lei Cheng1,
- Mingyun Li1,
- Yuqing Li1,
- Renke Wang1,
- Yi Yang2 and
- Xin Xu1Email author
© Liu et al.; licensee BioMed Central Ltd. 2013
Received: 16 July 2013
Accepted: 26 November 2013
Published: 1 December 2013
Oral streptococci metabolize carbohydrate to produce organic acids, which not only decrease the environmental pH, but also increase osmolality of dental plaque fluid due to tooth demineralization and consequent calcium and phosphate accumulation. Despite these unfavorable environmental changes, the bacteria continue to thrive. The aim of this study was to obtain a global view on strategies taken by Streptococcus mutans to deal with physiologically relevant elevated osmolality, and perseveres within a cariogenic dental plaque.
We investigated phenotypic change of S. mutans biofilm upon hyperosmotic challenge. We found that the hyperosmotic condition was able to initiate S. mutans biofilm dispersal by reducing both microbial content and extracellular polysaccharides matrix. We then used whole-genome microarray with quantitative RT-PCR validation to systemically investigate the underlying molecular machineries of this bacterium in response to the hyperosmotic stimuli. Among those identified 40 deferentially regulated genes, down-regulation of gtfB and comC were believed to be responsible for the observed biofilm dispersal. Further analysis of microarray data showed significant up-regulation of genes and pathways involved in carbohydrate metabolism. Specific genes involved in heat shock response and acid tolerance were also upregulated, indicating potential cross-talk between hyperosmotic and other environmental stress.
Hyperosmotic condition induces significant stress response on S. mutans at both phenotypic and transcriptomic levels. In the meantime, it may take full advantage of these environmental stimuli to better fit the fluctuating environments within oral cavity, and thus emerges as numeric-predominant bacterium under cariogenic conditions.
Dental plaque is a densely-packed microbial biofilm and the residents living inside lead a “famine and feast” life style due to the fluctuation of nutrients within the oral cavity . In addition to many commonly studied environmental stimuli such as acidic and hyperthermic conditions to which dental plaque residents are frequently exposed, osmotic stress is also believed to have a great impact on dental plaque ecology and the development of dental caries . Acidogenic bacteria within dental plaque are able to metabolize carbohydrate to produce organic acids, which not only decrease the environmental pH, but also increase ionic strength of the plaque fluid due to tooth demineralization and consequent calcium and phosphate accumulation . It has been reported that the ionic strength of plaque fluid is doubled after sugar challenges, increasing from roughly 150 mM to approximately 300 mM [3, 4]. Thus, persistent residents within dental plaque have likely evolved sophisticated molecular machineries to counter the detrimental effect of elevated osmolality on their growth.
S. mutans is normal resident in the dental plaque and has been considered as the primary causative agent of dental caries for decades. S. mutans is able to take advantage of low pH to emerge as numerically predominant resident in cariogenic plaque [1, 2]. In addition, S. mutans has developed intricate machineries to counter those detrimental environmental challenges such as hyperosmotic stress, in order to persevere within the dental plaque [1, 5]. Many microorganisms respond to hyperosmotic challenges by increasing the intracellular levels of K+ and accumulating compatible solutes [6, 7]. The complete genome sequence of S. mutans has revealed several genes sharing homology with K+ transporters and the Opu family of ABC transporters of Escherichia Coli[8, 9]. These findings suggest that S. mutans may rally a series of intricately regulated genes and pathways to internalize K+ and compatible solutes, and thus perseveres under hyperosmotic conditions. A previous study from Burne’s group has identified a few candidates involved in the hyperosmotic stress response of S. mutans, and a possible cross-talk between osmotic and oxidative stress responses in S. mutans has also been suggested . However, the traditional “hypothesis-driven approach” to investigating selected genes sharing homology with stress response related genes of model bacteria may not suffice to give a global knowledge about the strategies used by this caries pathogen to cope with hyperosmotic challenges.
Although a number of studies have described transcriptional responses of S. mutans under various conditions [11–15], the molecular response of this bacterium under physiologically relevant hyperosmotic condition has not been profiled at transcriptomic level. In this study, we used microarray to profile the transcriptome of S. mutans under hyperosmotic conditions. Several genes and pathways were identified and further correlated with phenotypic changes of the organism observed under hyperosmotic challenges. The aim of this work is to provide a comprehensive insight into the sophisticated machineries adopted by S. mutans to better fit the physiologically relevant elevated osmolality, and thus perseveres within the oral cavity.
Results and discussion
Hyperosmotic conditions initiate biofilm dispersal
Selected genes up- or down-regulated 2-fold or more under hyperosmotic stress
Putative ribosome-associated protein
Putative PTS system
Putative lactoylglutathione lyase
Putative Clp proteinase
Putative transcriptional regulator
Putative maltose/maltodextrin ABC transporter
Putative transcriptional regulator
Putative permease; multidrug efflux protein
Competence stimulating peptide
Putative Zn-dependent protease
30S ribosomal protein S20
Putative thioredoxin reductase (NADPH)
Transcriptional regulator Spx
Putative transcriptional regulator/aminotransferase
Putative MDR permease; multidrug efflux pump
Putative glycerol uptake facilitator protein
It’s noteworthy that a recent transcriptomic profiling of S. mutans in the presence of oxygen also showed significant down-regulation of gtfB and genes involved in ComCDE quorum sensing system . This suggests that a motile lifestyle may be a common strategy employed by S. mutans to adapt adversary conditions.
S. mutansincreases carbohydrates consumption in response to hyperosmotic challenge
Hyperosmotic challenge prepares S. mutansfor better fitness under multiple environmental stimuli
As mentioned above, several genes involved in the carbohydrate metabolism of S. mutans were up-regulated. S. mutans may take full advantage of this increased energy generation to cope with multiple environmental stimuli. Previously study from Burne’s group has shown that two oxidative stress genes, sodA and nox were induced during hyperosmotic stress, and certain up-regulated gene (Smu.2115) upon hyperosmotic challenges was also involved in acid/oxidative stress responses . These findings suggest a potential cross-talking between hyperosmotic stress responses and other environmental responses of S. mutans. In the current study, we found that Lactoylglutathione lyase (lgl, smu1603), and ClpB (smu1425) were significantly induced during hyperosmotic stress (Table 1 and Figure 3). lgl has been shown to play an essential role in the acid tolerance response of S. mutans by detoxifying cytotoxic metabolite methyglyoxal in the cytoplasm . Therefore, up-regulation of lgl under hyperosmotic conditions may enhance the aciduricity of S. mutans. ClpB encodes a chaperone subunit with two ATP-binding domains involved in heat shock response . Previous study from Burne’s group has also shown a significant up-regulation of ClpB in S. mutans during oxygen challenge . The up-regulation of ClpB upon hyperosmotic challenge may assist unfolding the denatured protein amassed during environment stimuli, thus promoting the fitness of S. mutans under other detrimental conditions such as oxidative and heat stresses.
On the other hand, it has been demonstrated that dispersal cells from bacterial biofilm can colonize different and/or more niches than the bacteria that initiated the original biofilm, leading to better fitness of those bacteria in the environment . The induced dispersal of S. mutans biofilm under hyperosmotic stress may to an extent enhance the colonizing capacity of S. mutans cells, leading to a potential “niche expansion”, and thus benefit its fitness under the fluctuating environment within the oral cavity.
Taken together, this study has investigated phenotypic and transcriptional effects of hyperosmotic stress on S. mutans, and revealed genes and pathways essential for the hyperosmotic tolerance in this caries associated bacterium. We believe that although hyperosmotic challenge may induce significant stress response on bacteria, S. mutans has evolved sophisticated molecular machineries to counter those elicited detrimental effects. Additionally, S. mutans can mobilize genes and pathways to take full advantage of these environmental stimuli to better fit the fluctuating environments within the oral cavity, and thus emerge as the numeric-predominant bacteria under cariogenic conditions such as frequent sugar uptake.
Bacteria strains and culture conditions
Streptococcus mutans UA159 was commercially obtained from the American Type Culture Collection (ATCC). Bacteria were grown in brain heart infusion broth (BHI; Difco, Sparks, MD, USA) at 37°C in a 5% CO2 atmosphere until the cells reached the mid-logarithmic phase (OD600nm = 0.5). To determine the sub-inhibitory level of hyperosmotic challenge, bacteria were grown in BHI supplemented with 0.05, 0.1, 0.2, 0.4, 0.5, 0.6, 0.8, 1.0 M of sodium chloride respectively. For in vitro biofilm establishment, bacterial cells were grown in BHI supplemented with 1% sucrose (wt/vol).
Bacteria susceptibility assays
The sub-inhibitory concentration of sodium chloride was determined by a microdilution method as described previously . Growth curves of S. mutans UA159 were further constructed by monitoring the optical density (OD600nm) of the cultures for 24 h using a Bioscreen C analyzer (Oy Growth Curves AB Ltd., Finland) . The formation of S. mutans biofilm under increasing concentrations of NaCl was quantified in a 96-well microtiter plate as described previously . Briefly, S. mutans UA159 (1 × 106 CFU/ml) was grown in BHI supplemented with 1% (wt/vol) sucrose and NaCl (0.05 M to 1.0 M) at 37°C for 24 h. The culture supernatant from each well was then decanted, and the adherent biofilm was washed three times with PBS, fixed with methanol for 15 min, and stained with 0.1% (wt/vol) crystal violet (Sigma-Aldrich Corp., St. Louis, MO, USA) for 5 min. Subsequently, the wells were rinsed with deionized water until the blank wells appeared colorless; 200 μl of 95% ethanol was added. The plates were shaken at room temperature for 30 min, and the absorbance at 595 nm was recorded. The short-term effect of hyperosmotic challenge on the pre-established biofilm was also determined by quantification of the biomass of 24 h S. mutans biofilm after exposure to 0.4 M NaCl for 15 min using the same method as described above. All the experiments were performed in three-replicates and the average was calculated.
Biofilm viability assays
24 h pre-established S. mutans biofilms were treated with 0.4 M NaCl for 15 min, gently harvested with a cell scraper, and suspended in PBS to an OD600nm of 1.0. The bacterial cells suspension was then serially diluted and plated in triplicate on BHI agar plates. After 48 hours incubation at 37°C (5% CO2), colony forming unit (CFU) of biofilms was enumerated.
The treated biofilms were also stained with a two-color fluorescence assay kit (LIVE/DEAD BacLight-Bacterial Viability Kit 7012, Invitrogen, Molecular Probes, Inc., Eugene, OR, USA) according to the manufacturer’s instructions. The biofilms images were captured using a Leica TCS SP2 confocal laser scanning microscope (Leica, Germany), and the percentage of viable cells was calculated by Image Pro-Plus 6.0 (Media Cybernetics Inc., Bethesda, MD, USA).
Microbial biofilm configuration
Scanning electron microscopy (SEM) was performed as described previously  to investigate the configuration of S. mutans biofilm under hyperosmotic condition. S. mutans biofilms were either established on glass slides in the presence of 0.4 M of NaCl for 24 h, or pre-established 24 h biofilm on glass slides and then treated with 0.4 M of NaCl for 15 min. Biofilm samples were gently washed two times with sterile PBS to remove planktonic cells and fixed with 2.5% glutaraldehyde at 4°C overnight. The samples then were dehydrated in a graded series of ethanol (50%, 60%, 70%, 80%, 90%, 95% and 100%), dried in a freeze dryer, gold coated and observed under a SEM (FEI, Hillsboro, OR, USA).
The biofilm samples were also double-labeled by the method as described by Koo et al. [27, 28]. In brief, the extracellular polysaccharides matrix of S. mutans biofilm was labeled by incorporating 2.5 μmol l-1 of Alexa Fluor 647-labelled dextran conjugate (10000 MW; absorbance/fluorescence emission maxima of 650/668 nm; Molecular Probes Inc., Eugene, OR, USA) into the newly formed glucan. The bacterial cells in biofilms were labeled by means of SYTO 9 green fluorescent nucleic acid stain (2.5 μmol 1-1, 480/500 nm; Molecular Probes Inc.). The biofilm images were captured using a Leica TCS SP2 confocal laser scanning microscope (Leica, Germany). The confocal image stacks were analyzed by the image-processing software COMSTAT as described previously . The three-dimensional architecture of the biofilms was visualized using AmiraTM5.0.2 (Mercury Computer Systems, Chelmsford, MS, USA).
Mid-logarithmic phase cells of S. mutans (OD600nm = 0.5) were incubated with 0.4 M of NaCl at 37°C for 15 min. Cells were collected and then treated with RNAprotect reagent (Qiagen, Valencia, CA, USA) immediately. Total RNA was extracted using RNeasy Mini kits (Qiagen) as described previously . Rnase-Free DNase Set (Qiagen) was used to remove genome DNA. A Nanodrop ND 1000 spectrophotometer (Thermo Fisher Scientific, Pittsburgh, PA, USA) was used to determine total RNA concentrations, and an Agilent 2100 Bioanalyser (Agilent Technologies, Santa Clara CA, USA) was used to evaluate the RNA quality (see Additional file 2 for RNA quality control). The isolated RNA was stored at −80°C before use.
Streptococcus mutans UA159 (NC004350) NimbleGen Genechip (4*72 K) whole-genome array was employed for transcriptional profiling in this study. The oligoarrays included 1949 S. mutans UA159 open reading frames with twelve 24-mer probe pairs (PM/MM) per gene, and each probe was replicated 3 times. The design also included random GC and other control probes. Array images were scanned by Gene Pix® 4000B Microarray Scanner (Axon Instruments, Union City, CA, USA). Raw data were normalized using RMA algorithm by Roche NimbleScan software version 2.6. We used the average value of each replicate probe as the signal intensity for the corresponding gene, and all the values were log2 transformed for further analysis. The normalized data with annotation information was processed by combining several different R/Bioconductor packages. We conducted a non-parametric statistical method contained in the RanProd package to detect the differentially expressed genes (DEG) . With 100,000 permutation test, genes having a minimum 2-fold change with the false discovery rate (FDR) smaller than 0.1 were considered as DEG, indicating a significant up- or down-regulation under hyperosmotic stress. For the pathway analysis, we firstly constructed the whole S. mutans UA159 pathway database based on the KEGG Pathway. Then gene set enrichment analysis (GSEA) was used to determine the pathways that changed significantly in response to hyperosmotic stress [32, 33].
The microarray results were further validated by quantitative RT-PCR for selected genes (see Additional file 3 for detailed primer sequences for qPCR). Quantitative RT-PCR assays were performed using a SYBR Green reverse transcription-PCR kit (TaKaRa, Dalian, China) according to the manufacturer’s instructions.
We used Student’s T-test to compare the non-treated control groups with treatment groups. All statistical procedures were conducted by R software . Data were considered significantly different if the two-tailed P-value was < 0.05.
Microarray data accession
All the microarray raw data have been submitted to the NCBI Gene Expression Omnibus database under the accession number GSE47170 (http://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE47170).
This work was supported by National Natural Science Foundation of China (grant number: 81170959), Doctoral Fund of Ministry of Education of China (grant number: 20120181120002) and National Natural Science Foundation of China (grant number: 81200782). The authors would like to thank Arne Heydorn from Section of Molecular Microbiology, the Technical University of Denmark, for proving image-processing software COMSTAT.
- Lemos JA, Burne RA: A model of efficiency: stress tolerance by streptococcus mutans. Microbiology. 2008, 154 (Pt 11): 3247-3255.PubMedPubMed CentralView ArticleGoogle Scholar
- Lemos JA, Abranches J, Burne RA: Responses of cariogenic streptococci to environmental stresses. Current Issues Mol Biol. 2005, 7 (1): 95-107.Google Scholar
- Gao XJ, Fan Y, Kent RL, Van Houte J, Margolis HC: Association of caries activity with the composition of dental plaque fluid. J Dental Res. 2001, 80 (9): 1834-1839. 10.1177/00220345010800091201.View ArticleGoogle Scholar
- Margolis HC, Duckworth JH, Moreno EC: Composition of pooled resting plaque fluid from caries-free and caries-susceptible individuals. J Dental Res. 1988, 67 (12): 1468-1475. 10.1177/00220345880670120601.View ArticleGoogle Scholar
- Liu YL, Nascimento M, Burne RA: Progress toward understanding the contribution of alkali generation in dental biofilms to inhibition of dental caries. Int J Oral Sci. 2012, 4 (3): 135-140. 10.1038/ijos.2012.54.PubMedPubMed CentralView ArticleGoogle Scholar
- Sleator RD, Hill C: Bacterial osmoadaptation: the role of osmolytes in bacterial stress and virulence. FEMS Microbiol Rev. 2002, 26 (1): 49-71. 10.1111/j.1574-6976.2002.tb00598.x.PubMedView ArticleGoogle Scholar
- Weber A, Jung K: Profiling early osmostress-dependent gene expression in escherichia coli using DNA microarrays. J Bacteriol. 2002, 184 (19): 5502-5507. 10.1128/JB.184.19.5502-5507.2002.PubMedPubMed CentralView ArticleGoogle Scholar
- Epstein W: The roles and regulation of potassium in bacteria. Prog Nucleic Acid Re. 2003, 75: 293-320.View ArticleGoogle Scholar
- Ajdic D, McShan WM, McLaughlin RE, Savic G, Chang J, Carson MB, Primeaux C, Tian RY, Kenton S, Jia HG, et al: Genome sequence of streptococcus mutans UA159, a cariogenic dental pathogen. P Natl Acad Sci USA. 2002, 99 (22): 14434-14439. 10.1073/pnas.172501299.View ArticleGoogle Scholar
- Abranches J, Lemos JA, Burne RA: Osmotic stress responses of streptococcus mutans UA159. Fems Microbiol Lett. 2006, 255 (2): 240-246. 10.1111/j.1574-6968.2005.00076.x.PubMedView ArticleGoogle Scholar
- Shemesh M, Tam A, Kott-Gutkowski M, Feldman M, Steinberg D: DNA-microarrays identification of streptococcus mutans genes associated with biofilm thickness. Bmc Microbiol. 2008, 8: 236-10.1186/1471-2180-8-236.PubMedPubMed CentralView ArticleGoogle Scholar
- Shemesh M, Tam A, Aharoni R, Steinberg D: Genetic adaptation of streptococcus mutans during biofilm formation on different types of surfaces. Bmc Microbiol. 2010, 10: 51-10.1186/1471-2180-10-51.PubMedPubMed CentralView ArticleGoogle Scholar
- Ahn SJ, Wen ZT, Burne RA: Effects of oxygen on virulence traits of streptococcus mutans. J Bacteriol. 2007, 189 (23): 8519-8527. 10.1128/JB.01180-07.PubMedPubMed CentralView ArticleGoogle Scholar
- Biswas I, Drake L, Erkina D, Biswas S: Involvement of sensor kinases in the stress tolerance response of streptococcus mutans. J Bacteriol. 2008, 190 (1): 68-77. 10.1128/JB.00990-07.PubMedPubMed CentralView ArticleGoogle Scholar
- Klein MI, DeBaz L, Agidi S, Lee H, Xie G, Lin AH, Hamaker BR, Lemos JA, Koo H: Dynamics of streptococcus mutans transcriptome in response to starch and sucrose during biofilm development. Plos One. 2010, 5 (10): e13478-10.1371/journal.pone.0013478.PubMedPubMed CentralView ArticleGoogle Scholar
- Koo H, Xiao J, Klein MI: Extracellular polysaccharides matrix–an often forgotten virulence factor in oral biofilm research. Int J Oral Sci. 2009, 1 (4): 229-234. 10.4248/IJOS.09086.PubMedPubMed CentralView ArticleGoogle Scholar
- Ahn SJ, Wen ZT, Burne RA: Multilevel control of competence development and stress tolerance in streptococcus mutans UA159. Infect Immun. 2006, 74 (3): 1631-1642. 10.1128/IAI.74.3.1631-1642.2006.PubMedPubMed CentralView ArticleGoogle Scholar
- Perry JA, Jones MB, Peterson SN, Cvitkovitch DG, Levesque CM: Peptide alarmone signalling triggers an auto-active bacteriocin necessary for genetic competence. Mol MicroBiol. 2009, 72 (4): 905-917. 10.1111/j.1365-2958.2009.06693.x.PubMedPubMed CentralView ArticleGoogle Scholar
- Perry JA, Cvitkovitch DG, Levesque CM: Cell death in streptococcus mutans biofilms: a link between CSP and extracellular DNA. Fems Microbiol Lett. 2009, 299 (2): 261-266. 10.1111/j.1574-6968.2009.01758.x.PubMedPubMed CentralView ArticleGoogle Scholar
- Kempf B, Bremer E: Uptake and synthesis of compatible solutes as microbial stress responses to high-osmolality environments. Arch Microbiol. 1998, 170 (5): 319-330. 10.1007/s002030050649.PubMedView ArticleGoogle Scholar
- Matsui R, Cvitkovitch D: Acid tolerance mechanisms utilized by streptococcus mutans. Future MicroBiol. 2010, 5 (3): 403-417. 10.2217/fmb.09.129.PubMedPubMed CentralView ArticleGoogle Scholar
- McDougald D, Rice SA, Barraud N, Steinberg PD, Kjelleberg S: Should we stay or should we go: mechanisms and ecological consequences for biofilm dispersal. Nat Rev Microbiol. 2012, 10 (1): 39-50.Google Scholar
- Xu X, Zhou XD, Wu CD: The Tea catechin epigallocatechin gallate suppresses cariogenic virulence factors of streptococcus mutans. Antimicrob Agents Ch. 2011, 55 (3): 1229-1236. 10.1128/AAC.01016-10.View ArticleGoogle Scholar
- Olsen B, Murakami CJ, Kaeberlein M: YODA: software to facilitate high-throughput analysis of chronological life span, growth rate, and survival in budding yeast. BMC Bioinformatics. 2010, 11: 141-10.1186/1471-2105-11-141.PubMedPubMed CentralView ArticleGoogle Scholar
- Reuter M, Mallett A, Pearson BM, van Vliet AHM: Biofilm formation by campylobacter jejuni is increased under aerobic conditions. Appl Environ Microb. 2010, 76 (7): 2122-2128. 10.1128/AEM.01878-09.View ArticleGoogle Scholar
- Hasan S, Danishuddin M, Adil M, Singh K, Verma PK, Khan AU: Efficacy of E. Officinalis on the cariogenic properties of streptococcus mutans: a novel and alternative approach to suppress quorum-sensing mechanism. Plos One. 2012, 7 (7): e40319-10.1371/journal.pone.0040319.PubMedPubMed CentralView ArticleGoogle Scholar
- Xiao J, Koo H: Structural organization and dynamics of exopolysaccharide matrix and microcolonies formation by streptococcus mutans in biofilms. J Appl Microbiol. 2010, 108 (6): 2103-2113.PubMedGoogle Scholar
- Xiao J, Klein MI, Falsetta ML, Lu BW, Delahunty CM, Yates JR, Heydorn A, Koo H: The exopolysaccharide matrix modulates the interaction between 3D architecture and virulence of a mixed-species oral biofilm. Plos Pathog. 2012, 8 (4): e1002623-10.1371/journal.ppat.1002623.PubMedPubMed CentralView ArticleGoogle Scholar
- Heydorn A, Nielsen AT, Hentzer M, Sternberg C, Givskov M, Ersboll BK, Molin S: Quantification of biofilm structures by the novel computer program COMSTAT. Microbiology. 2000, 146 (Pt 10): 2395-2407.PubMedView ArticleGoogle Scholar
- Xue XL, Tomasch J, Sztajer H, Wagner-Dobler I: The delta subunit of RNA polymerase, RpoE, is a global modulator of streptococcus mutans environmental adaptation. J Bacteriol. 2010, 192 (19): 5081-5092. 10.1128/JB.00653-10.PubMedPubMed CentralView ArticleGoogle Scholar
- Hong FX, Breitling R, McEntee CW, Wittner BS, Nemhauser JL, Chory J: RankProd: a bioconductor package for detecting differentially expressed genes in meta-analysis. Bioinformatics. 2006, 22 (22): 2825-2827. 10.1093/bioinformatics/btl476.PubMedView ArticleGoogle Scholar
- KEGG: Kyoto Encyclopedia of Genes and Genomes. http://www.genome.jp/kegg/,
- Subramanian A, Kuehn H, Gould J, Tamayo P, Mesirov JP: GSEA-P: a desktop application for gene Set enrichment analysis. Bioinformatics. 2007, 23 (23): 3251-3253. 10.1093/bioinformatics/btm369.PubMedView ArticleGoogle Scholar
- The R Project for Statistical Computing. http://www.r-project.org/,
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.