Food waste from household consumption contributes largely to food loss as demonstrated in the study of Garcia-Herrero et al. . Besides measures for reducing production of food waste, methods for its valorisation is important. Urban farming is an example of alternative food production systems that provides locally produced food, thereby contributing to a reduction in long-distance transportation of foodstuffs into cities. However, for effective food production in cities, where possibilities for cultivation in soil are scarce, alternative production systems with low area requirements are needed. One solution to this constraint is to turn to hydroponic farming, where the food crop is grown directly in a nutrient solution , minimizing the space needed for cultivation. These production systems enable the possibility of farming in varying system designs such as horizontal, vertical or in several layers, and in diverse locations such as in basements, on rooftops and in containers, are thus attractive and promising systems to explore further in an urban context. Hydroponics have been used in traditional large-scale production of vegetables over the last three to four decades, however the nutrient supply is almost exclusively based on industrially manufactured, mineral fertilizers that challenge the pursuit for sustainable and renewable nutrient and resource loops . In Sweden, selected organic wastes, including food waste, are largely used for biogas production, with a nutrient-rich liquid digestate remaining as the by-product after the anaerobic digestion process. Using this anaerobic digestate as the nutrient solution in hydroponic cultivation systems could pave the way for a circular urban food production system as well as valorising food waste. Evidently, the use of recycled and biobased fertilizers constitutes an advantage from an environmental perspective compared to the mineral fertilizers used in conventional hydroponic production.
One major point of attention is however the close contact between the crop and the nutrient solution in hydroponic production systems. It is therefore paramount to primarily investigate and establish whether the anaerobic digestate is microbiologically safe to use for food production . In Sweden,flood and drain tray anaerobic digestate based on selected waste originating from the food and/or feed chain can be certified as biofertilizer according to SPCR 120 , a Swedish national regulation that needs to meet the criteria of the EU-regulation EC No. 1069/2009 regarding the treatment of biowaste . In order to fulfil the requirements for this certification, the feedstock used in the biogas process is initially hygienized by heat treatment . Previous studies have concluded that the combination of thermal pre-treatment followed by anaerobic digestion is successful in reducing Salmonella, Enterococci and Escherichia coli to acceptable/non-detectable levels as required by EU-regulation . However, while the presence and survival of these specific bacteria have been closely investigated, more in-depth studies into the overall biosecurity and pathogen content in anaerobic digestate from biowaste are encouraged . Regarding extended utilization in shorter nutrient cycles such as a hydroponic setup, which omits the natural processes occurring in contact with organic compartments such as soil, a thorough risk assessment becomes even more relevant. In a pilot study preceding the present work, three different, geographically distributed biogas plants in Sweden were sampled and the microbiological quality of the biofertilizer studied. In addition to the requirements in the certification, control of spore-forming species and presence of antibiotic resistance were conducted. The results confirmed that all formal criteria were met, however, biofertilizer from all plants had unsanitary levels of the food-borne spore-forming pathogen Bacillus cereus . This is in agreement with a previous study conducted on the hygiene aspects of biofertilizers where high levels of Bacillus spp. were detected , and it was deduced that neither the hygienization treatment nor the following anaerobic digestion affected the number of Bacillus spp. The overall scope of this study was to assess microbial risks related to the use of SPCR120 certified anaerobic digestate as a nutrient source in the hydroponic production of vegetables. The microbial viability and activity in the biofertilizer, before use in a hydroponic system, was initially studied over time with cultivation-based viable count, and cultivation-independent isothermal calorimetry.
Challenge testing with the three major food-borne pathogens B. cereus, Salmonella enterica ser. Typhimurium, and Listeria monocytogenes was also performed to investigate the biofertilizer’s susceptibility to contamination and ability to support microbial survival and growth. For assessment in hydroponic production settings, samples of circulating nutrient solution, based on either biofertilizer or inorganic fertilizer, were collected during a growth cycle in a greenhouse experiment and 16S rRNA gene amplicon sequencing was used to study the bacterial community composition over time. IC was utilized to measure the heat developed over time in samples with or without supplementation of nutrients. The heat developed is a result of metabolic activity of the organisms in the sample, and it was thus utilized as a cultivation-independent method of investigating microbiological viability and/or growth, an asset when assessing complex samples that may contain viable but not culturable cells. It also has the advantage of monitoring microbiological viability and growth without the introduction of bias that the agar plates selected in traditional standard plate count may account for, and the calorimetric measurement also gives an on-line and continuous output. At the same time, VC analysis was performed. Obtained VC results of samples without supplementation indeed pointed towards an actively growing microflora present, since a rich number of colonies was obtained on the plates at each sampling point. Contradictory to these results, there was no heat generation detected within the same samples when utilizing IC. The IC thus provided a presentation of the microbiological state of the biofertilizer without the bias that the introduction of nutrients from an agar plate may introduce. Fig. 2A and B shows the analyses from VC and FC after supplementation of BHI broth, and from IC after supplementation of BHI broth and glucose to the biofertilizer during the full duration of the accelerated microbial activity assessment experiment. As can be seen in Fig. 2B, the supplementation of glucose generates no metabolic activity. Regarding the supplementation of BHI broth, the first supplementation generates a heat production of around 100 J, with the following two supplementations generating a heat production of around 50 J, and the last two supplementations generating around 25 J. If it is assumed that the headspace of the vials is filled with air each time a vial is opened, the first supplementation generates more than the 50 J that aerobic metabolism can give, so this part does include anaerobic processes, but the lower heats indicate that the processes may be mainly aerobic.
Since the last two supplementations do not reach 50 J of heat produced, it is hypothesized that maximum growth capacity in the matrix has been reached, possibly due to restrictions in water activity or antagonistic behavior within the microbial community. The corresponding VC shows a 2.5 log increase in CFU mL− 1 after the first supplementation, and the following supplements induce no substantial increase in growth with either VC or FC . When performing microbial food safety risk assessments, not only presence/absence but also levels or concentrations of microorganisms, are valuable pieces of information needed to be able to evaluate food safety risks. From these results it appears however that the heat produced from metabolic activity is challenging to correlate with the number of CFU mL− 1 and cell count mL− 1 . According to a review by Braissant et al., the heat production of creating a cell should be rather constant, and it is also stated that if cell lysis takes place, this will cause a discrepancy between the heat generated and cell count . The fact that the sample contains a complex, mixed microbial community complicates the interpretation of the heat flow and heat generated after multiple supplementations, as the metabolism of different bacteria will generate different heat flows, and also there might be a succession of bacteria or a decline in some species caused by metabolites produced by the predecessor. This hypothesis was further hinted at when the visual inspection of the agar plates from the VC indeed revealed varying colony morphology and appearance after the different supplementations, andalso revealed some colonies exerting antimicrobial behaviour on their surroundings . This study has highlighted the difficulties of correlating VC with calorimetric data, which has already been observed in studies of other complex natural matrices . However, it can be stated regarding both the calorimetric data and the VC that while bacteria are indeed present in the biofertilizer, no metabolic activity is generated until the supplementation of an external complex nutrient source. As seen in Fig. 2A, the supplementation of glucose did not lead to subsequent metabolic activity while BHI broth did, indicating that an accessible source of carbon was not the limiting factor for microbiological activity, but potentially a combination of a carbon source and necessary trace elements that the BHI broth provides. In terms of the microbiological safety of utilizing this biofertilizer for hydroponic vegetable production, it is thus essential to avoid the addition of a nutrient source that can allow for the establishment of pathogenic bacteria. Although cultivation-based VC is a commonplace method of evaluating food safety in a matrix, it may allow for the introduction of false negative results when viable but not culturable cells remain undetected. As presented in this study, it might also produce false positive results when the agar plates provide the nutrients necessary to allow for microbial growth, nft hydroponic while the biofertilizer on its own does not provide the necessary factors for growth, rendering its natural microflora dormant. Studying the microbial community of complex natural matrices is in general difficult as several parameters of the matrix and its microbial processes are unknown.
As pointed out by Wads¨ o in a work using IC for studying the microbial activity in soil , IC is useful due to heat measurements being non-specific, and might thus be a preferable tool when investigating the total sum of complex microbial activities within a natural sample. Although IC is insufficient on its own for determining the microflora, separating the processes of one microorganism from the other, or separating microbial metabolic activity from microbial growth, IC and VC are excellent complementary tools when investigating the presence, viability and activity of complex microbial communities in their natural matrix. Microbial contamination of biofertilizer in hydroponics could cause serious consequences as the biofertilizer is recirculated and the plants are exposed to it during their entire growth cycle. A previous review, assessing the internalization ability of bacteria present in nutrient solution in hydroponic setups, concluded that present pathogenic bacteria and viruses internalize readily and more frequently compared to soilbased systems . This endorses the need for a deeper and more systematic understanding of how pathogenic bacteria would behave in the biofertilizer, in case of a contamination scenario, to assess the risks of using biofertilizers from anaerobic digestate for food production in hydroponic systems. As Bacillus cereus occurs naturally in the biofertilizer, and Salmonella and Listeria are able to internalize into growing crops , a challenge test experimental setup was performed to simulate contamination with the food-borne pathogens S. enterica serovar Typhimurium, L. monocytogenes and B. cereus, and assess their establishment, survival and growth in the nitrified biofertilizer over time. Fig. 3 shows the outcome of the inoculation of the food-borne pathogens B. cereus, S. enterica and L. monocytogenes in the biofertilizer. In the two biological replicates performed, S. enterica and L. monocytogenes were no longer detectable through selective plating within 48 h of incubation. Throughout the two biological replicates, B. cereus was steadily present in the control of non-inoculated biofertilizer and estimated at 1 log10 CFU mL− 1 . The biofertilizer inoculated with B. cereus decreased to these levels within 24 h after incubation. Previous microbiological controls at three Swedish biogas production plants had shown that the biofertilizer product after hygienization and anaerobic digestion contained up to 4.3 log10 CFU B. cereus per gram of biofertilizer . This level should be given attention, since the majority of food-borne outbreaks caused by B. cereus have been implicated with concentrations of 5–8 log10 CFU g− 1 of food of emetic toxin producing B. cereus. Occasional outbreaks of both emetic and diarrhoeal B. cereus illnesses with even lower levels have also been reported.