Identification of cellular pathways affected by Sortin2, a synthetic compound that affects protein targeting to the vacuole in Saccharomyces cerevisiae
© Norambuena et al; licensee BioMed Central Ltd. 2008
Received: 30 June 2007
Accepted: 07 January 2008
Published: 07 January 2008
Sortin2 is a low mass compound that interferes with vacuolar delivery of proteins in plants and yeast. The Sortin2 phenotype was tested in Arabidopsis thaliana and found to be reversible upon drug removal, demonstrating the ability of chemical genomics to induce reversible phenotypes that would be difficult to achieve using conventional genetics . However, standard genetic methods can be used to identify drug target pathways in a high-throughput manner.
In this study, we analyzed structure-function relationships of Sortin2 using structural analogues. The results show the key roles of sulphite substitution and a benzoic acid group. A Sortin 2 hypersensitivity screen for the induced secretion of a vacuolar cargo protein was done utilizing a yeast haploid deletion library. Using bioinformatics approaches, we highlighted functional information about the cellular pathways affected by drug treatment which included protein sorting and other endomembrane system-related processes.
Chemical, genomic and genetics approaches were used to understand the mode of action of Sortin2, a bioactive chemical that affects the delivery of a vacuolar protein. Critical features of Sortin2 structure necessary for bioactivity suggest a binding pocket that may recognize two ends of Sortin2. The genome-wide screen shows that Sortin2 treatment in yeast affects primarily components within the endomembrane system. This approach allowed us to assign putative functions in protein sorting for fifteen genes of previously unknown function.
Endomembrane trafficking in eukaryotes is essential for the intracellular delivery of cargoes and membranes. As the site of protein degradation, nutrient recycling and storage of biological components, the vacuole is one of the key compartments of the endomembrane system. In addition, the mechanisms implicated in the delivery of cargoes to the vacuole have been delineated mainly by yeast genetic screens. A relatively new approach for discovering cellular pathways that takes advantage of yeast genetics is chemical genomics, which uses small bioactive molecules that can affect biological pathways [2, 3]. The physiological effects of such drugs may be appreciated and exploited long before the corresponding targets are identified. However, the eventual identification of target proteins within drug-sensitive pathways is necessary to discover and characterize intracellular networks [4–10].
Using forward chemical genomics, we identified fourteen low mass compounds that interfere with the delivery of the vacuolar resident protein carboxypeptidase Y (CPY) in Saccharomyces cerevisiae . These compounds are able to trigger the secretion of CPY, mimicking the vacuole protein sorting (vps) phenotype . Two of these compounds, Sortin1 and Sortin2, result in the secretion of CPY in Arabidopsis thaliana . To understand Sortin2 activity in more detail, we focused on two objectives: to understand the structural features of Sortin2 necessary for its activity and to analyze S. cerevisiae cellular pathways affected by the chemical. The bioactivity of different Sortin2 analogues highlighted structural domains important for Sortin2 activity. In order to identify the cellular pathway(s) affected by Sortin2, a genetic screen was performed for strains hypersensitive to this chemical in a S. cerevisiae haploid deletion collection. Among 4,800 clones screened, 217 strains were hypersensitive to Sortin2. Bioinformatics analysis indicated that the Sortin2 dataset was enriched in components related to protein trafficking and localized primarily to the endomembrane system, particularly endosomes. Our approach permitted an initial functional assignment of fifteen genes of previously unknown function to protein trafficking pathways.
Sortin2 structure-bioactivity relationships
In order to understand the structural determinants of Sortin 2 necessary for bioactivity, a panel of chemicals that were similar in structure to Sortin2 were examined. The Sortin2 contains discrete domains: a chlorobenzene, a furan, a thiazolidine ring, and a sulphite group. The activity of structural analogs as promoters of CPY secretion is shown in Figure 2. For convenience, the potency of Sortin 2 and its analogues was defined as the minimum concentration at which activity was observed.
Chemical 1 (Fig. 2) represented the Sortin2 structure minus the sulphite group. This analogue was inactive indicating that the sulphite group was critical for binding of Sortin2 to its target. Chemical 2 which had a benzoic ring substitution instead of the sulphite group had the same activity as Sortin2. Chemical 3 had a carboxyl group instead of the sulphite group. It was bioactive; however, it had a lower potency (40 μM) than Sortin2 and chemical 2 (5 μM). Taken together, these data suggest that the interaction between Sortin2 and its target probably required a dense electron cloud in order to affect target activity and function. This could be direct or indirect by affecting the overall conformation of Sortin2.
Screening for yeast mutants hypersensitive to Sortin2
Vacuolar protein sorting (vps) mutants and hypersensitivity to Sortin2
Sixty-one yeast deletion strains were known previously to have impaired sorting of CPY; these are referred to as vps mutants. Forty-one vps mutants were identified as drug-dependent in our screen with respect to Sortin2. These included mutants in all six classes of VPS genes (Additional file 1). Class E is involved in protein sorting to lumenal vesicles of the MVB (multi-vesicular body). This family contains 16 members plus VTA1 and HSE1; however, the latter two genes were not described as VPS genes . Remarkably, fourteen out of sixteen members (88.9%) of the VPS Class E were hypersensitive to Sortin2, the class with highest representation among VPS genes. Interestingly, the remaining 20 vps mutants displayed no secretion of CPY when treated with Sortin2. The screen, thus, discriminated between deletion and Sortin2-induced vps phenotypes. This indicated that the Sortin2 phenotype was specific and distinct, rather than a broad effect on secretion, when compared to the majority of the vps mutants. As part of this screen, we analyzed the drug sensitivity of 148 strains identified previously in a screen for mutants that secreted CPY . Among these mutants, 91 were hypersensitive to Sortin2 when compared to the untreated control. The absence of a hypersensitive response from the other 57 mutants indicated that Sortin2 was selective for specific elements of the vacuole targeting machinery.
Analyses of identified ORFs
FunCat  was used to classify the Sortin2 hypersensitive dataset into functional categories relative to the Saccharomyces genome (Table 1, Additional file 2). Some categories showed statistically significant over-representation such as "interaction with the environment" and "biogenesis of cellular components" which were enriched 1.7- and 1.5-fold, respectively (p ≤ 0.04) (Table 1). The greatest enrichment were the categories "protein fate" and "cellular transport, transport facilities and transport routes" which displayed 2.5- and 3-fold enrichment, respectively (p = 1.51 E-19 and 2.99 E-29; Table 1). In terms of subcellular location, the greatest over-representation corresponded to the category "endosomes" with a 17.5-fold enrichment compared to the yeast genome (p = 6.2 E-35) (Table 1). The "Golgi apparatus", "transport vesicles", "vacuole" and "punctuate composite" categories were enriched between 3.5- and 5.4-fold with statistically significant over-representation as shown by the p-values in Table 1. These results revealed that processes within the endomembrane system are strongly and specifically affected by Sortin2 treatment.
Forty-seven percent of the Sortin2 dataset was associated with the GO term "Transport". The set was analyzed by Term Finder tools, and significantly, the sole descendant terms included "Protein Transport", "Intracellular Transport", "Vesicle-mediated Transport" and "Secretory Pathway" highlighting the involvement of the majority of identified ORFs in protein trafficking pathways (data no shown). We also applied the Term Finder tool to analyze the association of the Sortin2 dataset with granular GO terms that were linked to a cellular component (Additional file 3). Within the Sortin2 hypersensitive dataset, we identified members of crucial Golgi and endosomal sorting complexes, their interacting partners and other resident proteins associated all directly in CPY sorting (Additional file 3). In the endosomal compartment, which is the destination for CPY after exiting the late Golgi, we identified members of the three known endosomal and the MVB sorting complexes. Systematic deletion of ESCRT (Endosomal Sorting Complex Required for Transport) genes was reported to cause reduced telomere length . Interestingly, we identified ESCRT-unrelated telomere length mutants as hypersensitive in our screen. In addition several mutants linked to chromatin remodelling were identified (Additional file 3). Overall, we concluded that the Sortin2 hypersensitive ORFs were located preferentially within the endomembrane system and included many molecular components related to the CPY delivery. Our screening results also supported the hypothesis previously proposed that interactivity occurs between the endomembrane system, chromatin remodelling, and telomere maintenance processes.
Functional and locational gene product categorization of Sortin2 hypersensitive ORFs.
Representation on dataset
Enrichment on dataset (-fold)
Cellular transport, transport facilities and transport routes
Protein fate (folding, modification, destination)
Biogenesis of cellular components
Interaction with the environment
Unknown function ORF of the corresponding deletion strains identified as Sortin2 hypersensitive mutants.
Unknown genes categories
CPY secretor deletion strain
Not previously associated to sorting pathway
Other unknown genes were found in high-throughput screens that focused on genetic/physical interactions and protein localization [21–23]. This annotation supports putative roles for the corresponding gene products in protein trafficking. Ygr206wp, for example, was identified as endosomal and more recently as a component of the ESCRT-1 complex [23–25]. YGR206wp and Ylr426wp physically interact with the V-ATPase subunit Vma6p . Their corresponding genes together with YNL080C, SNA2 and TMS1 were the only uncharacterized ORFs from our dataset that overlapped previously unknown genes whose deletion resulted in CPY secretion . Excluding the 10 dubious ORFs described above, the remaining 15 unknown genes have not been associated previously with protein sorting pathways. Among these unknowns, Yil041wp (Gvp36p) was localized to Golgi vesicles . The mutants for the unknown ORFs YDR357C and YGL079W were hypersensitive to Sortin2 and their gene products were identified as interacting partners in a two-hybrid assay . The GFP fusion of Ygl079wp co-localizes with an endosome marker . In addition, Ydr357cp and Yll049wp, both from the Sortin2 dataset, physically interact with the autophagy protein Atg17p  suggesting a role in the endomembrane system. Similarly, YIL039W and YLR361C (DCR2) genetically interact with the Golgi residents Ric1p and Ypt6p , both of which mediate vesicular trafficking at the Golgi apparatus. Tvp38p (Figure 3b) is an integral membrane protein localized to late Golgi vesicles along with the v-SNARE Tlg2p .
To test the interaction of mutants in unknown ORFs with Sortin2, active and inactive analogues were tested for their ability of trigger CPY secretion in the corresponding deletion mutants for YDR525W-A/SNA2, YDR105C/TMS1, YGL079W, YLL049W, YKR088C/TVP38 (data no shown). All of the mutants were more hypersensitive compared to their parental strain when grown in the presence of bioactive compounds 2 and 5. In contrast, none of the strains tested were hypersensitive to the inactive Sortin2 analogues 1, 4 and 7. Therefore, the Sortin2 hypersensitivity of mutants in these genes of unknown function was specific for analogues possessing features important for Sortin2 bioactivity. There were no detectable ALP or PGK proteins in the culture media for any Sortin2 analogs (data no shown) indicating that the CPY secretion was not due to the cell lysis.
In summary, we assigned a putative role in protein trafficking to 15 ORFs of unknown function in S. cerevisiae based on their Sortin2-induced vps phenotypes. The actual role of these ORFs will be addressed in future studies.
In this study, we analyzed the structure-activity relationships and the cellular pathways affected by Sortin2 which triggers the secretion of the soluble vacuolar protein CPY in S. cerevisiae and A. thaliana. Our Sortin2 hypersensitive screen identified genes whose loss of function resulted in no vacuolar sorting defects; however, a sorting defect occurred when these mutants were challenged with the drug. Furthermore, our analysis implicated a number of genes of unknown function in the process of vacuolar protein sorting. Therefore, this approach could be of general use for discovering the functional identity of uncharacterized genes.
The use and analysis of Sortin2 structural analogs allowed us to define the sulphite substitution and benzoic acid group on Sortin2 as key determinants for bioactivity and perhaps target binding. Interaction with a target would probably require both ends of the molecule suggesting a binding pocket that recognizes one face of Sortin2. However, the possibility of multiple binding pockets or targets can not rule out. It is possible to speculate that the dimensions of the binding pocket in the cognate target are very exacting for activity resulting in lower potency or no activity as observed for several Sortin2 analogs. The compounds tested are within the parameters set by Lipinski  indicating that compound permeability would not be an issue.
We screened for enhanced secretion in the presence of Sortin2 using a haploid deletion library. If the drug has a single target, deletion of the gene encoding the target would result in a strain whose phenotype could not be enhanced further by the drug. Thus, our approach would not identify a cognate target. However, we anticipated that this hypersensitive screen would identify proteins that were either in the same pathway or in a pathway connected to the Sortin2 target. The main pathway affected by Sortin2 is the endomembrane system although the effect could be indirect. The intracellular location displaying the greatest enrichment was the endosomes suggesting that the Sortin2 target may be affecting mainly the endosomal compartments which precede the final localization of CPY. Consistent with this idea, almost 90% of the vps Class E mutants, which have an aberrant MVB (the "Class E compartment"), were hypersensitive to Sortin2. Sortin2 is able to mimic the vps phenotype; however, its effect on MVB structure has not been tested yet.
Sortin2 will be useful for probing conserved processes such as endomembrane trafficking in other organisms, and the powerful genomics available will permit the identification of cognate targets.
In this study, we combined chemical, genomic and genetics approaches to understand the mode of action of Sortin2, a bioactive chemical that affects the delivery of a vacuolar protein. Several features of Sortin2 structure are critical for bioactivity suggesting a binding pocket that recognizes one face of the molecule.
The genome-wide screen in yeast showed that Sortin2 mainly affected components within the endomembrane system. Other cellular functions affected by Sortin2 may highlight interactions between cellular processes. Our approach allowed us to assign a putative function in protein sorting for 15 genes of unknown function that were not associated previously with protein trafficking pathways.
Protein secretion assay
The BY4742 haploid parental line (MATalpha his3Δ1 leu2Δ0 lys2Δ0 ura3Δ0) was grown in yeast extract/peptone/dextrose (YPD) medium supplemented with the designated compounds or 1% DMSO as a negative control. Cultures were grown for 72 hours at 30°C. Growth medium was analyzed using monoclonal antibodies against CPY, ALP and 3-phosphoglycerate kinase (Molecular Probes) by dot-blot as described previously .
Sortin2 analogs and substructures
Sortin2 and its analogues and substructures were obtained from Chembridge (San Diego, CA) and referred to by the identification numbers assigned by the manufacturer. Analogue and substructure searches were done using the Hit2Lead database (Chembridge), SciFinder database, and the ChemMine comparative chemical database [29, 30].
Sortin2 Hypersensitivity Screening
The S. cerevisiae haploid deletion library was generated using the BY4742 parental strain containing 4,800 yeast clones (Open Biosystems, Huntsville, AL). The primary screen utilized the microplate format of the deletion library. The library was grown on 2 mg/l (4.7 μM) Sortin2 and 1% DMSO. Growth medium was analyzed for secreted CPY as described previously . The mutant strains identified in the primary screen were further analyzed for CPY secretion at various concentrations of Sortin2 (0.5, 1, 2, 5, and 10 mg/l) in a verification screen. Mutants were considered verified as hypersensitive if significant CPY secretion was detected at 2 mg/l of Sortin2.
The Functional Categorization (FunCat) and Cellular Location (CellLoc) of the identified genes were acquired form The M unich I nformation Center for P roteins S equences . To retrieve information about each deleted ORF, we queried the S accharomyces G enome D atabase (SGD, ). The SGD tools were also used to assign the granular or general Gene Ontology (GO) terms to identified ORFs, using the GO Term Finder or the GO Slim Mapper, respectively.
Vacuole protein sorting
Open reading frame
Multi vesicular body
Endosomal Sorting Complex Required for Transport
We thank Thomas Girke, Marcela Rojas-Pierce and Marci Surpin for critical comments on the manuscript. We acknowledge helpful discussions with members of the Natasha Raikhel's laboratory. This work was supported by a grant from the National Science Foundation (MCB-0515963) to NVR.
- Zouhar J, Hicks GR, Raikhel NV: Sorting inhibitors (Sortins): Chemical compounds to study vacuolar sorting in Arabidopsis. Proc Natl Acad Sci USA. 2004, 101: 9497-9501. 10.1073/pnas.0402121101.PubMed CentralView ArticlePubMed
- Lokey RS: Forward chemical genetics: progress and obstacles on the path to a new pharmacopoeia. Current Opinion in Chemical Biology. 2003, 7: 91-96. 10.1016/S1367-5931(02)00002-9.View ArticlePubMed
- Schreiber SL: Small molecules: the missing link in the central dogma. Nat Chem Biol. 2005, 1: 64-66. 10.1038/nchembio0705-64.View ArticlePubMed
- Peyroche A, Antonny B, Robineau S, Acker J, Cherfils J, Jackson CL: Brefeldin A Acts to Stabilize an Abortive ARF-GDP-Sec7 Domain Protein Complex: Involvement of Specific Residues of the Sec7 Domain. Molecular Cell. 1999, 3: 275-285. 10.1016/S1097-2765(00)80455-4.View ArticlePubMed
- Lum PY, Armour CD, Stepaniants SB, Cavet G, Wolf MK, Butler JS, Hinshaw JC, Garnier P, Prestwich GD, Leonardson A: Discovering Modes of Action for Therapeutic Compounds Using a Genome-Wide Screen of Yeast Heterozygotes. Cell. 2004, 116: 121-137. 10.1016/S0092-8674(03)01035-3.View ArticlePubMed
- Giaever G, Flaherty P, Kumm J, Proctor M, Nislow C, Jaramillo DF, Chu AM, Jordan MI, Arkin AP, Davis RW: Chemogenomic profiling: Identifying the functional interactions of small molecules in yeast. Proc Natl Acad Sci USA. 2004, 101: 793-798. 10.1073/pnas.0307490100.PubMed CentralView ArticlePubMed
- Cheng Y, Dai X, Zhao Y: AtCAND1, a HEAT-repeat protein that participates in auxin signaling in Arabidopsis. Plant Physiol. 2004, 135: 1020-1026. 10.1104/pp.104.044495.PubMed CentralView ArticlePubMed
- Butcher RA, Schreiber SL: Identification of Ald6p as the target of a class of small-molecule suppressors of FK506 and their use in network dissection. Proc Natl Acad Sci USA. 2004, 101: 7868-7873. 10.1073/pnas.0402317101.PubMed CentralView ArticlePubMed
- Stearns T, Hoyt MA, Botstein D: Yeast Mutants Sensitive to Antimicrotubule Drugs Define Three Genes That Affect Microtubule Function. Genetics. 1990, 124: 251-262.PubMed CentralPubMed
- Chan T-F, Carvalho J, Riles L, Zheng XFS: A chemical genomics approach toward understanding the global functions of the target of rapamycin protein (TOR). Proc Natl Acad Sci USA. 2000, 97: 13227-13232. 10.1073/pnas.240444197.PubMed CentralView ArticlePubMed
- Bowers K, Stevens TH: Protein transport from the late Golgi to the vacuole in the yeast Saccharomyces cerevisiae. Biochimica et Biophysica Acta (BBA). 2005, 1744: 438-454. 10.1016/j.bbamcr.2005.04.004.View Article
- Lenney JF, Matile P, Wiemken A, Schellenberg M, Meyer J: Activities and cellular localization of yeast proteases and their inhibitors. Biochemical and Biophysical Research Communications. 1974, 60: 1378-1383. 10.1016/0006-291X(74)90350-7.View ArticlePubMed
- Schwaiger H, Hasilik A, von Figura K, Wiemken A, Tanner W: Carbohydrate-free carboxypeptidase Y is transferred into the lysosome-like yeast vacuole. Biochemical and Biophysical Research Communications. 1982, 104: 950-956. 10.1016/0006-291X(82)91341-9.View ArticlePubMed
- Vida TA, Huyer G, Emr SD: Yeast vacuolar proenzymes are sorted in the late Golgi complex and transported to the vacuole via a prevacuolar endosome-like compartment. J Cell Biol. 1993, 121: 1245-1256. 10.1083/jcb.121.6.1245.View ArticlePubMed
- Bonangelino CJ, Chavez EM, Bonifacino JS: Genomic Screen for Vacuolar Protein Sorting Genes in Saccharomyces cerevisiae. Mol Biol Cell. 2002, 13: 2486-2501. 10.1091/mbc.02-01-0005.PubMed CentralView ArticlePubMed
- Ruepp A, Zollner A, Maier D, Albermann K, Hani J, Mokrejs M, Tetko I, Guldener U, Mannhaupt G, Munsterkotter M, Mewes HW: The FunCat, a functional annotation scheme for systematic classification of proteins from whole genomes. Nucl Acids Res. 2004, 32: 5539-5545. 10.1093/nar/gkh894.PubMed CentralView ArticlePubMed
- Askree SH, Yehuda T, Smolikov S, Gurevich R, Hawk J, Coker C, Krauskopf A, Kupiec M, McEachern MJ: From the Cover: A genome-wide screenfor Saccharomyces cerevisiae deletion mutants that affect telomere length. Proc Natl Acad Sci USA. 2004, 101: 8658-8663. 10.1073/pnas.0401263101.PubMed CentralView ArticlePubMed
- Parsons AB, Brost RL, Ding H, Li Z, Zhang C, Sheikh B, Brown GW, Kane PM, Hughes TR, Boone C: Integration of chemical-genetic and genetic interaction data links bioactive compounds to cellular target pathways. Nat Biotech. 2004, 22: 62-69. 10.1038/nbt919.View Article
- Calero M, Winand NJ, Collins RN: Identification of the novel proteins Yip4p and Yip5p as Rab GTPase interacting factors. FEBS Letters. 2002, 515: 89-98. 10.1016/S0014-5793(02)02442-0.View ArticlePubMed
- Geissler SSK, Schiebel E: A novel protein complex promoting formation of functional alpha- and gamma-tubulin. Embo J. 1998, 17: 952-966. 10.1093/emboj/17.4.952.PubMed CentralView ArticlePubMed
- Ho Y, Gruhler A, Heilbut A, Bader GD, Moore L, Adams S-L, Millar A, Taylor P, Bennett K, Boutilier K: Systematic identification of protein complexes in Saccharomyces cerevisiae by mass spectrometry. Nature. 2002, 415: 180-183. 10.1038/415180a.View ArticlePubMed
- Ito T, Chiba T, Ozawa R, Yoshida M, Hattori M, Sakaki Y: A comprehensive two-hybrid analysis to explore the yeast protein interactome. Proc Natl Acad Sci USA. 2001, 98: 4569-4574. 10.1073/pnas.061034498.PubMed CentralView ArticlePubMed
- Huh W-K, Falvo JV, Gerke LC, Carroll AS, Howson RW, Weissman JS, O'Shea EK: Global analysis of protein localization in budding yeast. Nature. 2003, 425: 686-691. 10.1038/nature02026.View ArticlePubMed
- Curtiss M, Jones C, Babst M: Efficient Cargo Sorting by ESCRT-I and the Subsequent Release of ESCRT-I from Multivesicular Bodies Requires the Subunit Mvb12. Mol Biol Cell. 2007, 18: 636-645. 10.1091/mbc.E06-07-0588.PubMed CentralView ArticlePubMed
- Oestreich AJ, Davies BA, Payne JA, Katzmann DJ: Mvb12 Is a Novel Member of ESCRT-I Involved in Cargo Selection by the Multivesicular Body Pathway. Mol Biol Cell. 2007, 18: 646-657. 10.1091/mbc.E06-07-0601.PubMed CentralView ArticlePubMed
- Inadome H, Noda Y, Adachi H, Yoda K: Immunoisolaton of the Yeast Golgi Subcompartments and Characterization of a Novel Membrane Protein, Svp26, Discovered in the Sed5-Containing Compartments. Mol Cell Biol. 2005, 25: 7696-7710. 10.1128/MCB.25.17.7696-7710.2005.PubMed CentralView ArticlePubMed
- Tong AHY, Lesage G, Bader GD, Ding H, Xu H, Xin X, Young J, Berriz GF, Brost RL, Chang M: Global Mapping of the Yeast Genetic Interaction Network. Science. 2004, 303: 808-813. 10.1126/science.1091317.View ArticlePubMed
- Lipinski CA, Lombardo F, Dominy BW, Feeney PJ: Experimental and computational approaches to estimate solubility and permeability in drug discovery and development settings. Advanced Drug Delivery Reviews. 1997, 23: 3-25. 10.1016/S0169-409X(96)00423-1.View Article
- Girke T, Cheng L-C, Raikhel N: ChemMine. A Compound Mining Database for Chemical Genomics. Plant Physiol. 2005, 138: 573-577. 10.1104/pp.105.062687.PubMed CentralView ArticlePubMed
- ChemMine. [http://bioweb.ucr.edu/ChemMineV2/]
- Munich Information Center for Proteins Sequences. [http://mips.gsf.de]
- Saccharomyces Genome Database. [http://www.yeastgenome.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.