Transcriptome Profile Analysis of Twisted Leaf Disease Response in Susceptible Sugarcane with Narenga porphyrocoma Genetic Background
- 1.6k Downloads
Sugarcane is an important industrial crop with a high sugar yield that has become a leading energy crop worldwide. It is widely cultivated in tropical and subtropical regions. Various diseases beset the cultivation of sugarcane. The molecular study of disease resistance in sugarcane is limited by its complicated genome. In our study, RNA-seq was employed to detect the mechanism of twisted leaf disease tolerance in modern cultivar sugarcane, which derived from Narenga porphyrocoma. We completed high-throughput transcriptomic sequencing of 12 samples, including three stages of a susceptible (NSBC1_T “H3–8”) and an unsusceptible cultivar (NSBC1_CK “H-19”) with two biological repeats, respectively. Using the Saccharum spontaneum genome as reference, the average mapping ratio of the clean data was over 70%. Among the differentially expressed genes between H3–8 and H3–19, we focused on the analysis of hormone pathways and resistance (R) genes. The results showed that twisted leaf disease triggers hormone networks and around 40% of R genes conditioned lower expression in the susceptible cultivar. One of the possible reasons for H3–8 being susceptible to twisted leaf disease might be the null/retarded response of R genes, especially in pre-onset stage (46% down-regulated) of pathogens infection.
KeywordsHormones R genes RNA-seq Sugarcane
Non-redundant protein sequences in NCBI
Clusters of orthologous group
Kyoto encyclopedia of genes and genomes databases
TIGR gene indices clustering tools
Basic Local Alignment Search Tool
Sugarcane belongs to the Poaceae family, which includes maize, wheat, rice, sorghum, and many types of grass (Li 2010). Cultivated sugarcane is an important industrial crop with a high sugar yield and has become a major energy crop worldwide (Grivet and Arruda 2001; Yang et al. 2010). It is cultivated in ~26 million hectares in tropical and subtropical regions of the world, producing up to 1.8 billion metric tons of crushable stems (Zhang et al. 2018). Sugarcane provides around 80% of the world’s sugar, with the secondary production as raw materials for pulp, ethanol and bioplastics (Lam et al. 2009; Li and Yang 2015). Modern sugarcane cultivars are interspecific hybrids which were generated from the cross between Saccharum officinarum and Saccharum spontaneum, followed by backcrossing into Saccharum officinarum to select sugar-poor relative traits (Roah 1972; Branes and Sartoris 1936). In China, modern sugarcane is mainly distributed in the provinces of Guangxi, Yunnan, Guangdong, and Hainan. Some cultivars were also improved disease resistance by crossing with Saccharum barberi Jesweit or/and Narenga porphyrocoma (Gao et al. 2012; Liu et al. 2012a, b).
Sugarcane diseases, such as smut, white leaf, and wilt/top rot/Pokkah Boeng, are critical limitations of production, causing serious losses in yield and quality among susceptible cultivars (Hameed et al. 2015; Paulo et al. 2016; Su et al. 2016). Traditionally, the resistant traits of Saccharum spontaneum, Saccharum barberi Jesweit or/and Narenga porphyrocoma were introduced into Saccharum officinarum by hybridization to improve the disease resistance of cultivated sugarcane (Gao et al. 2012; Liu et al. 2012a, b). With the development of breeding approaches, molecular breeding with precise genome information accelerates the collection of disease resistance genes in one cultivar. However, applying this to sugarcane is difficult due to its high polyploidy and complex genome, with ploidy levels ranging from 5× to 16×, and chromosome numbers from 2n = 40–128, with some even as high as 200 (Sreenivasan et al. 1987; Liu et al. 2012a, b). The estimated polyploid genome size of sugarcane ranges from 3.36 to 12.64 Gb, and the monoploid genome size ranges from 760 to 985 Mb (D’hont et al. 1996; Zhang et al. 2012), which is larger than the rice (400 MB) and the sorghum (760 MB) genomes (Soderlund et al. 2011). Such complex genetic background blocks the sequencing of the whole sugarcane genome. Studies on biological traits, such as biomass yield, sugar accumulation, and stress tolerance, have focused on transcriptome analysis (Kido et al. 2012; Fracasso et al. 2016; Huang et al. 2016). The transcriptome reveals specific transcripts produced under biotic stress (e.g., smut) and abiotic stress (e.g., drought) conditions in sugarcane (Iskandar et al. 2011; Mattielllo et al. 2015).
With the Saccharum spontaneum genome, which is relative high quality, as reference (Zhang et al. 2018), we employed transcriptome profiling to analyze the genes and pathways involved in twist leaf disease in modern sugarcane cultivar. Twisted leaf disease, caused by Phoma sp., which is one of the largest fungal genera, was first reported in Guangxi, China, in 2014, when more than 5% of sugarcane was infected in the field. Twisted leaf disease is somewhat similar to Pokkah Boeng disease (caused by Fusarium moniliforme Sheldon). The symptoms begin with yellowing on the midribs and leaf margins, then spread to the entire leaf, along with twisting and curling of the crown leaves (Lin et al. 2014). The modern sugarcane cultivar used in this study was generated from the BC1 generation offspring of Narenga porphyrocoma via crossing and backcrossing with cultivated sugarcane varieties, so that it harbored the Narenga porphyrocom genetic background. The RNA-seq (two biological repeats) comparison between un-susceptible and susceptible cultivars in different stages of infection provided a reference for understanding the mechanism of twisted leaf disease. Analysis revealed that twisted leaf disease triggered the hormone network. R genes expression profile showed ~40% R genes expressed lower in susceptible cultivar. The null/retarded response of R genes might be one of the possible reasons for H3–8 being susceptible to twisted leaf disease.
RNA Sequencing and Mapping to the Reference Genome
Overview of the transcriptome sequencing and de novo assembly results
Total nucleotides (nt)
Genome mapping summary of all samples
Total Clean Reads
Total Mapping Ratio
Gene Expression Statistics
To elucidate gene expression profile of un-susceptible and susceptible cultivars, we annotated the aligned reads according to the reference genome. The result showed that there were 91,386 genes expressed that included 65,391 known genes and 25,995 novel genes. In the 96,101 annotated transcripts, 53,481 transcripts with novel alternative splicing subtypes encode known proteins, 27,151 transcripts were defined as novel protein coding genes, and 15,469 transcripts were classified into long non-coding RNAs. Details of each sample were exhibited in Table S1 and Fig. S1.
Differentially Expressed Genes (DEGs) in Un-Susceptible and Susceptible Cultivars
Ingenuity Pathways Analysis
Also, we analyzed the specific biosynthesis genes in SA, JA and ET pathway, respectively. In details, the phenylalanine ammonia-lyase, an enzyme of phenylalanine ammonium lyase (PAL) pathway in SA synthesis (Yokoo et al. 2018), was increased in expression at the pre-onset of Phoma sp. infection. The isochorismate synthase 1, which is involved in the synthesis of SA (Yokoo et al. 2018), showed lower expression level in three stages of the susceptible cultivar (Table S1). In the JA pathway, 12-oxophytodienoate reductase, which catalyses the reduction of 10, 11-double bonds of 12-oxophytodienoate to 3-oxo-2-(2′-pentenyl)-cyclopentane-1-octanoic acid (OPC-8:0) (Tani et al. 2008), conditioned relative higher expression in the susceptible cultivar (Table S1). The aminotransferase, a gene family including VAS1 which participates in auxin and ethylene biosynthesis (Zhang et al. 2012), varied in expression at different stages of the twisted leaf disease. These findings supported that the infection of Phoma sp. induced the response of the hormones network.
Evaluation of Differentially Expressed Disease-Resistant Genes
Classification of annotated R genes and representatives for ingenuity analysis
Disease resistance protein (NBS-LRR class) family
Pyridoxal phosphate-dependent transferases superfamily protein
HOPZ-ACTIVATED RESISTANCE 1
NB-ARC domain-containing disease resistance protein
TIR-NBS-TIR-TIR-WRKY type disease resistance protein
S-locus lectin protein kinase family protein
cysteine-rich RLK (RECEPTOR-like protein kinase)
LRR and NB-ARC domains-containing disease resistance protein
NADPH HC toxin reductase-like protein
NAC domain containing protein
Seven transmembrane MLO family protein
Protein phosphatase 1 regulatory subunit
ADR1-like 1 disease resistance protein
The transcriptome, which is influenced by external environmental conditions, can effectively reveal the response mechanism of biotic and abiotic stress in plants (Wei et al. 2011). In our study, the natural infected cultivar H3–8 was defined as susceptible cultivar and the un-susceptible cultivar (H3–19) was used as control. With two biological repeats, we used transcriptome profiling to detect the genes associated with twisted leaf disease and obtained 65,780,017,250 bp data from three stages of the twisted leaf disease. Using the genome of Saccharum spontaneum L. as reference (Zhang et al. 2018), the average mapping ratio of twelve samples was 76.74%. The annotation of aligned reads showed 91,386 genes expressed (65,391 known and 25,995 novel), and 96,101 annotated transcripts, including 53,481 novel alternative splicing subtypes encode known proteins, 27,151 novel protein coding genes, and 15,469 long non-coding RNAs. With these data, we analyzed the DEGs, including hormone pathways and R genes, between the susceptible cultivar and the control. Also, these data provided a foundation for the further analysis of the twisted leaf disease, such as specific gene characterization, resistant genes selection, and immune network description.
The comparison between susceptible cultivar and the control identified 16,242, 12,892 and 34,924 DEGs in pre-onset, early stage and serious symptom stage, respectively (Fig. 3). These DEGs are involved in critical biological activities that are essential for disease resistance (Fig. 4). Ingenuity pathway analysis revealed hormone biosynthesis genes variations in the susceptible cultivar (Fig. 5). The SA and ET pathways were induced in the infection of Phoma sp. at the pre-onset stage. Along with the spread of the disease, JA and ET synthesis increased and maintained high levels in the serious symptom stages. As three major hormones involved in stress resistance, SA is a biotrophic and hemi-biotrophic pathogen-triggered signaling pathway, and JA combined with ET signaling pathways are induced by necrotrophic pathogens (Glazebrook 2005). Therefore, twisted leaf disease, caused by fungal pathogens, first triggers the SA pathway as response to pathogens infection. The high level of JA and ET in serious symptoms might partially result from tissue damage in disease. Other hormones, including ABA, auxin, and CK, crosstalk with major disease response hormones (SA, JA, and ET) and construct a network to respond to disease stress (Verma et al. 2016).
Plant disease resistance (R) genes are prevalent in all plant species and harbor a conserved LRR domain (Meyers et al. 1998). R genes are defined as gene-for-gene interaction and specifically recognize an avirulence protein encoded by a pathogen with a hypersensitive response (Flor 1956). In the R gene analysis of the susceptible cultivar, around 40% R genes, including those that play essential functions, were down-regulated, especially in pre-onset stage. For example, belongs to LRR and NB-ARC domains-containing proteins of which members participate in disease resistance (Fischer et al. 2016). is predicted to encode RGA protein which contains NB-ARC domain and functions disease resistance (Cesari et al. 2013). is descripted as CRKs (cysteine-rich protein kinase), which involves in the ROS signaling pathway and plays important roles in the elimination of fungal growth damage (Niina Idänheimo 2015). The low expression of these R genes in the twisted leaf disease would drag down the rapid response of R genes defensive system. Besides ~40% down-regulated R genes, there were still a number of R genes conditioned normal expression level or even higher expression level in the susceptible cultivar. Unfortunately, these high expressed R genes were unable to eliminate the pathogens successfully. These increased expressions might be a result of universal reactions of sugarcane under pathogens stress rather than the specific response of Phoma sp. infection. Such phenomenon was commonly reported in the studies of other species (Heath 2000; van Loon 2015).
Materials and Methods
Plant Materials and Sampling
Narenga porphyrocoma, which is an important relative genus of sugarcane, was collected from a barren mountainous area in Guangxi province, China. We obtained the BC1 generation offspring of the Narenga porphyrocoma via crossing and backcrossing with sugarcane varieties several years. Among the BC1 offspring, H3–8 was susceptible to twisted leaf disease at the sugarcane elongating stage, whereas H3–19 was not. H3–8 was proven to affect stable twisted leaf disease occurrence after 3 years based on field and greenhouse observations.
Plantlets at the same growth stage were selected and planted in pots filled with a mixture of peat soil and washed sand, and grown in a greenhouse at the Sugarcane Research Institute, Guangxi Academy of Agricultural Sciences. The natural infected plants were selected and defined as the susceptible cultivator. The leaf samples of the susceptible sugarcane clone H3–8 were collected at three stages corresponding to the pre-onset (named NSBC1_T1), early (named NSBC1_T2), and serious symptom (named NSBC1_T3) stages of the disease, respectively. The leaf samples of the un-susceptible sugarcane clone H3–19 were simultaneously collected, labeled NSBC1_CK1, NSBC1_CK2, and NSBC1_CK3, respectively. Two biological repeats of each sample were collected. All samples were immediately frozen in liquid nitrogen and stored at −80 °C.
RNA Extraction and Quality Determination
The total RNA was extracted using TRIzol reagent (Invitrogen, Carlsbad, CA, USA) and treated with RNase-free DNase I and RNA integrity number (RIN > 8.0). RNA quality and quantity were verified using a NanoDrop 1000 spectrophotometer (Wilmington, DE, CA, USA) and an Agilent 2100 Bioanalyzer (Santa Clara, CA, USA) prior to the library construction. No sign of degradation was found.
cDNA Library Construction and Sequencing
The poly(A) RNA was fragmented into approximately 300-nt fragments using RNA Fragmentation Reagents (Ambion, Austin, TX, USA). Using these short fragments as templates, the double-stranded cDNA was synthesized with random primers (Invitrogen, Carlsbad, CA, USA). The end repair of these cDNA fragments was subsequently performed with Klenow polymerase, T4 DNA polymerase, and T4 polynucleotide kinase (NEB,Ipswich, MA, USA). Illumina adapters (containing primer sites for sequencing and flow cell surface annealing) were ligated to the end-repaired fragments using T4 DNA ligase (Invitrogen, Carlsbad, CA, USA). The products were enriched for the cDNA fragments using a Qiaquick Gel Extraction Kit (Qiagen, Duesseldorf, Germany) and amplified with polymerase chain reaction (PCR) to prepare the sequencing library. Agilent 2100 Bioanalyzer (Agilent, Beijing, China) was used to detect the quantity and quality of the cDNA. Then, the paired-end RNA-seq libraries were prepared following Illumina’s protocols and sequenced on the Illumina HiSeq 2500 platform (Illumina, San Diego, CA, USA). Sequencing was performed in a flowcell on an Illumina HiSeq 2500 sequencer using the TruSeq Paired-End Cluster Kit v3 (Illumina PE-401-3001) and the TruSeq SBS HS Kit v3200 cycles (Illumina FC-401-3001), generating 2 × 125 bp.
Data Filtering and Genome Mapping
The raw reads were filtered according to the following criteria: (1) reads containing adaptor, (2) reads with unknown nucleotides larger than 5%, and (3) low-quality reads (the rate of reads with a quality value of ≤10 was more than 20%. The clean data was mapped to the genome of Saccharum spontaneum L. (Zhang et al. 2018) via HISAT (v2.0.4; −-phred64 --sensitive --no-discordant --no-mixed -I 1 -X 1000) (Kim et al. 2015). The transcripts were re-constructed via StringTie and predicted using Cuffcompare tool in Cufflink (Pertea et al. 2003; Trapnell et al. 2010; Pertea et al. 2016).
Differential Gene Expression
The DEGs analysis was performed as the description of DEGseq (Wang et al. 2010) and corrected the P-values to Q-values based on Benjamini-Hochberg method (Benjamini and Hochberg 1995) and Storey-Tibshirani method (Storey and Tibshirani 2003). GO (gene ontology) terms were assigned based on the best-hits BLASTx resulted from NR alignments that were derived from Blast2GO (v2.5.0) against GO database (release-20,120,801). The DEGs were aligned against KEGG (Kyoto Encyclopedia of Genes and Genomes) by BLASTx package with threshold of E-value <=10–5.
Ingenuity Analysis of DEGs
The rice hormones proteins were used as reference to identify the hormone genes in sugarcane (Cohen et al. 2017). The longest transcript of genes was selected as representative for analysis. All predicted protein sequences were aligned against the sugarcane gene models by BLASTp (Identity>50%; Coverage>50%; E-value: 1 × 107), and defined as homologs of hormones biosynthesis genes. Also, we aligned all predicted hormone genes against the sugarcane genome via tBLASTn (E-value: 1E-10), and used GeneWise (v2.2.0) to identify the structure of candidate hormone genes. Among these genes, the ones with non-significant different expression in three stages were eliminated. The final candidates were aligned against KEGG database (version:84). The log fold of each gene was obtained from the case-control group. The patterns of the gene in each group were compared. Similarly, predicted proteins were aligned against the PRG database using BLASTp. The results were filtered (E-value: 1e10), and the best hit of predicted protein was retained. The fpkm (Reads per kilobase of exon per million reads mapped) of each predicted resistant gene in the case sample was filtered out and a downstream analysis was performed.
We thank Mick Rose in Primary Industry Department of Australia for the assistance of writing paper.
Hongwei Tan, and Xihui Liu conceived and designed the experiments. Jinju Wei and Huiping Ou performed the experiments. Xiaoqiu Zhang, Ronghua Zhang, Hui Zhou, Yiyun Gui,Haibi Li, Yangrui Li, Rongzhong Yang and Dongliang Huang performed cross and backcross of Narenga porphyrocoma, planted sugarcane and collected samples. Zhihui Xiu, Junhui Chen and Huayan Jiang analyzed the data. Zhihui Xiu and Xihui Liu wrote the manuscript. All authors have read and approved the final manuscript.
This work was financially supported by the National Science Foundation of China (31760368, 31101195), Guangxi Fund (GKAA17202042–6), Fund of Guangxi Academy of Agricultural Sciences (GNK2018YT02 and GNK2018YM01) and Fund of Modern Agriculture Technology (CARS-170105, gjnytxgxcxtd-03).
Compliance with Ethical Standards
Conflict of Interest
The authors declare no conflict of interest.
The data that support the findings of this study have been deposited in the CNSA (https://db.cngb.org/cnsa/) of CNGBdb with accession code CNP0000220, and NCBI with accession code SAMN05428728 to SAMN05428739.
- Bari R, Jones JD (2009) Role of plant hormones in plant defence responses. Plant Mol Boil 69:473–488Google Scholar
- Benjamini Y, Hochberg Y (1995) Controlling the false discovery rate: a practical and powerful approach to multiple testing. J R Stat Soc Ser B 57:289–300Google Scholar
- Branes EW, Sartoris GB (1936) Sugarcane: its origin and improvement. Yerab US Dep Agric:561–624Google Scholar
- Cesari S, Thilliez G, Ribot C, Chalvon V, Michel C, Jauneau A, Rivas S, Alaux L, Kanzaki H, Okuyama Y, Morel JB, Fournier E, Tharreau D, Terauchi R, Kroj T (2013) The rice resistance protein pair RGA4/RGA5 recognizes the Magnaporthe oryzae effectors AVR-pia and AVR1-CO39 by direct binding. Plant Cell 25(4):1463–1481PubMedPubMedCentralGoogle Scholar
- Flor HH (1956) The complementary genic system in flax and flax rust. Adv Genet 8:29–54Google Scholar
- Fracasso A, Trindade LM, Amaducci S (2016) Drought stress tolerance strategies revealed by RNA-Seq in two sorghum genotypes with contrasting WUE. BMC Biol 16:115Google Scholar
- Gamalero E, Glick BR (2012) Ethylene and abiotic stress tolerance in plants. In: Environmental adaptations and stress tolerance of plants in the era of climate change. Springer, New York, pp 395–412Google Scholar
- Gao Y, Fang F, Liu XH et al (2012) Identification of progeny from crosses between sugarcane (Saccharum spp.) and intergeneric hybrid complex (Erianthus arundinaceus × Saccharum spontaneum) with molecular markers. J Plant Genet Res 13:912–916Google Scholar
- Glazebrook J (2005) Contrasting mechanisms of defense against biotrophic and necrotrophic pathogens. Annu Rev Phytopathol 43:205–227Google Scholar
- Grivet L, Arruda P (2001) Sugarcane genomics: depicting the complex genome of an important tropical crop. Curr Opin Plant Biol 5:122–127Google Scholar
- Hameed U, Pan YB, Iqbal J (2015) Genetic analysis of resistance gene analogues from a sugarcane cultivar resistant to red rot disease. J Phytopathol 163:755–763Google Scholar
- Huang DL, Gao YJ, Gui YY, Chen ZL, Qin CX et al (2016) Transcriptome of high sucrose sugarcane variety GT35. Sugar Tech 18:520–528Google Scholar
- Idänheimo N (2015) The role of cysteine-rich receptor-like protein kinases in ROS signaling in Arabidopsis thaliana. Department of Biosciences, University of Helsinki, FinlandGoogle Scholar
- Iskandar HM, Casu RE, Fletcher AT, Schmidt S, Xu JS et al (2011) Identification of drought-response genes and a study of their expression during sucrose accumulation and water deficit in sugarcane culms. BMC Plant Bio 11:12Google Scholar
- Kido EA, Ferreira NJRC, de Oliveira Silva RL, Pandolfi V, Guinaraes ACR et al (2012) New insights in the sugarcane transcriptome responding to drought stress as revealed by supersage. Scientific World J 2012:821062Google Scholar
- Lam E, Shine J, Da SJ, Lawton M, Bonos S et al (2009) Improving sugarcane for biofuel: engineering for an even better feedstock. GCB Bioenergy 1:251–255Google Scholar
- Li YR (2010) Modern sugarcane science. Beijing: China Agriculture Press 1–125Google Scholar
- Li YR, Yang LT (2015) Research and development priorities for sugarcane industry of China: resent research hignlights. Sugar Tech 17:9–12Google Scholar
- Liu XH, Fang FX, Gao YJ, Zhang RH, Song HZ et al (2012a) Identification and genetic analysis of hybrid from cross between Erianthus arundinacius (Retz.) Jesws. And Saccharum spontaneum L. Acta Agron Sin 38:914–920Google Scholar
- Liu XH, Fang FX, Zhang RH, Song HZ, Gao YJ et al (2012b) Identification of progenies from sugarcane × Narenga porphyrocoma (Hance) Bor. by SSR marker. Southwest China J Agric Sci 25:38–43Google Scholar
- Loake G, Grant M (2007) Salicylic acid in plant defence--the players and protagonists. Curr Opin Plant Biol 10:466–472Google Scholar
- Mattiello L, Riaño-Pachón DM, Martins MCM, da Cruz LP, Bassi D et al (2015) Physiological and transcriptional analyses of developmental stages along sugarcane leaf. BMC Plant Bio 15:300Google Scholar
- Paulo C, Sentelhas DL, dos Santos Leonardo A et al (2016) Agro-climatic favorability zones for sugarcane orange rust as a tool for cultivar choice and disease management. Crop Prot 84:88–97Google Scholar
- Roah BT (1972) Nobilization of sugarcane. Proc Int Soc Sugar Cane Technol 14:206–216Google Scholar
- Soderlund C, Bomhoff M, Nelson WM (2011) SyMAP v3.4: a turnkey synteny system with application to plant genomes. Nucl Acids Res 2011:1–9Google Scholar
- Sreenivasan TV, Ahloowalia BS, Heinz DJ (1987) Cytogenetics: in: Heinz DJ (ed) sugarcane improvement through breeding. Elsevier, New York, pp 211–253 CytogeneticsGoogle Scholar
- Su YC, Guo JL, Ling H, Chen SS, Wang SS et al (2016) Isolation of a novel peroxisomal catalase gene from sugarcane, which is esponsive to biotic and abiotic stresses. PLoS One 9(1):e84426Google Scholar
- Trapnell C, Williams BA, Pertea G, Mortazavi AM, Kwan G, van Baren MJ, Salzberg SL, Wold B, Pachter L (2010) Transcript assembly and quantification by RNA-Seq reveals unannotated transcripts and isoform switching during cell differentiation. Nat Biotechnol 28:511–515PubMedPubMedCentralGoogle Scholar
- Verma V, Ravindran P, Kumar PP (2016) Plant hormone-mediated regulation of stress responses. BMC Plant Boi 16:86Google Scholar
- Zhang J, Nagai C, Yu Q, Pan YB, Ayala-Silva T, Schnell RJ, Comstock JC, Arumuganathan AK, Ming R (2012) Genome size variation in three Saccharum species. Euphytica. 185:511–519Google Scholar
- Zhang J, Zhang X, Tang H et al (2018) Allele-defined genome of the autopolyploid sugarcane Saccharum spontaneum L. Nat GenetGoogle Scholar
Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made.