Preservation of H2 production activity in nanoporous latex coatings of Rhodopseudomonas palustris CGA009 during dry storage at ambient temperatures

Summary To assess the applicability of latex cell coatings as an ‘off-the-shelf’ biocatalyst, the effect of osmoprotectants, temperature, humidity and O2 on preservation of H2 production in Rhodopseudomonas palustris coatings was evaluated. Immediately following latex coating coalescence (24 h) and for up to 2 weeks of dry storage, rehydrated coatings containing different osmoprotectants displayed similar rates of H2 production. Beyond 2 weeks of storage, sorbitol-treated coatings lost all H2 production activity, whereas considerable H2 production was still detected in sucrose- and trehalose-stabilized coatings. The relative humidity level at which the coatings were stored had a significant impact on the recovery and subsequent rates of H2 production. After 4 weeks storage under air at 60% humidity, coatings produced only trace amounts of H2 (0–0.1% headspace accumulation), whereas those stored at < 5% humidity retained 27–53% of their H2 production activity after 8 weeks of storage. When stored in argon at < 5% humidity and room temperature, R. palustris coatings retained full H2 production activity for 3 months, implicating oxidative damage as a key factor limiting coating storage. Overall, the results demonstrate that biocatalytic latex coatings are an attractive cell immobilization platform for preservation of bioactivity in the dry state.


Introduction
The encapsulation of living cells to create living hybrid materials for use as biocatalysts, biosensors and bioremediation shows tremendous promise and has been the subject of extensive research for decades (Scott, 1987;Bjerketorp et al., 2006;Wang et al., 2010;Michelini and Roda, 2012). One of the fundamental challenges facing the successful commercialization of immobilized cell devices is how to preserve biological activity while retaining functionality and affordability (Bjerketorp et al., 2006;Wang et al., 2010;Michelini and Roda, 2012). Storage conditions that minimize metabolic activity, such as low temperatures or low relative humidity (i.e. freeze-dried, -80°C, under vacuum, etc.), are widely used and have a long history for microbial preservation; however, some of these require continuous cold storage and very little attention has been paid to the application of non-refrigerated conditions for cell preservation in modern immobilization matrices such as sol-gels or latex (Bjerketorp et al., 2006;Morgan et al., 2006;Tessema et al., 2006;Soltmann and Böttcher, 2008;Kuppardt et al., 2009). Furthermore, almost all immobilization matrices require liquid immersion or a humid atmosphere in order to maintain bioactivity of the immobilized cells (Bjerketorp et al., 2006;Michelini and Roda, 2012;Pannier et al., 2012). A notable exception has been the development of freeze-gelation techniques for biologically active biocers (biological ceramic composites) (Koch et al., 2007;Soltmann and Böttcher, 2008;Pannier et al., 2012).
Adhesive latex binders are a stable, non-toxic, nanoporous matrix that have advantages compared with other immobilization matrices, such as alginate and sol-gel, because it is adhesive, economical (produced in very large quantities at low cost for the water borne coating industry), does not collapse upon drying, and can be used to immobilize very high concentrations of cells . Experimental biocatalytic latex coatings have been designed for a variety of applications including: mercury detection, microbial fuel cell technology, high intensity chiral oxidations, and biocatalysis by thermophiles (Lyngberg et al., 1999;Fidaleo et al., 2006;Srikanth et al., 2008). Adhesive latex-based coatings are also ideal for photosynthetic H2 production because they provide a high surface area to volume ratio for incident light, efficient rates of gas diffusion, and can significantly increase cell longevity by protecting the microorganisms from mechanical degradation and contamination (Lyngberg et al., 2001;Flickinger et al., 2007). Indeed, latex coatings have been successfully developed for H2 production using the purple non-sulfur phototroph, Rhodopseudomonas palustris, with stable rates of H2 production (2.08 Ϯ 0.01 mmol H2 m -2 day -1 ) observed for > 4000 h when cell/latex coatings were hydrated in liquid medium that was periodically refreshed (Gosse et al., 2007;. Rhodopseudomonas palustris produces H2 via nitrogenase as an obligate product of enzyme turnover (N2 + 8e -+ 8H + + 16ATP → 2NH3 + H2 +16ADP + 16Pi), and in the absence of N2, nitrogenase acts as an ATP-powered hydrogenase where all electrons shuttled to the nitrogenase are available for H2 production (2H + + 2e -+ 4ATP → H2 + 4ADP + 4Pi) . Rhodopseudomonas palustris is an excellent candidate for biological H2 production because it can grow phototrophically and utilize a broad range of organic molecules as a source of electrons to support nitrogenase activity (i.e. organic acids, lignin monomers and other agricultural and food wastes) . Indeed, it was recently demonstrated that an inorganic electron source, thiosulfate, can also support nitrogenase activity in this versatile microorganism (Huang et al., 2010). H2 production efficiency is constrained in growing cultures of R. palustris as most electrons are routed for biosynthesis rather than to nitrogenase. However, R. palustris can generate up to 80% of the theoretical H2 yields when sustained in a non-growing state (Akkerman et al., 2002). Importantly, cells of R. palustris can remain active in a non-growing state through light-driven cyclic photophosphorylation, which maintains membrane H + gradients and intracellular ATP levels (Melnicki et al., 2008).
Preservation methods to ensure that latex-embedded cells can undergo long-term storage while maintaining activity have not been investigated. Coating preparation involves a controlled coalescence/drying step and, if needed, a subsequent short-term storage period, both affected by temperature and usually carried out at 60% relative humidity (Gosse et al., 2007). While this strategy has proven successful when coatings are rehydrated for use within several days of preparation, for industrial 'offthe-shelf' applications of microbial coatings, economics and practicality necessitate that coatings be stored in a dry state with stable reactivity over a much longer time frame (months-years). Storage conditions (temperature, humidity, O 2 tension, illumination or dark, etc.) must be defined to enable long-term dry storage of latexembedded microbial cells, preferably without the requirement for refrigeration.
Cells in coatings are subjected to desiccation stress, including osmotic shock, as the latex emulsion dries forming particle-particle coalescence and adhesion to the substrate resulting in cell immobilization. To moderate this stress, cell/latex formulations have been supplemented with sucrose, glycerol and other osmotic stabilizers (Yoo and Lee, 1993;Leslie et al., 1995;Lyngberg et al., 2001). Additionally, glycerol and carbohydrates serve to arrest polymer particle coalescence during film formation and thus increase nanoporosity within the coatings. Previous studies have demonstrated that these additives are required to maintain optimal levels of cell viability and activity in coatings upon rehydration 24-48 h after formation Gosse et al., 2007). However, the potential osmoprotective role of these compounds over long-term periods of dry storage has not been examined in bioactive latex coatings.
In order to advance the development of latexembedded R. palustris cells as lightweight, portable catalysts for H 2 production, we examined the effectiveness of (i) osmotic stabilizers such as glycerol, sucrose, trehalose and sorbitol and (ii) modifications to temperature, humidity and atmospheric O2 concentration during the critical film formation coalescence and storage periods to extend the shelf life of R. palustris coatings. The goal of our work here was to identify conditions, including the addition of trehalose, sorbitol or succinate, that preserve the H2 production capacity of dry R. palustris coatings over monthlong storage periods at room temperature. Here, we establish that biocatalytic latex coatings of R. palustris can be stored in a dry state at room temperature for up to 3 months, while retaining their original H2 production activity. To our knowledge, this is the first report of a method to preserve photobiological activity of embedded cells during dry storage.

Effect of osmotic stabilizers on H2 production by R. palustris latex coatings
The relative efficacy of three osmotic stabilizers for maintaining H2 production in R. palustris coatings following the initial dehydration period was examined. Latex formulations were supplemented with sucrose, sorbitol or trehalose, with or without glycerol, and stored for 2 days at 22°C with 60% humidity. Regardless of the osmolyte combination tested, coatings of R. palustris maintained H2 production for at least 85 days and after multiple media replacements (Fig. 1A). As observed previously (Gosse et al., 2007;, H2 production rates were highest immediately after hydration, and lower, but stable, after subsequent medium replacement/headspace flushing events (0.82 mmol H2 m -2 h -1 over~80 days; Fig. 1B inset). Rates of H2 production were similar among the stabilizer formulations tested (Fig. 1B).

Coalescence and storage temperature of R. palustris coatings and H2 production
Here, we tested the hypothesis that coatings coalesced and stored at cold temperatures would maintain higher H2 production levels following hydration. Rhodopseudomonas palustris latex coatings stabilized with glycerol and sucrose, sorbitol or trehalose were dried and stored at either 22°C or 4°C (7 days, dark, 60% humidity). Initially (up to 5 days post hydration), H2 production was much higher (6-to 10-fold) in coatings prepared and stored at 4°C than those treated at 22°C ( Fig. 2A). However, H2 accumulation was similar between all treatments following 19 days of incubation ( Fig. 2B), and 3 days after a subsequent medium replacement/flushing event (Fig. 2C).

H2 production by R. palustris coatings after long-term storage
To assess the ability of latex-embedded R. palustris cells to maintain activity over long-term storage, coatings containing different osmotic stabilizers were stored up to 4 weeks at 60% humidity then assayed for H2 production (Table 1). Coatings stored for 14 days exhibited comparable H2 yields to fresh coatings that were rehydrated and assayed < 24 h after coalescence. After 28 days of storage, two of three sucrose coatings, one of three trehalose coatings and all three sorbitol coatings failed to produce H2. The single active sucrose coating exhibited decreased H2 production yields (9.1% headspace H2) compared with fresh coatings (29% H2 headspace), whereas the two active trehalose coatings exhibited modest losses of H2 production activity compared with H2 production by R. palustris latex coatings containing different osmotic stabilizers (sucrose, sorbitol or trehalose; with or without glycerol). Freshly prepared R. palustris coatings were stored at 22°C under 60% humidity for 2 days prior to hydration in PM(NF) medium and initiation of the H2 production assays. A. H2 production is presented as per cent accumulation in the headspace over time where the headspace of each is tube is refreshed periodically with argon upon medium replacement (~every 2 weeks). B. Cumulative H2 production from the same coatings with arrows representing each medium replacement/headspace-flushing event. The inset in (B) shows the average rate of H2 production of all coatings, 0.82 mmoles H2 m -2 h -1 . Symbols represent average values obtained from duplicate experiments. fresh coatings (25% and 23% H2 headspace). None of the coatings generated H2 after 56 days of storage at 60% humidity.
To determine if R. palustris coatings could maintain greater H2 production capability when stored under conditions of low relative humidity, coatings stabilized with glycerol and either sucrose, trehalose or sorbitol were stored for 28 days at < 5% or 60% relative humidity.
Sucrose and trehalose coatings retained 67% and 59% of their respective H 2 production activity when stored at < 5% humidity over this time period. Importantly, all of the sucrose and trehalose coating replicates (three of each) retained H2 production activity when stored at < 5% humidity, and there was little strip to strip coating (technical replicate) variability (data not shown). Sorbitol coatings did not produce H2 after 28 days of storage, either at 60% or at < 5% humidity.
Additional experiments were conducted, with sucrose and trehalose as stabilizers (+ glycerol), to more fully assess the effects of storage humidity on H2 production by R. palustris coatings. As in our previous experiment, coatings stored for 28 days or longer exhibited very little or no H2 production activity when stored at 60% humidity and 22°C (Table 2). In contrast, when stored at < 5% humidity the sucrose and trehalose coatings retained significant activity, even after 56 days of storage (27% and 53% of the original activity respectively). When stored at < 5% humidity, each coating replicate (three of each) exhibited activity and the strip to strip variability was relatively small. It is also important to note that when stored at < 5% humidity, very little activity was lost from the coatings between 4 and 8 weeks of storage. Coatings initially dried under an atmosphere of < 5% relative humidity (48 h) then placed at 60% humidity behaved similarly to coatings that were coalesced and stored at 60% humidity -with complete (or nearly complete) loss of H 2 production activity after 28 days of storage (Table 2).

Respiratory activity of latex-embedded R. palustris after long-term storage
Rhodopseudomonas palustris cells embedded in latex were examined for anaerobic respiratory activity after various storage periods using 5-cyano-2, 3-ditolyl tetrazolium chloride (CTC). Rhodopseudomonas palustris coatings stabilized with sucrose (+ glycerol) that had been stored in the dark for 2, 14, 28 or 56 days under  either < 5% humidity or 60% humidity at 22°C were rehydrated anaerobically in PM(NF) medium and stained with CTC. CTC-stained coatings were examined with a confocal laser-scanning microscope to evaluate cell anaerobic respiratory activity of the embedded R. palustris cells (Fig. 3). Image analysis revealed dense, actively respiring cells through the z-plane when the coatings were stored at < 5% humidity for 2 days, while a less dense population was observed in coatings stored at 60% humidity (1.3-to 3.4-fold less CTC-stained cells based on image analysis; Table 3). Longer storage times resulted in a marked decrease in cell respiratory activity, which was strongly dependent on the storage humidity. Coatings stored at 60% humidity contained very few CTC-stained cells after 14 days of storage, whereas coatings stored at < 5% humidity still contained an appreciable number of active cells (on average,~15-fold more than the coatings stored at 60% humidity). CTC-stained cells were not detected in coatings stored at 60% humidity after 56 days storage, and < 1% of the number of CTC-stained cells enumerated in the 2-day-old coatings remained in the coatings stored at < 5% humidity (Table 3). Interestingly, the highest density of CTC-stained cells observed in the 2-day-old coatings was located in the upper portion of the coating, but after long-term storage, CTC-stained cells tended to cluster further into the coating interior (Fig. 3). These results led us to hypothesize that cellular damage incurred in coatings over longterm storage periods was related to oxygen exposure. Thus, sucrose or trehalose (+ glycerol) coatings were stored under argon or air for a period of 8-12 weeks at room temperature, and then assayed for H2 production and CTC response. After 8 weeks of storage at < 5% humidity, R. palustris coatings stored in an argon atmosphere exhibited two-to fourfold higher H2 production activity than coatings stored in air (Table 4). After 12 weeks of storage H2 accumulation was markedly higher (7-to 17-fold) in coatings stored in argon versus air. Notably, the H2 production activity measured in R. palustris coatings that had been stored in argon for 12 weeks was equal to or greater than the activity observed in freshly prepared coatings (Tables 1, 2 and 4). In accordance with these results, numerous CTC-stained cells were observed in coatings stored at < 5% humidity in argon following an 8to 12-week storage period (Fig. 4), whereas few actively respiring cells were detected in coatings that were treated similarly but stored under air. Additionally, the CTCstained cells did not appear to cluster preferentially in the interior of the coating after storage in argon. These results demonstrate that H 2 producing coatings of R. palustris can be stored dry, at room temperature, and maintain full activity for up to 3 months.

Discussion
An important technical hurdle that must be addressed before biocatalytic latex coatings can be used as 'off-theshelf' catalysts for H2 production, or other applications, is a. H2 accumulation in argon 10 days post rehydration in PM(NF) medium. b. H2 accumulation in argon (10-day incubation) following the initial H2 production period (20 days) and one flushing/medium refresh event.
c. Values are averages from three replicate strips Ϯ SD. Data are presented as the number of fluorescent spots detected in each plane of analysis. Values represent triplicate measurements (Ϯ SD), where fluorescent spots were enumerated for three randomly selected fields of view for each plane of analysis (coating depth).

Fig. 4.
Three-dimensional views of R. palustris coatings (+ sucrose/glycerol) stained with CTC (the axes distances are shown in mm). Coatings were stored at room temperature in the dark at < 5% humidity under argon (A, C) or air (B, D) for 56 or 84 days. It is important to note that although the images for Figs 3 and 4 were taken identically, the brightness and contrast of the images in Fig. 4 were adjusted to enable better visualization; thus, comparisons of fluorescent intensity between Figs 3 and 4 is not appropriate. However, each of the Fig. 4 images was treated identically so that comparisons between these images are still valid. that of maintaining bioactivity stability during long-term storage. In this study, latex-embedded cells of R. palustris were stored in a dry state at room temperature for up to 3 months while maintaining their original H2 production activity. Successful preservation of cell activity required the addition of select osmotic stabilizers, i.e. sucrose or trehalose, to the coating mixture, low relative humidity (< 5%) and anaerobic conditions during storage. It is important to note that we did not determine the dry storage lifetime of R. palustris coatings in this study. Yet, based on the observation that H2 production activity remained relatively stable over the final two sampling periods of this study (8 and 12 weeks), it seems likely that activity could be preserved in R. palustris coatings for greater periods of time.
Rhodopseudomonas palustris coatings dried and stored at 4°C produced greater amounts of H2 than those treated similarly at 22°Cin the days immediately following hydration. Over time (10-20 days post hydration), however, H2 production activity was similar between R. palustris coatings that had been stored at 4°C and 22°C. These results could signify that coatings stored at 4°C are capable of initiating H2 production activity quicker than those stored at 22°C, but that latex-embedded cells stored at 22°C can eventually recover full activity after periods of short term (1 week) storage. Alternatively, diminished particle coalescence at 4°C could lead to greater rates of acetate diffusion through the latex matrix upon hydration, thus greater H2 yields for R. palustris coatings prepared and stored at this temperature. As the coatings age under hydration, the sugars tend to leach into the medium and particle coalescence resumes, thus the permeability decreases and acetate accessibility would become more uniform for both treatment temperatures (Lyngberg et al., 2001).
The primary goal of this particular experiment was to determine if decreased temperatures during the film formation and polymer particle coalescence process and initial dry storage period would have an impact on H 2 production activity following hydration of the coatingsnot to determine if low temperatures could be used as a long-term storage strategy. In our view, storage at room temperature is more compatible with many of the shipping, storage and handling constraints that would make latex embedded cells attractive as a lightweight, off-theshelf biocatalyst or biosensor technology. The observation that R. palustris coatings exhibit a shorter lag in hydrogen production response time when prepared and stored at 4°C has implications for applications, such as biosensors, where a ready-to-use system exhibiting immediate activity is essential.
Freeze-drying and controlled drying without freezing are the methods of choice used by industry to preserve microbial cells, and while these techniques are primarily applied to cell suspensions or pastes, they have also been investigated as techniques to preserve immobilized cells. For example, bacterial sensor cells targeting molecules as diverse as N-acylhomoserine or arsenite/arsenate have been air dried onto filter paper, lyophilized, and stored at 4°C for 3 months or 30°C for 2 months without appreciable loss of reporter activity (Stocker et al., 2003;Struss et al., 2010). Freeze drying has also been applied to cells immobilized in sol-gels (Tessema et al., 2006;Meunier et al., 2010). However, freeze-dried cells must not be exposed to moisture and, despite high initial cell suspensions (greater than 10 8 ), survival rates of the original cell population can be as low as 0.1% (Bozoglu et al., 1987;Miyamoto-Shinohara et al., 2000;. These cell viability rates may be acceptable for propagation of the strain but are incompatible with biocatalytic latex coatings that are engineered for a high reactivity per unit of surface area (high intensity) and where cell growth is limited. Nonetheless, because our results demonstrate that latex-embedded cells of R. palustris retain considerable activity when stored under low relative humidity and respond quicker when prepared at lower temperatures, the applicability of freeze drying biocatalytic latex coatings for long-term storage (> 1 year) should be evaluated [latex embedded R. palustris cells were previously shown to maintain activity after storage for 1 year at -80°C ].
Regarding osmotic stabilizers, rates of H 2 production by R. palustris coatings were quite similar for each of the formulations tested (sucrose, sorbitol and trehalose Ϯ glycerol) when the strips were fresh or stored for Յ 2 weeks. Coatings prepared with either trehalose or sucrose retained 31-67% of their H2 production activity through 28 days under low humidity and 27-53% activity through 56 days of storage. In contrast, sorbitol-stabilized coatings were inactive regardless of the relative humidity levels beyond 2-week storage time. Although we did not investigate the underlying factors responsible for the differences in stabilizer performance, other studies have concluded that the efficacy of sorbitol as an osmoprotectant is quite variable (de Valdez et al., 1983;Carvalho et al., 2003), resulting in a greater emphasis on sucrose and trehalose as stabilizers (Leslie et al., 1995;Lyngberg et al., 2001).
The two elements found to be critical for preserving H 2 production activity in R. palustris coatings were (i) low relative humidity and (ii) low O2 levels during the storage period. Since slower drying rates allow for greater polymer particle mobility, coatings that are dried at 60% humidity have greater permeability than those dried at lower humidity (Lyngberg et al., 1999) resulting in greater porosity for gas and nutrient diffusion (Sperry et al., 1994;Ma et al., 2005). Yet, in this study we detected essentially no H2 production activity in R. palustris coatings stored at 60% humidity under air after 1 month. In addition to this loss of H2 production activity (Table 2) and general cell respiratory activity (Fig. 3), visual inspection of the R. palustris coatings also revealed differences between coatings stored under low or high humidity (data not shown). The characteristic red-purplish pigmentation of R. palustris dulls to a light red/orange colour over time during dry storage at 60% humidity, suggesting loss of light-harvesting bacteriochlorophyll or carotenoids. Coatings stored at < 5% humidity were much more resilient to pigmentation loss. We suspect that long-term storage at 60% humidity may provide the latex-embedded cells enough moisture to support low levels of metabolic activity, which could result in energy depletion over time or accumulation of metabolites in the pore space adjacent to cells that could become either toxic or exert osmotic stress on the embedded cells.
Oxidative stress is well known as a cause of cell damage and death during long-term storage (Dimmick et al., 1961;Meng et al., 2008). Oxidative damage to DNA, proteins, and particularly the cell membrane, has been implicated as a major contributor to the viability losses often observed when dried microorganisms are exposed to air for extended periods (Marshall et al., 1974;Israeli et al., 1975;Teixeira et al., 1996;Vriezen et al., 2007;Scherber et al., 2009). Water-deficient cells are unable to actively neutralize or excrete oxygen radicals or repair oxidative damage, thus cellular injury would inevitably and slowly accumulate until a threshold is reached beyond which cell recovery is improbable. In this study, we provide two lines of evidence that oxidative stress is an impediment to the long-term storage of R. palustris coatings. First, coatings stored under an argon atmosphere retained > 10 times greater H 2 production activity than those stored under air, even under conditions of low relative humidity where metabolic activity should be minimal. Second, as storage time elapsed under air, respiratory activity (assayed under anaerobic conditions) in rehydrated coatings was detected in R. palustris cells that tended to be clustered in the interior of the latex coating (Fig. 3), where O2 exposure during storage would be expected to be less than at the edges. In contrast, active cells were detected relatively evenly throughout the vertical profile of coatings that had been stored under argon (Fig. 4).
The enzyme responsible for H2 production in R. palustris, nitrogenase, is highly sensitive to oxygen (Gallon, 1992); therefore, exposure to O2 during long-term storage could result in longer lag times associated with H2 production (time required for repair or de novo synthesis of nitrogenase). Longer lag periods were noted for R. palustris coatings stored under air versus argon (data not shown), which could account for the large difference in H2 yields exhibited by coatings stored under these two conditions immediately following hydration (Table 4). Nonetheless, because coatings stored under argon continued to produce H2 at rates > 10¥ that of air-stored coatings up to 40 days after rehydration and after medium replacement, nitrogenase inactivation is almost certainly not the only damage incurred by latex-embedded R. palustris cells upon long-term storage in the presence of air. Accordingly, the CTC staining experiments revealed that general cell respiratory activity was compromised to a much greater extent in latex-embedded R. palustris stored in air.
Overall, this study demonstrates that biolatex coatings have great potential as an 'off-the-shelf' catalyst, considering that, with very little effort towards optimization, consistent retention of R. palustris activity was achieved following dry storage of coatings for at least 3 months at room temperature. If successfully applied to other microorganisms, the long-term storage properties of biocatalytic latex coatings would make it an attractive technology for a myriad of applications. For example, cell immobilization technologies applied to bioconversion, alternative fuel production, bioremediation, solar energy trapping, and food processing could benefit from long-term storage at room temperature using the methods described herein (Junter and Jouenne, 2004).

Bacterial strain, media, growth conditions and latex characteristics
Rhodopseudomonas palustris CGA009 was kindly provided by Dr Caroline Harwood, University of Washington. This strain produces hydrogen via three isozymes of the nitrogenase protein in an anaerobic environment at higher yields than the wild type due to an inactive uptake hydrogenase caused by a spontaneous frameshift mutation in the hydrogen sensor protein, hupV (Oda et al., 2005;Rey et al., 2006). Rhodopseudomonas palustris was cultured anaerobically in nitrogen fixing photosynthetic medium, PM(NF), supplemented with 20 mM acetate (unless otherwise noted, PM(NF) used throughout this study contained 20 mM acetate) in sealed glass serum bottles under a N2 atmosphere (Gosse et al., 2007). Rhodopseudomonas palustris was incubated statically under constant illumination at 60 mE with 60 W incandescent light bulbs at 31°C.
Latecies KAK4391 and Rhoplex™ SF012 (Rohm and Hass, Philadelphia, PA) latex formulations, both adjusted to pH 7.0, were used for this study. Latex KAK3941 is a vinyl acetate-co-acrylate that does not include biocide or hydroxyethylcellulose surface grafting, has a glass transition temperature (Tg) of 8.1°C, an average particle size of 280 nm, and a per cent solids of 52.5%. Rhoplex SF012 is a commercially available, acrylic co-polymer binder without biocide containing a solids content of 43.5%. Fig. S1) Rhodopseudomonas palustris cells were harvested in early stationary phase (OD660~0.8) by centrifugation at 7600 g for 15 min at room temperature. Cell pellets were suspended in 50 ml of PM(NF) medium without acetate and transferred to pre-weighed 50 ml Falcon tubes. After centrifugation (as above), the supernatant was removed and the wet weight of the cell pellet was determined. Prior to latex addition, the bacterial cell paste was first mixed with the indicated amount of glycerol and/or sucrose, sorbitol or trehalose. The latex emulsions for coatings were prepared based on the formulation ratio of 1.2 g wet cell weight, 350 ml of 1.7 M sucrose, sorbitol or trehalose, 150 ml of 100% glycerol (exceptions are noted) and 1 ml of latex. The two initial experiments (Figs 1 and 2) were conducted using KAK4391 latex; however, due to a discontinuation of this product by the manufacturer, the remaining studies were performed with Rhoplex emulsion SF012. Importantly, when using the SF012 latex, the polyester template was first cleansed with a small amount of 1 M HCl to minimize hydrophobic tension in the formulation upon spreading.

Preparation of R. palustris latex coatings (Supplementary
Rhodopseudomonas palustris latex coatings were prepared as strips using a template design consisting of a glass support, a 125-mm-thick polyester sheet (DuPont Melinex 454, Tekra Corp, NJ), and an adhesive vinyl mask (84 mm thick, Con-Tact, Stamford, CT). The polyester sheet was precut with parallel lines separated by 1 cm to define the width of each strip, and attached to a glass support covered with double-sided Scotch tape (strips perpendicular to the tape). An adhesive vinyl mask with a pre-cut rectangle (5 cm long to define the length of each strip; modified to 3.5 cm) was placed on top of the polyester so that its parallel pre-cut lines were in the centre of the mask opening. The width of the mask opening was determined by the number of polyester strips plus an additional 0.5 cm on each end [for template design details ].
Coatings were prepared under aerobic conditions in an acrylic glove box (Plas By Labs, Lansing, MI) at 22°C and 60% humidity, unless otherwise noted. Humidity was measured using a dew point, wet-bulb humidity thermometer (Fisher Scientific, Pittsburgh, PA). The latex/cell formulation was transferred from a Falcon tube onto an assembled template mask where it was then spread across the top of the mask along the width of the polyester window with a pipette, minimizing bubble formation (see Supplementary Fig. S1 in Supporting information). A 26-wire wound Mayer rod (Paul N. Gardner, Pompano Beach, FL) was drawn by hand down the template mask in order to spread the formulation. The coatings were allowed to dry for 24 h (unless otherwise noted) in the glove box before removing the mask. Each individual polyester strip (1 ¥ 5 cm, unless otherwise noted) 'painted' with embedded cells was then removed from the glass support and hydrated with 10 ml of PM(NF) medium in Balch tubes (Bellco Biotechnology, Vineland, NJ) (Gosse et al., 2007). Tubes were sealed with butyl septum stoppers, and flushed with argon for 30 min to produce an anaerobic environment for H2 production (16.5 ml of headspace). Rhodopseudomonas palustris coatings in sealed Balch tubes were incubated statically under a light intensity of 60 mE at 31°C.

Preparation of latex coatings under modified storage conditions
Coatings were prepared from a single culture of R. palustris using a separate template mask for each stabilizer combina-tion. After the latex/cell/stabilizer mixture was allowed to coalesce for 24 h, triplicate coatings from each treatment were removed from their respective masks and placed into separate Balch tubes containing PM(NF) medium to assay H 2 production (argon atmosphere; H2 accumulation was measured at 5-day intervals over 20 days). The remaining coatings were removed from the mask, placed in Petri dishes, covered with foil, and stored at 22°C under 60% humidity for 14, 28 or 56 days before assaying for H2 production. One set of triplicate coatings was stored at < 5% humidity for 28 days.

Gas analysis
Headspace gas analysis was performed using an Agilent 6890 gas chromatograph (Agilent Technologies, Santa Clara, CA) equipped with a thermal conductivity detector and a HP-Molseive column (30 m ¥ 0.32 mm ¥ 25 mm). Argon was the carrier gas and the oven, detector and inlet temperature settings were 50°C, 275°C and 105°C, respectively, yielding a RT of 6.02 min for H2 and baseline separation of N2 and O2. H2 was quantified by comparison of peak areas to standard curves constructed from known amounts of H2 gas (Gosse et al., 2007).

CTC staining and confocal microscopy
The respiratory activity of R. palustris CGA009 cells in latex coatings was assessed under anaerobic conditions using 5-cyano-2,3-ditolyl tetrazolium chloride dye (CTC; Sigma-Aldrich, St. Louis, MO). CTC is internalized and reduced by actively respiring cells to a fluorescent CTC-formazan that can be detected by epifluorescence microscopy (Rodriguez et al., 1992;Yu and McFeters, 1994).
Coatings of R. palustris CGA009, prepared with a latex/ sucrose/glycerol formulation, were dried at 60% humidity for 3 h and then stored in the dark at either 60% or < 5% humidity for 2, 14, 28 or 56 days. Additionally, coatings, prepared with sucrose or trehalose, were stored in septasealed Balch tubes containing drierite for 8-12 weeks under aerobic (a needle attached to a 0.2-micron syringe filter was passed through the septum) or anaerobic conditions (tubes were sealed and flushed extensively with argon; low O2 concentrations were confirmed by GC-TCD). Following the storage period, coatings were hydrated in PM(NF) medium and pre-incubated in sealed Balch tubes under an argon headspace for 4 days, after which CTC dye was injected to a final concentration of 4 mM. Coatings were incubated anaerobically with CTC in the dark for 1 h with constant shaking (100 r.p.m.) at 31°C. Heat-treated (85°C for 0.5 h) and unstained latex coatings were also examined as controls.
A confocal laser-scanning microscope (LSM 510 Meta, Carl Zeiss Microimaging) equipped with a HeNe1 laser was used to view the CTC-stained latex coatings. Images were collected at an excitation wavelength of 543 nm, a master gain of 633 V and with an alpha Plan-Fluor 100¥/1.45 oil objective. The confocal microscope facilitated viewing the cells at different depths (z-axis) within the latex coating matrix. Fluorescent spots were enumerated for three randomly selected fields of view for each plane of analysis