Please download to get full document.

View again

of 10

A modular yeast biosensor for low-cost point-of-care pathogen detection

The availability of simple, specific, and inexpensive on-site detection methods is of key importance for deployment of pathogen surveillance networks. We developed a nontechnical and highly specific colorimetric assay for detection of
8 views10 pages
All materials on our website are shared by users. If you have any questions about copyright issues, please report us to resolve them. We are always happy to assist you.
Documenttranscript
  HEALTH AND MEDICINE  Copyright © 2017 The Authors, somerights reserved;exclusive licenseeAmerican Associationfor the Advancementof Science. No claim tosrcinalU.S.GovernmentWorks. Distributedunder a CreativeCommons AttributionLicense 4.0 (CC BY). A modular yeast biosensor for low-cost point-of-carepathogen detection Nili Ostrov, 1 * † Miguel Jimenez, 1 * ‡ Sonja Billerbeck, 1 * James Brisbois, 1  Joseph Matragrano, 1 Alastair Ager, 2,3 Virginia W. Cornish 1,4§ The availability of simple, specific, and inexpensive on-site detection methods is of key importance for deploy-ment of pathogen surveillance networks. We developed a nontechnical and highly specific colorimetric assayfor detection of pathogen-derived peptides based on  Saccharomyces cerevisiae — a genetically tractable modelorganism and household product. Integrating G protein – coupled receptors with a visible, reagent-free lycopenereadout, we demonstrate differential detection of major human, plant, and food fungal pathogens with nano-molar sensitivity. We further optimized a one-step rapid dipstick prototype that can be used in complexsamples, including blood, urine, and soil. This modular biosensor can be economically produced at large scale,is not reliant on cold-chain storage, can be detected without additional equipment, and is thus a compellingplatform scalable to global surveillance of pathogens. INTRODUCTION Global surveillance of pathogens is critical for human health, food se-curity, bioterrorism, and maintenance of biodiversity ( 1 ,  2 ). Althoughmonitoring of global pathogen burden has been traditionally limited toasmallnumberof specializedcenters,more effective detection couldbeperformed in real time by making accurate diagnostics accessible at thepointofcare( 3 ).However,thesetestsareonlyavailableforahandfulof pathogens and often rely on costly reagents, cold-chain distribution,specialized equipment, and technical personnel ( 4 ,  5 ). In the absenceofinnovativealternativestoexpensiveantibody-andnucleicacid – basedassays,lengthyculturingmethodsremainasthedominantapproachformonitoring pathogens in resource-limited areas ( 3 ).The emerging field of synthetic biology has the potential to providenovel diagnostic platforms to overcome global health challenges ( 6  ,  7  ),muchlike advancesin molecularbiology gave rise to antibody diagnos-tics. Although synthetic biology has thus far been leveraged primarily foreconomical fermentation of industrial and biomedical commodities via metabolic engineering ( 8 ), significant applications in public health,animal health, and agriculture remain untapped.Here,wedemonstratethathighlyspecialized Saccharomycescerevisiae cells can be generated to create transformative on-site diagnostic de- vices unattainable via traditional engineering methods. Of the growing number of engineered biological systems used for biosensing ( 7  ,  9  – 11 ), S. cerevisiae  stands out because of its marked robustness, genetic trac-tability, and established human safety ( 12 – 14 ). Capitalizing on the or-thogonality and specificity of its G protein – coupled receptor (GPCR)mating pathway ( 15 ), it has also shown potential as a platform for drug discovery by engineering mammalian GPCRs ( 16  ,  17  ).To overcome current bottlenecks in deployment of biosensors, wedeveloped a modular, single-component biosensor that uses fungalmating GPCRs for detection of pathogen-specific peptides and a redplant pigment as readout (Fig. 1). Expanding on well-establishedmethods for scalable and cost-effective distribution of   S. cerevisiae , weenvision that this portable, safe biosensor can enable routine on-sitepathogen surveillance.We demonstrate the utility of our platform for detection of fungalpathogens, a rising global public health burden particularly acute indeveloping countries ( 2 ,  18 ,  19  ). Fungal pathogens cause an estimated2 million deaths annually and inflict devastating losses of plant cropsand population decline in animal wildlife ( 2 ,  19  ). However, efforts toabate fungal infections prevalent in resource-limited areas are ham-pered by the dearth of fungal diagnostics ( 20 ). We thus used the largereservoir of natural fungal GPCRs to construct  S. cerevisiae – based bio-sensors for detection of fungi responsible for major human disease, ag-ricultural damage, and food spoilage. RESULTS S. cerevisiae  as a biosensor visible to the naked eye We started by building a prototype biosensor for detection of   Candidaalbicans , a commensal human pathogen and the leading cause of life-threateninginfectionsinimmunocompromisedpatients( 18 ). C.albicans is among the most genetically tractable pathogenic fungi, with a well-studiedmatingsystem( 21 ).Forspecificdetectionofthe C.albicans mating peptide, we replaced the  S. cerevisiae  mating receptor (Sc.Ste2) withthatof  C.albicans (Ca.Ste2)(tablesS1andS2).Wethenusedsynthetic C. albicans  mating pheromone to test receptor activation using a fluo-rescent reporter that is induced by the mating signaling pathway ( 17  ).Consistentwithpreviousreports( 22 ),wefoundCa.Ste2tobehighly sensitivetoitscognateligand,withanEC 50  value(concentrationofpep-tiderequiredforhalf-maximalactivation)of~4nM(Fig.2A).Toprobereceptor specificity further, we also tested Ca.Ste2 activation using nineheterologousfungalmatingpeptidesandfoundittobehighlyspecifictoits mating peptide (Fig. 2B).Tomeettheneedsofon-sitedetection,wethenreplacedthefluores-cent readout with a robust signal that is easily visible to the naked eye.We chose lycopene, an antioxidant carotenoid pigment naturally producedinplantsandbacteriaandwidelyusedformetabolicengineer-ing ( 23 ). We introduced the lycopene production pathway into ourstrain byplacingthe firsttwobiosyntheticgenes,  Erwinia herbicola crtE  1 Department of Chemistry, Columbia University, New York, NY 10027, USA.  2 Depart-ment of Population and Family Health, Mailman School of Public Health, ColumbiaUniversity, New York, NY 10032, USA.  3 Institute for Global Health and Development,QueenMargaretUniversity,Edinburgh,UK. 4 DepartmentofSystemsBiology,ColumbiaUniversity, New York, NY 10032, USA.*These authors contributed equally to this work. † Present address: Department of Genetics, Harvard Medical School, Boston, MA02115, USA. ‡ Present address: Koch Institute for Integrative Cancer Research, MassachusettsInstitute of Technology, Cambridge, MA 02139, USA.§Corresponding author. Email: vc114@columbia.edu SCIENCE ADVANCES | RESEARCH ARTICLE Ostrov  et al  .,  Sci. Adv.  2017; 3 :e1603221 28 June 2017  1 of 9   on J  ul   y 1  ,2  0 1 7 h  t   t   p:  /   /   a d v  an c  e s . s  c i   en c  em a g. or  g /  D  ownl   o a d  e d f  r  om   Fig. 1.  S. cerevisiae  biosensor for detection of fungal pathogens.  ( A ) Overview of biosensor components. Highly specific fungal receptors provide sensitive response tomatingpeptidessecretedbypathogenicfungi.Activationofthedownstreammatingsignalingpathwayinducestranscriptionalactivationofbiosyntheticgenesforproductionof redlycopenepigmentvisibletothenakedeye.FMN,flavinmononucleotide;FAD,flavinadeninedinucleotide;FPP,farnesylpyrophosphate;GGPP,geranylgeranylpyrophosphate.( B ) Color signal as shown in paper-based dipstick assay. Scale bars, 0.5 cm. Fig. 2. Biosensorfunctionalityandlycopeneoptimization.  ( A )Activationof  C. albicans matingreceptor(Ca.Ste2)in S. cerevisiae byitscognatematingpeptide.Fluorescence(black)andlycopeneabsorbance(red)wereusedasatranscriptionalreadoutforreceptoractivation.( B )SpecificityofCa.Ste2andSc.Ste2receptors.Fluorescencewasdeterminedafter9hoursusing5 m Msyntheticfungalpeptides.( C )Optimizationoflycopeneproduction.Maximallycopeneyieldwasmeasuredafterinductionwith10 m Msynthetic S.cerevisiae mating peptide. Null, parental strain (no lycopene genes); Lyco-1, parental strain with single-copy CrtE, CrtB, and CrtI; 2xCrtI, Lyco-1 with additional plasmid-borne copy of CrtI;Fad1, Lyco-1 with additional plasmid-borne copy of Fad1; Lyco-2, all genes genomically integrated into Lyco-1. ( D ) Time course of lycopene production in lycopene-producingstrains.Inductionasin(C).( E )Representativephotosofcellpellets(5×10 7 cells)correspondingtostrainsin(D).Lycopenepercellwasdeterminedbyspectroscopy(see Supplementary Methods). SCIENCE ADVANCES | RESEARCH ARTICLE Ostrov  et al  .,  Sci. Adv.  2017; 3 :e1603221 28 June 2017  2 of 9   on J  ul   y 1  ,2  0 1 7 h  t   t   p:  /   /   a d v  an c  e s . s  c i   en c  em a g. or  g /  D  ownl   o a d  e d f  r  om   (geranylgeranyl diphosphate synthase) and  crtB  (phytoene synthase),under constitutive promoters from  ADH1  and  TEF1 , respectively.To tie lycopene production to activation of the mating receptor, weplaced the last pathway gene  crtI   (lycopene synthase) under controlof the pheromone-inducible promoter from  FUS1  (Fig. 1 and fig. S1).To circumvent the low-throughput chemical extraction that is tra-ditionally used for measurement of lycopene yield ( 23 ), we developed aquantitativeabsorbance-basedmethodtomeasurelycopeneproductionper cell (LPC) (Supplementary Methods). We determined ≥ 3.5 LPCunits to be a stringent threshold for a robust yes/no lycopene readout visible to the naked eye.Lycopene production in this initial strain reached a maximum of 3 LPC units, just below the visible threshold (Fig. 2, C to E). We thusoptimized lycopene production by adding another copy of the en-dogenous flavin adenine dinucleotide synthase ( FAD1 ) ( 24 ) and ly-copenesynthase( crtI  )(fig.S1).Theenhancedstrainexhibitedstrong lycopene production of more than 10 LPC units, crossing the visiblelycopene threshold in only 3 hours (Fig. 2, D and E).Using the lycopene readout, we measured the limit of detection(LoD) of our  C. albicans  biosensor in liquid culture to be 1 to 10 nMmating peptide, with lycopene sensitivity closely matching that of thefluorescent reporter (Fig. 2A). The LoD and strong lycopene readoutremained stable across pH and temperature and in clinically relevantsamplessuchasurineandserum(fig.S2). Together,these resultsestab-lish the feasibility of using   S. cerevisiae  as a specific and sensitive detec-tor, competitive with culture-based diagnostic methods. Fungal GPCRs as modular detection elements for invasivefungal pathogens Next,wesoughttoexpandtherangeoftargetpathogens.However,con-trary to  C. albicans , the mating systems for most fungal pathogens arenot well studied ( 25 ). To identify candidate mating receptors, wesearched the large reservoir of available fungal genomes for receptorsand putative mating peptides homologous to  S. cerevisiae STE2  and  MF  a 1 , respectively (table S1 and fig. S3A) ( 26  ). We introduced codon-optimized or wild-type receptor genes into an  S. cerevisiae  straincontaining a fluorescent reporter (yMJ183) and measured receptor ac-tivationusing the cognate fungal matingpeptides(Fig. 3A,fig.S4,andtable S2).By simply swapping the Ste2 receptor, we generated functional bio-sensors for 10 major human, agricultural, and food spoilage pathogens: Candida albicans ,  Candida glabrata ,  Paracoccidioides brasiliensis , Histoplasma capsulatum ,  Lodderomyces elongisporus ,  Botrytis cinerea , Fusarium graminearum ,  Magnaporthe oryzae ,  Zygosaccharomyces bailii ,and  Zygosaccharomyces rouxii . Notably, no additional engineering wasnecessary for activation of these heterologous receptors, as is often re-quired for mammalian GPCRs, possibly due to higher homology withthe endogenous receptor (fig. S3).All receptors exhibited high sensitivity to their cognate ligands, withEC 50  values ranging from 14  m M to 4 nM (Fig. 3B and fig. S4). Codonoptimization of the receptor gene provided a slight increase in sensitiv-ity (fig. S4B).Most tested receptors were also highly specific to their cognate mat-ingpeptide,asexpectedformediatorsofspecies-specificsignals(Fig.3Candfig.S5).Notably,ourresultsshowdifferentialdetectionof  C.albicans , C.glabrata , L. elongisporus ,and P.brasiliensis ,majorhumanpathogenswith distinct susceptibility to antifungal drugs ( 27  ).Wethenturnedtoimplementingourlycopenereadoutforthesenew fungal targets. Comparison of the fluorescent activation profile of thesebiosensors with that of our initial  C. albicans  biosensor suggested thatthe new set of receptors could be readily implemented with the opti-mized lycopene readout strain to give a strong readout visible to thenakedeye(Figs.2Aand3A).Wechosethebiosensorfor P.brasiliensis for further characterization. As with the  C. albicans  biosensor, we ob-served a  P. brasiliensis  pheromone LoD of <0.01 to 10 nM and a 3.9 ±0.1 fold activation of lycopene production, yielding a robust readoutacross pH and temperature and in urine and serum (fig. S6). Detection of fungal pathogen clinical isolates Next, we challenged our biosensor for detection of naturally secretedmating peptides using clinically isolated  Paracoccidioides  strains. Para-coccidioidomycosis (PCM), an invasive fungal infection endemic toLatinAmerica,isoneofmanyneglectedtropicaldiseasesthatprimarily affect poor populations and lack systematic surveillance ( 28 ). PCM iscaused by inhalation of airborne conidia produced by mycelium of the soil ascomycete  P. brasiliensis  ( 29  ). Recent identification of the ge-netic components underlying its mating system ( 30 ) enabled us to pur-suespecificyeast-baseddetectionof  P.brasiliensis ,whichcouldfacilitatedetection of its environmental reservoir.Specifically, we challenged our biosensor to detect cultured mycelial P. brasiliensis  isolated from human patients (see Materials andMethods). Biosensor cells expressing   P. brasiliensis  mating receptor,which exhibited specific and sensitive detection of its synthetic mating peptide (figs. S5A and S6), were mixed with spent supernatants fromtwo clinically isolated  Paracoccidioides  strains (table S2). In response,we observed lycopene production well above the visible threshold(Fig.3E).Secretedmatingpeptidesweresimilarlydetectedfromclinicalisolates of   C. albicans  and  H. capsulatum  (Fig. 3E). The peptideproduced by   H. capsulatum  ( 30 ), the causative agent of histoplasmosis( 18 ), is identical to that of   P. brasiliensis  and could be detected using both biosensor strains (fig. S7). A low-cost, low-tech dipstick assay Encouraged by these results, we then focused our efforts towardtranslation of our assay to the field by developing a simple dipstick test.For this aim, biosensor and control cells were spotted onto filter paper,and detection was performed by simply dipping the paper into liquidsamples containing synthetic mating peptides (Fig. 4A and fig. S8). Inaddition to visual inspection, we quantified lycopene accumulation onpaper using pixel color analysis (Supplementary Methods).Usinga P.brasiliensis dipstickassay,weobservedarobustandhighly reproduciblesignalthatsurpassedthevisiblelycopenethresholdtogivea clear yes/no readout (Fig. 4B and movie S1). Similar results wereachieved using a  C. albicans  dipstick assay (Fig. 4C and fig. S9). Asexpected,nocross-reactivitywasobservedbetweenthesetwopathogens(movie S1). Last, to ensure that the signal remains visible in complex samples, we performed dipstick tests in soil, urine, serum, and bloodsupplemented with synthetic mating peptides. In all sample types, mi-cromolar levels of peptide were successfully detected (Fig. 4D, fig. S9,and movies S2 to S5). The dipstick assay retained its functionality afterbeing stored for 38 weeks at room temperature (fig. S10). DISCUSSION Insummary,ourresultsshowthefeasibilityandrobustnessofpathogendetection using engineered  S. cerevisiae . We achieved sensitivity andspecificitylevelscomparabletothoseofmammalian,whole-cell,antibody,and nucleic acid assays, all of which are significantly more expensive SCIENCE ADVANCES | RESEARCH ARTICLE Ostrov  et al  .,  Sci. Adv.  2017; 3 :e1603221 28 June 2017  3 of 9   on J  ul   y 1  ,2  0 1 7 h  t   t   p:  /   /   a d v  an c  e s . s  c i   en c  em a g. or  g /  D  ownl   o a d  e d f  r  om   and technically demanding. Our assay can be constructed cheaply on alarge scale by leveraging existing infrastructure for yeast culture, widely distributedasastabledriedproductforhouseholduse,robustlyappliedto complex samples, and readily detected by eye.Although the lycopene readout already provides a robust yes/no re-sponse, that response can be further enhanced through promoter opti-mization, signal amplification feedback loops, alternate lycopenebiosynthetic enzymes, or engineering of receptor/G protein interaction.Additionally,althoughtheGPCRsalreadyshowmarkedspecificity,totalassay specificity can be increased by incorporating additional positiveandnegativebiosensorspotsforothertargetandoff-targetpheromonesor by engineering the pheromone specificity of individual receptors.Fungal pheromones represent a promising class of species-specificbiomarkers that may be used to overcome the time-consuming methods in place today for the definitive diagnosis of fungal pathogens( 3 ,  18 ). The biosensors developed in this work will simplify the valida-tion of these peptide biomarkers, which remain poorly characterized inclinical and environmental samples ( 31 ). While at an early stage of im-plementation,thesebiosensorscanbeimmediatelyadoptedintheclinicto shorten the time required for diagnosis of fungal pathogens fromblood cultures (see Supplementary Methods). Furthermore, pathogencharacterization can be expanded to other fungi by adding new GPCR-pheromone pairs, identified through fungal genome mining,to our biosensor platform. Fig. 3. Yeast biosensor for multiple fungal peptides.  ( A ) Activation of fungal mating receptors in  S. cerevisiae  by the corresponding cognate synthetic matingpeptides (40  m M) (see also fig. S4). Dotted line denotes the effective visible threshold from Fig. 2A. ( B ) EC 50  values calculated for fungal receptors in  S. cerevisiae  usingcognate ligands. ( C ) Specificity of heterologous fungal receptors. Receptors were activated by 5  m M synthetic peptides. Response was measured by fluorescence andnormalized for each receptor (see fig. S5). ( D ) Comparative scoring of all biosensors. ( E ) Lycopene production induced by culture supernatant from clinically isolatedfungal pathogens. Lycopene per cell measured by spectroscopy at 9 hours (see Supplementary Methods). ** P   ≤  0.01, *** P   ≤  0.001;  n  = 3. SCIENCE ADVANCES | RESEARCH ARTICLE Ostrov  et al  .,  Sci. Adv.  2017; 3 :e1603221 28 June 2017  4 of 9   on J  ul   y 1  ,2  0 1 7 h  t   t   p:  /   /   a d v  an c  e s . s  c i   en c  em a g. or  g /  D  ownl   o a d  e d f  r  om   The range of biosensor targets can be further extended beyond fun-gal pheromones by leveraging established GPCR-directed evolutionmethodsinyeast( 16  , 17  ).FungalGPCRs,aswellasotherGPCRclasses,can be engineered to recognize novel peptide, protein, polysaccharide,andsmall-moleculeligandsaspotentialtargetsforyeast-baseddetectionof bacteria, viruses, toxins, and other diseases.Although early success has established the impact of synthetic bi-ologyinthefieldofmetabolicengineering,theapplicationofsyntheticbiology to grand challenges in biomedicine, materials science, andother fields is still at an early stage of development. Here, we show that yeast engineeredwith the growing toolbox of synthetic biology canhave a major impact on the field of biosensors. While sophisticated,highly technical diagnostics remain limited to clinical and industriallaboratories, reliable low-tech tools accessible to the general populationcan provide disruptive cheap alternatives to broadly transform the fieldof diagnostics. MATERIALS AND METHODS Materials Synthetic mating peptides ( ≥ 95% purity) were obtained from GenScriptor Zymo Research. Polymerases, restriction enzymes, and GibsonassemblymixwereobtainedfromNewEnglandBiolabs.Culturemediumcomponents were obtained from BD Bioscience and Sigma-Aldrich.PrimersandsyntheticDNAwerepurchasedfromIntegratedDNATech-nologies(IDT).Plasmidswereclonedandamplifiedineither Escherichiacoli  TG1 or C3040 (New England Biolabs). Human urine (catalog no.IR100007P) and single donor human whole blood (catalog no. IPLA-WB1) were purchased from Innovative Research. Human serum,normal off the clot (frozen) (catalog no. HSER-2ML) was purchasedfromZen-Bio.ProfessionalpottingmixsoilwaspurchasedfromFafard. S. cerevisiae  general cloning methods All strains were derived from parental  reiterative recombination  ac-ceptor strain LW2591 (  MATa-inc   genotype), and cloning of expres-sion modules into the  HO  locus was performed using   reiterativerecombination  ( 32 ). Scarless gene deletions and gene replacementswere carried out using   delitto perfetto  ( 33 ). Endogenous yeast promo-ters, terminators, and open reading frames (ORFs) were obtained by polymerase chain reaction (PCR) from strain FY251 [American TypeCulture Collection (ATCC) 96098] or LW2591 ( 32 ). Yeast transfor-mations were carried out using the lithium acetate method ( 34 ). Allplasmids are derivatives of the pRS41x series of centromeric shuttleplasmids, and were cloned using standard molecular biology protocolsand Gibson assembly ( 35 ). Yeast strains used in this study are listed intable S2. Plasmids used in this study are listed in table S3. Expression Fig. 4. Paper-baseddipstickassayfordetectionoffungalpathogens. ( A )Dipstickdevice.Inset: “ +, ” positivebiosensorstrain; “ − , ” negativecontrolstrain.( B )Quantitativeanalysisof lycopeneproductionusingdipstickassay,asscoredbytime-lapsephotography(seeSupplementaryMethods)fordetectionof1 m Msynthetic P.brasiliensis matingpeptide.Individualrunsareshownin light color,andaverageresponseis shown indark color.Shading indicatesvisiblethreshold.( C ) P. brasiliensis and C. albicans matingpeptideswerereproducibly detectedusingthedipstickassay.Maximalresponsewasachievedby12hoursafterexposuretotherespectivepeptides(1 m M).( D )Detectionof  P.brasiliensis matingpeptideincomplexsamples.Liquidsamplesweresupplementedwithsynthetic P.brasiliensis matingpeptide(blue)orwater(gray),andscoredasin(B).YPD,mediumonly;soil,standardpottingsoil;urine,50%pooledhuman urine; serum, 50% human serum; blood, 2% whole blood. All experiments were performed using 1  m M peptide and supplemented with YPD medium. AU, arbitrary units. SCIENCE ADVANCES | RESEARCH ARTICLE Ostrov  et al  .,  Sci. Adv.  2017; 3 :e1603221 28 June 2017  5 of 9   on J  ul   y 1  ,2  0 1 7 h  t   t   p:  /   /   a d v  an c  e s . s  c i   en c  em a g. or  g /  D  ownl   o a d  e d f  r  om 
Advertisement
MostRelated
View more
We Need Your Support
Thank you for visiting our website and your interest in our free products and services. We are nonprofit website to share and download documents. To the running of this website, we need your help to support us.

Thanks to everyone for your continued support.

No, Thanks
SAVE OUR EARTH

We need your sign to support Project to invent "SMART AND CONTROLLABLE REFLECTIVE BALLOONS" to cover the Sun and Save Our Earth.

More details...

Sign Now!

We are very appreciated for your Prompt Action!

x