Incorporating microorganisms into polymer layersprovides bioinspired functional living materialsLukas C. Gerber, Fabian M. Koehler, Robert N. Grass, and Wendelin J. Stark1

Institute for Chemical and Bioengineering, Department of Chemistry and Applied Biosciences, Eidgenössiche Technische Hochschule Zurich, 8093 Zurich,Switzerland

Edited by Gabor A. Somorjai, UC Berkeley, Berkeley, CA, and approved November 10, 2011 (received for review September 23, 2011)

Artificial two-dimensional biological habitats were prepared fromporous polymer layers and inoculated with the fungus Penicilliumroqueforti to provide a living material. Such composites of classicalindustrial ingredients and living microorganisms can provide anovel form of functional or smart materials with capability forevolutionary adaptation. This allows realization of most complexresponses to environmental stimuli. As a conceptual design, weprepared a material surface with self-cleaning capability whensubjected to standardized food spill. Fungal growth and reproduc-tion were observed in between two specifically adapted polymerlayers. Gas exchange for breathing and transport of nutrientthrough a nano-porous top layer allowed selective intake of foodwhilst limiting the microorganism to dwell exclusively in betweena confined, well-enclosed area of the material. We demonstrateda design of such living materials and showed both active (eating)andwaiting (dormant, hibernation) states with additional recoveryfor reinitiation of a new active state by observing the metabolicactivity over two full nutrition cycles of the living material (active,hibernation, reactivation). This novel class of living materials canbe expected to provide nonclassical solutions in consumer goodssuch as packaging, indoor surfaces, and in biotechnology.

adaptable ∣ artificial biological niche ∣ biomimetics ∣functional surface ∣ self-cleaning

Living surfaces occur in all organism’s outer shell, like tree bark(1, 2) but also as lichen on rocks (3, 4) and have long been used

in food preservation (5, 6). Living surfaces show most complexresponses to a variety of external stimuli and thereby providean attractive target to the development of functional materials:The fluffy white surface of the French gourmet cheese Camem-bert, for example, is a living, functional biomaterial with all prop-erties of a classical biotechnological process: Inoculation of theraw cheese’s surface with the edible mold Penicillium camembertileads to its exquisite aroma and taste [i.e. a biotechnologicaltransformation (7–9)]. It further prevents the contamination withother microorganisms (i.e. defense) that may spoil the cheese(9, 10) and trigger diseases (11), while the cheese supplies nu-trients to the mold (selective transport). At present, industrialbiotechnology predominantly uses fungi to produce valuableproducts (12) or to delignify wood (13) in large tank reactors.

Transferring the complex response of microorganisms intoconsumer goods would require the use of microorganism as amaterial component (Fig. 1). An example of a complex behavioris active defense of a food source like in cheese rind (10) or drymeat (9, 14) through a specific mold. Inspired by this concept, wetherefore designed surface structures combining inert (classicalpolymers) and living parts (biological organisms).

As a model organism, we choose Penicillium roqueforti con-fined to a planar habitat between a polymer support foil (mechan-ical integrity) and a nano-porous top layer that allowed nutrienttransport and gas exchange (O2 uptake and CO2 release). Thelimited pore size did not allow the fungal hyphae to grow out ofits predefined habitat, yet enabled exchange of small molecules(diffusion). As a conceptual demonstration, we showed that theseliving layers could metabolize model food spills and withstand

harsh handling and prolonged periods of severe stress. Livingorganisms’ capability to adapt to changing environments distin-guishes them from classical anthropogenic materials—we demon-strated adaptation by showing hibernation (a food shortageinduced waiting state of fungi; some fungal spores can outlivemillions of years [15)] and reinitiation of growth and proliferationupon renewed exposure to humidity and nutrients.

ResultsThe design of a living material demands three functional layers:base, living, and cover layer (Fig. 2A and 2B). The base layer islinked to the living layer and acts as an independent support, orlinks an existing substrate with the living layer as used in classicalpolymer science. The living layer is defined as the habitat of theliving fungi. Within this part the fungi are maintained and allowedto grow. The cover layer protects the living layer from environ-mental influences (mechanical, chemical, biological) and viceversa. The cover layer also acts as the interconnection betweenthe living layer and the surrounding world. Here we used poly-vinyl chloride as a base layer, a living layer based on agar, and aporous polycarbonate membrane as a cover layer to allow the dif-fusion of gases and nutrients necessary for the growth. In preli-minary tests it was observed that membranes with pore diametersof around 400 nm avoided the transfer of fungal spores andhyphae out of the living layer while allowing nutrients to pass.

Two types of living surfaces were used during this work. Letter-sized specimen for qualitative proof of principal and smallsurfaces (7 cm2) for quantitative experiments. The productionof large living surfaces is shown in Fig. S1. The fungi were firstimplemented into an agar layer, which can be modified throughapplication temperature (affects the liquid viscosity) and roll ordoctor blade coating conditions. Large demonstrator foils had asurface of approximately 500 cm2 and a thickness of 300 μm. Thesmall specimens used for mechanistic investigation had an acrylicbase layer which was transparent, flexible, and had a thickness ofaround 100 μm. The living layer itself had a thickness of around1.4 mm directly after application, and about 300 μm when dry(Fig. 2). The living layer was transparent as long as few colonyforming units (CFU, <300 CFU∕cm2) were present in the layer,and became opaque when the mold proliferated (photographs inFig. 3). A membrane (10 μm; 400 nm pore size) was used as thetop layer responsible for the transportation of food and gasesfrom and to the living layer (Fig. 2C). A top view of a partiallygrown living surface is shown in Fig. 2D.

Author contributions: L.C.G., F.M.K., R.N.G., andW.J.S. designed research; L.C.G. and F.M.K.performed research; L.C.G. analyzed data; L.C.G., F.M.K., R.N.G., and W.J.S. wrotethe paper.

The authors declare no conflict of interest.

This article is a PNAS Direct Submission.1Towhom correspondencemay be addressed at: HCI E107,Wolfgang-Pauli-Strasse 10, 8093Zurich, Switzerland. E-mail: wendelin.stark@chem.ethz.ch.

This article contains supporting information online at www.pnas.org/lookup/suppl/doi:10.1073/pnas.1115381109/-/DCSupplemental.

90–94 ∣ PNAS ∣ January 3, 2012 ∣ vol. 109 ∣ no. 1 www.pnas.org/cgi/doi/10.1073/pnas.1115381109

Food Consumption. The growth and food consumption rate ofexternally provided food depended on the type of food and itsconcentration: Testing four concentrations of potato dextrosebroth (PD) with glucose concentrations ranging from 0 to 48 g∕Lshowed increasing growth rates (CFUend, Fig. 3). A starting layerwith 2;000 CFU∕layer was first transparent and fungal hyphaewere only recognizable under an optical microscope. After 8 d,layers with no food provided (model spills consisting of physio-logical saline solution; no metabolizable food) showed no changein colony forming unit (CFU) count nor an associated change in

transparency. With increasing nutrient concentration, samplesdisplayed higher fungal hyphae density and became more opa-que. The glucose consumption increased similarly with increasingglucose supply. When 0.68 mg glucrose was provided, 0.64 mg wasconsumed within 8 d. This corresponds to a 94% reduction. Fromthe correlation between CFU and glucose consumption, we mayconclude that fungal growth was predominantly limited by theamount of provided food and not by the type of habitat, gastransport, or other transport limitation (see below). For further

Fig. 1. Design of a living surface. The simplest form of a living surface is composed of three layers. The base layer (light blue) may be an inert support foil (e.g. aPVC film), a linker—e.g. an adhesive—or a surface that shall be coated directly with a living surface. The second layer (white, with black CFU) is the living layerhosting microorganisms. The top layer (black, wireframe) is responsible for the confinement of the fungi, for nutrient, gas, and product transportation, andmechanical, chemical, and biological stability. (A) In the waiting state the number of potent/living fungi slowly degrades, but may survive for a long time e.g. inthe form of spores. (B) When a nutrient [e.g. a food spill (yellow)] is on the surface, the organisms within the living layer start growing and consuming the food.(C) Active phase: Here the microorganisms grow until all food is consumed. The membrane is responsible for the transport of food, gas, and secondary me-tabolites. (D) Secondary metabolites (e.g. penicillin) interacting with the surrounding allow complex functions of such materials.

Fig. 2. Layered structure of a living material on (A) macroscopic and (B) mi-croscopic level. The base layer consists of a sprayed 100 μm thick polyacryliclayer. The living layer mainly consists of agar and Penicillium roqueforti asactive organism. The top layer is a thin porous polycarbonate membrane.(C) Scanning electron microscopy image of the porous top membrane. (D)Microscopic top view of a partially grown living layer showing the two-dimensional hyphae network.

Fig. 3. Behavior of a living layer exposed to differently concentrated potatoglucose broths. Initially 2,000 CFU were present in a 7 cm2 sized living surface(left). Exposure to 2 mL of indicated media (diluted PD from 0.00–47.5gglucose∕L) led to the here shown CFU values after 8 d. On the secondaryy-axis the amount of consumed glucose is shown. Insets: Photographs of theextracted 8 d old living layers. The more food was provided, the more CFUwere found. This correlates to the increasing optical density shown. A con-stant ratio glucose consumed to CFU formed indicates that the growth is lim-ited by the available food and not by the habitat itself.

Gerber et al. PNAS ∣ January 3, 2012 ∣ vol. 109 ∣ no. 1 ∣ 91

APP

LIED

BIOLO

GICAL

SCIENCE

S

experiments a model food spill solution with 5 g glucose∕L (PD)was chosen to characterize the fungal growth properties undervarious conditions.

Long-term Survival and Hibernation.The evolution of fungal growth(CFU count) within the living layer and the glucose level on topof the layer (“spill-side,” user exposed) was observed over a timeperiod of over 1 mo (Fig. 4). At the beginning, each sample had12,000 CFU and the nutrient had a glucose level of 4.41 g∕L. Asthe fungal habitat was large enough, the growth was exponentialuntil most (>95%) provided glucose was consumed after about11 d. At this point a surface showed a CFU loading around45,000. The remaining 5% of glucose were consumed within3 d down to below detection limit and initiated a 7 d starvationperiod. Such conditions favor formation of spores (16, 17)and the experimentally measurable CFU load almost quad-rupled. To model a next use cycle, the living surfaces were againsupplied with a model food spill on day 21. The glucose leveldecreased again in a similar way as in the first cycle with asso-ciated increase or growth (higher CFU). As in the previouscycle, starvation-induced sporulation heavily increased the totalamount of CFU.

Robustness and Organism Confinement. In order to demonstratepotential application of such living surfaces in consumer exposedobjects, robustness of the encapsulated fungi and its fitness/performance is required to withstand aggressive cleaning or treat-ment procedures. Representative surface treatments includedshort sterilization with ethanol (disinfection; surface sweep withan ethanol based hand disinfection agent), standard soap, or 10 dwithout any nutrients under humid respectively wet conditions.Only surfaces exposed to severe dryness stress simulated througha 10 d period with no humidity supply were affected. These livingsurfaces were permanently harmed and no proliferation or glu-cose consumption was observed even after renewed supply ofwater or glucose even though potent fungi were found within theliving layer. Extraction of the fungal material from the living layerconfirmed that the fungi remained viable. Thus, the inhibitionof the glucose consumption function after severe dryness may beattributed to a lack of water, potentially due to very slow rehy-

dration of the dried-out agar support. None of the other harshtreatments resulted in a significantly reduced performance (i.e.model food spill cleaning) or fungal growth (Fig. 6). A startingCFU of 104.95 always led to 105.0–5.2 after all provided glucose(start: 8.3 mg) had been consumed (11 d). This considerable ro-bustness was attributed to the porous top layer, which acts as amembrane for food transportation and as a protective layer, re-ducing the effect of external stress and moderating changes inhabitat conditions, thereby allowing the organism to adapt (e.g.sporulation or reduced metabolism in case of dryness).

The importance of transport in this system was demonstratedby a simple experiment: Exposure of living surfaces to model foodspills of limited extension (e.g. spilled droplets) resulted in localactivation of the living layer (Fig. 5A). This localized nutrientconsumption resulted in local, dense two-dimensional hyphaenetworks within the microenvironment and imaging the originalarea of food supply (Fig. 5B). No fungal growth was observedin untreated regions (no model spills) or regions treated withphysiological saline (NaCl 0.9% in water, no metabolizable food).

DiscussionAfter the confinement of fungi into the two-dimensional micro-environment only few nutrients are present. Therefore, the fun-gus is close to a dormant state (Fig. 1A). In this phase, fungi are“waiting” and reduce energy intensive interaction with sur-roundings. By maintaining adequate conditions like humidity, theencapsulated microorganisms and especially their spores surviveand can be reactivated after several months when a nutrientsource (food spill, bacterial contamination) comes into contactwith the surface (Fig. 1B). The fungus consumes the “food” andproliferates within its domain (Fig. 1C). As the fungus is confinedto its thin two-dimensional polymer-based microenvironment,it cannot spread out of this habitat at any time and contaminateits surroundings. When all nutrients are consumed, the fungi maysporulate, switch back to a dormant state, and wait for new food.

The nutrient triggered growth of the fungi is conceptuallyrelated to the microorganism triggered release of pathogenicsilver concentrations from polymer surfaces [self-sterilizing sur-faces (18, 19)] or modified surfaces that kill bacteria on contact(20). Triggering enables timely delivery of action. The presentexample is the first eating material, an otherwise difficult to rea-lize target structure if restricting development to conventionalmaterial science strategies.

In this work we have shown a flat, two-dimensional, flexiblesupported preparation of a fungi-based living material, whichenables encapsulation of the fungi between polymer foils whilemaintaining nutrient and metabolite transfer with the exteriorenvironment (porous top layer). The fungal development insidethe living material depends only on available nutrients (from out-side) and space. Depending on the nutrients on the exterior of themembrane, the fungi consume the food and grow depending on

Fig. 4. Cyclic behavior of a living surface (7 cm2). At time zero, the livinglayer had 1;700 CFU∕cm2 and was fed with 2 mL diluted PD(4.41 gglucose∕L). For every time point two samples were analyzed. Whenthe glucose level reached 0.0 g∕L, one week of regeneration time wasgranted. At day 21 the surfaces were fed again with 2 mL PD (4.34 g∕L).As in the first cycle, the glucose content goes to zero, the CFU count goesup. The fast increase of CFU after that the renewed addition of glucose isattributed to spore formation due to starvation. The glucose consumptionrate and therefore the fungal activity were even higher in the second cycle.At the end of the second cycle, the CFU count increased disproportionatelystrong due to sporulation.

Fig. 5. (A) A ready-to-use living surface is activated locally by the applicationof a PD drop (model for a food spill). Exclusively under the drop the livinglayer is activated. (B) Photograph and microscopic magnifications of a300 cm2 sized living surface (top layer removed for images) after 5 d feedingwith PD (a food spill area, 6 cm in diameter, was placed in the bottom leftcorner). The fungi grew exclusively where food (PD) was present. No growthwas observed if only sodium chloride solution was applied.

92 ∣ www.pnas.org/cgi/doi/10.1073/pnas.1115381109 Gerber et al.

the amount of food available. In addition, the encapsulated fungiswitch to a sporulation state (rapidly increasing CFU) when thefood source is fully consumed. This behavior is well known frommost habitats with changing food supply—here, microorganismsalso survive by switching to an intermediate, dormant state. With-out available nutrients the fungi can survive for several weeksusing the here generated artificial biological habitat. Furtherimprovement may prolong this period (i.e. a measure for a ma-terial’s shelf life) significantly. The base layer and porous toplayer (membrane) provide a species barrier in two directions: theyprotect the environment (e.g. a user) from direct contact with thefungi and possible contamination, but also effectively protectthe encapsulated fungi from standard washing procedures withethanol or detergents (typical hospital routine operations), or,from invading species that may compete for the here generatedartificial habitat.

ConclusionUntil now, only few microorganisms are used in material science.Incorporating microorganisms in materials by providing anartificial habitat can have many advantageous effects. Here wedescribed how a fungus can be confined in a thin environment toobtain surfaces with self-cleaning (“eating”) properties. This mostprimitive first demonstration of a living material may later beextended to much more complex response, like bacteria triggeredrelease of antibiotics. In that case, the here proposed systemswould in principle mimic the natural ecological processes wherenumerous species compete for a resource. Using such competi-tion as a design element in material sciences offers creativedesign opportunities. Incorporation of well-known species likefungi (P. roqueforti, Saccharomyces cerevisiae), or even algae,bacteria, or symbiotic organism like lichen would extend theenormous technical benefits of such microorganisms since thediscovery of beer through fermentation of starches (malted bar-ley, wheat, maize) by yeast (21), or Fleming’s discovery of peni-cillin (22). Depending on the organism, its nutrition, secondarymetabolites, or growth conditions, various forms of such livingmaterials can lead to a plethora of application-oriented materials:Penicillin producing molds [e.g. Penicillium jensenii (23)] maylead to antibacterial surfaces, where the mold competes againstexternal organisms. Penicillium species are further able to pro-

duce a range of toxins including PR toxin (24), roquefortin C(25), mycophenolic acid (26), cyclopiazonic acid (27), and ochra-toxin A (28). Incorporating these species into foils could generatesurfaces displaying toxins on demand. In case of foils, fungi pro-ducing antibiotics could create near independent, self-sterilizingsurfaces, producing antibiotics on demand and therefore lower-ing the total amount of antibiotics. On a long-term perspective,living materials may show evolutionary adaptation to complexstimuli, thereby providing adaptable behavior under a numberof application conditions.

Living materials could be used in different ways. In a “dormantstate competent” form they await action (e.g. eating or disinfec-tion, see above) or may operate under steady state conditionsfor the production of biochemical agents, air cleaning, food pro-duction, or other harvesting systems. Due to the encapsulation,separation of the desired product is easy. Two-dimensional envir-onments increase the available surface to the exterior and mole-cules may rapidly diffuse in or out of the foil, enabling efficientproduction.

Materials and MethodsProduction of Living Surfaces/Test Samples. Cell culture inserts (Millipore,PIHP03050, diameter 31.5 mm, polycarbonate membrane, pore size 0.4 μm)were adhered by the aid of acetone to the cover standard 6-well plate(1.5 mm borehole for food supply). For the living layer an agar solution (15 gagar (AppliChem) in 985 mL water (ddH2O, Millipore, 18.2 MΩ cm at 25 °C),autoclaved 15 min at 121 °C) was heated until boiling to completely liquefythe agar. After cooling to 45 °C, 6 mL of a fungi suspension was added to40 mL liquid agar. The fungi suspension was obtained by high-shear mixing(IKA, Ultra-Turrax T18, 15′500 rpm) a fully grown Penicillium roqueforticulture in 600 mL ddH2O for 15 min cooled with an ice bath. To removenutrients, the suspension was centrifuged for 5 min at RCF ¼ 1;300 (Mistral3,000E, 2,500 rpm), the supernatant removed, and the fungi resuspendedin NaCl 0.9% (B. Braun). Mixed with the liquid agar, this resulted in a disper-sion containing 104–5 CFU∕mL. For blank samples (no fungi), no fungi suspen-sion was added at this point. Directly after the preparation of the agar-fungimix, 1 mL of that mix was applied to the bottom of a cell culture insertand allowed to stand for 2 h at ambient conditions. A thin layer of polymer(Dupli-Color, Aerosolart) was then sprayed onto the living layer. After 5 hdrying, the samples were turned over and used for testing. This procedureled to a structured design: Base/polymer—living layer—top/membrane.

Large Living Surfaces. Ten mL of an agar-fungi mix as above was poured at45 °C onto a polyvinyl chloride film (Perlen Converting AG, Switzerland, thick-ness 250 μm) and a 20 × 25 cm polycarbonate membrane (Sterlitech,PCT0420030) was applied with constant low pressure (1,000 Pa) to achievea thin spread of the agar-fungi mix. After 2 h, 2 mL potato dextrose broth(PD, Fluka), respectively 2 mL NaCl 0.9% (no food experiments) was place onthe surface covering approximately 20 cm2. A petri dish was placed overthe drops, and then the surfaces were placed in a constant climate chamber(Binder, KBF 240 ICH) at 25 °C and 80% relative humidity. After 6 d thesurfaces were analyzed visually.

Fungi. Penicillium roqueforti (Thom, anamorph deposited as Penicilliumroquefortii, ATCC 42294, LGC Standards) was grown in 75 cm2 T-flasks at22 °C on 50 mL PD for two weeks before usage.

Colony Forming Unit (CFU) Count. To quantify the amount of fungi in a livingsurface, the agar-fungi layer was pealed out and high-shear mixed (IKA,Ultra-Turrax T10, 30,000 rpm) in 20 mL deionized water (dH2O) for 1 minat maximum speed. A dilution row (100 to 10−5 in NaCl 0.9%) was plated onpotato dextrose agar plates (PDA, VWR BDH Prolabo). The plates were incu-bated at ambient conditions for 2–3 d before readout. A detailed procedureand validation of the method is described in the supporting information.

Glucose Measurement. The glucose level was measured with a glucose andsucrose assay kit (abcam, ab65334). Briefly, a sample was diluted to thetestable range (2–10 nmol∕well). 10 μL of the test solution was diluted with40 μL of glucose assay buffer and incubated for 30 min at 37 °C. After addi-tion of 46 μL glucose assay buffer, 2 μL glucose probe, and 2 μL glucoseenzyme mix, the sample was incubated for another 30 min at 37 °C. Theoptical density was then measured at 550 nm for colorimetric assay (platereader Tecan infinite F200).

Fig. 6. Demonstration of robustness through pretreatment of a ready-to-use living layer with disinfection agents, soap, or a 10 d starvation under hu-mid conditions which had no influence on the glucose consumption capacity(i.e. a measure for fungal proliferation). Control experiments: If no food wasprovided, the CFU count slowly decreased. With no fungi in the living layer,no glucose was consumed. Complete dehydration over 10 d killed the func-tion (no glucose consumption) of the living surface although potent fungiwere detected within the living layer. *Below detection limit (CFU < 10;glucose < 0.1 mg).

Gerber et al. PNAS ∣ January 3, 2012 ∣ vol. 109 ∣ no. 1 ∣ 93

APP

LIED

BIOLO

GICAL

SCIENCE

S

Food Consumption. Four food types were prepared by diluting PD withNaCl 0.9%. In duplicate, 2 mL of every food type was applied to small livingsurfaces and incubated at ambient conditions. After 8 d the supernatantbroth was analyzed in respect of the glucose content and the CFU countin the living layer was determined.

Long-term Survival and Hibernation. 22 small ready-to-use surfaces were fedwith each 2 mL 10% PD (in NaCl 0.9%) and incubated at ambient conditions.At every measuring point, the supernatants of two surfaces were analyzedwith respect to the glucose levels, and the living lavers were extracted todetermine the CFU load. After 21 d, all supernatants were replaced byfresh 10% PD (in NaCl 0.9%). All experiments were conducted in duplicate.

Robustness and Organism Confinement. Small ready-to-use surfaces wereexposed to pretreatments where 5 mL hand disinfection (Lysoform,AHD2000, 80 wt% EtOH), respectively 5 mL dishwasher (Handy, 1∶1 dilutedwith dH2O) was applied to the surface. After 5 s the agents were decanted,

the surface rinsed with 10 mL ddH2O, and rubbed dry with a Kleenex tissueapplying 20 kPa.

“Ten days wet” samples were covered with 2mL NaCl 0.9% and incubatedat ambient conditions for 10 d. “Ten days dried-out” samples were incubatedat ambient conditions for 10 d without any previous treatment. Directly afterthe pretreatment (where applicable, supernatant was removed) 2 mL of 10%PD (in NaCl 0.9%) was applied to induce grow of the fungi. A control withno pretreatment and a control with no fungi were fed with 2 mL 10% PD(in NaCl 0.9%). A control was fed with NaCl 0.9%. The surfaces were incu-bated at 25 °C for 11 d before the glucose level in the supernatant andCFU load within the living layer was determined. All experiments wereconducted in duplicate or triplicate.

Scanning electron microscopy (SEM) images were recorded on a ZeissLEO Gemini 1530. Optical microscopy images were recorded on a Zeiss AxioImager.M2m equipped with an Axiocam MRc.

ACKNOWLEDGMENTS. We acknowledge the financial support by ETH Zurichand the electron microscopy center at ETH Zurich (EMEZ).

1. Pearce R (1996) Tansley Review No. 87. Antimicrobial defences in the wood of livingtrees. New Phytol 132:203–233.

2. Stephenson SL (1989) Distribution and ecology of myxomycetes in temperate forests.II. Patterns of occurrence on bark surface of living trees, leaf litter, and dung. Myco-logia 81:608–621.

3. Burkholder PR, Evans AW, McVeigh I, Thornton HK (1944) Antibiotic activity of lichens.Proc Natl Acad Sci USA 30:250–255.

4. Cislaghi C, Nimis PL (1997) Lichens, air pollution and lung cancer. Nature 387:463–464.5. Cleveland J, Montville TJ, Nes IF, Chikindas ML (2001) Bacteriocins: Safe, natural

antimicrobials for food preservation. Int J Food Microbiol 71(1):1–20.6. Paul Ross R, Morgan S, Hill C (2002) Preservation and fermentation: Past, present and

future. Int J Food Microbiol 79:3–16.7. Gripon JC (1993) Cheese: Chemistry, Physics and Microbiology, ed PF Fox (Chapman

and Hall, London), pp 111–136.8. Molimard P, Spinnler H (1996) Review: Compounds involved in the flavor of surface

mold-ripened cheeses: origins and properties. J Dairy Sci 79:169–184.9. Kück U, Nowrousian M, Hoff B, Engh I (2009) Molds: living, benefit, hazard, control

(Springer-Verlag, Berlin), 3 Ed.10. Nielsen MS, Frisvad JC, Nielsen PV (1998) Protection by fungal starters against growth

and secondarymetabolite production of fungal spoilers of cheese. Int J FoodMicrobiol42:91–99.

11. Nielsen KF, et al. (1999) Production of mycotoxins on artificially and naturally infestedbuilding materials. Mycopathologia 145:43–56.

12. Synowiecki J, Al-Khateeb NA (2003) Production, properties, and some new applica-tions of chitin and its derivatives. Crit Rev Food Sci Nutr 43:145–171.

13. Blanchette RA (1991) Delignification by wood-decay fungi. Annu Rev Phytopathol29:381–403.

14. Berdagué J, Monteil P, Montel M, Talon R (1993) Effects of starter cultures on theformation of flavour compounds in dry sausage. Meat Sci 35:275–287.

15. Vreeland RH, Rosenzweig WD, Powers DW (2000) Isolation of a 250 million-year-oldhalotolerant bacterium from a primary salt crystal. Nature 407:897–900.

16. Piggot P, Coote J (1976) Genetic aspects of bacterial endospore formation. BacteriolRev 40:908–962.

17. Schaeffer P, Millet J, Aubert JP (1965) Catabolic repression of bacterial sporulation.Proc Natl Acad Sci USA 54:704–711.

18. Loher S, Schneider OD, Maienfisch T, Bokorny S, Stark WJ (2008) Micro organismtriggered release of silver nanoparticles from biodegradable oxide carriers allowspreparation of self sterilizing polymer surfaces. Small 4(6):824–832.

19. Gerber LC, et al. (2011) Incorporation of reactive silver tricalcium phosphate nanopar-ticles into polyamide 6 allows preparation of self disinfecting fibers. Polym Eng Sci51:71–77.

20. Tiller JC, Liao CJ, Lewis K, Klibanov AM (2001) Designing surfaces that kill bacteria oncontact. Proc Natl Acad Sci USA 98:5981–5985.

21. Homan MM (2004) Beer and its drinkers: An ancient Near Eastern love story. Near EastArchaeol 67:84–95.

22. Fleming A (1929) On the antibacterial action of cultures of a penicillium, with specialreference to their use in the isolation of B. influenzae. Br J Exp Pathol 10:226–236.

23. Farber P, Geisen R (1994) Antagonistic activity of the food-related filamentous fungusPenicillium nalgiovense by the production of penicillin. Appl Environ Microbiol60:3401–3404.

24. Wei R, Schnoes H, Hart P, Strong F (1975) The structure of PR toxin, a mycotoxin fromPenicillium roqueforti. Tetrahedron 31:109–114.

25. Scott PM, Merrien MA, Polonsky J (1976) Roquefortine and Isofumigaclavine-a,Metabolites from Penicillium-Roqueforti. Experientia 32:140–142.

26. Lafont P, Siriwardana MG, Combemale I, Lafont J (1979) Mycophenolic-Acid inMarketed Cheeses. Food Cosmet Toxicol 17(2):147–149.

27. Lebars J (1979) Cyclopiazonic acid production by Penicillium-Camemberti Thom andnatural occurrence of this mycotoxin in cheese. Appl Environ Microbiol 38:1052–1055.

28. Creppy EE (2002) Update of survey, regulation and toxic effects of mycotoxins inEurope. Toxicol Lett 127:19–28.

94 ∣ www.pnas.org/cgi/doi/10.1073/pnas.1115381109 Gerber et al.