By Will Roberts, June 2018
Minimizing the populations of potentially harmful bacteria and fungi is a key goal of maintaining effective sanitation practices in our nursery. A new method to measure microbial populations was developed using small Petri dishes called contact plates. Using this measurement method, two experiments were performed to compare the effectiveness of 10% bleach and Physan 20 for sterilizing humidity domes and 7-gallon pots. 10% bleach was found to be much more effective than Physan 20 at sterilizing both surface types. Physan 20 did little better than no treatment at all in sterilizing pots. The contact plate method can be adapted to measure the effectiveness of nearly any disinfectant or cleaning process in managing microbial populations.
Sanitation is the foundation of any successful plant nursery and must be carefully managed to control the inception and spread of pests and disease. Common microbial diseases that threaten greenhouse crops include Botrytis, Fusarium, Pythium, and Pseudomonas. Diseases can spread via many pathways such as insects, tools, machinery, hands and dirty pots and trays. The easiest and most effective means of controlling these types of pathogens is prevention through systematic, effective and consistent sanitation practices. That has always been our goal at DHN but we have always lacked a means to directly measure the impact of our sanitation practices.
Complete sterility is usually not required (or at all practical) for most of our operations and the vast majority of microbes present in the environment pose no threat to our plants. However, there are many microbial pathogens in the environment that are potentially harmful to our plants and to our end customers that buy our clones. Thus the goal of our sanitation practices is to keep microbial levels to a low, manageable level.
The experiments described here focus on cleaning surfaces, with the goal to develop a method for quantifying the number of microorganisms present on a surface before and after cleaning. By measuring the microbial populations’ response to different combinations of physical washing (i.e. scrubbing and rinsing), disinfectants, and other processes, we can work towards optimizing our sanitation practices for maximum effectiveness.
Microbes, of course, cannot usually be seen or measured with the naked eye. Thus an indirect method of measuring microbial populations must be used. Contact plates are commonly used to sample the population of microbes in a small area (Figure 1).
Figure 1. Contact plates with colonies of microbes picked up during ‘contact’ with the sampling area.
Contact plates are small Petri dishes pre-filled with a sterile culture medium meant to culture a specific type of microorganism. Two types of contact plates were used in these experiments: Tryptic Soy Agar, which selectively cultures bacteria. The second, Sabdex Agar, selectively cultures yeasts and fungi. Using both types of contact plate allows us to gauge the effectiveness of a particular cleaning practice in killing specific types of microorganisms.
When sampling a surface, the contact plate is opened carefully, pressed against a flat surface for 2-3 seconds, and then immediately re-sealed with the lid. Great care is taken to sample only the surface of interest and not contaminate the culture medium with outside microbes. This could include accidentally breathing on it or exposing it to the air for long periods of time.
Once the samples have been collected, the plates are placed in a constant temperature incubator at 100 degrees Fahrenheit. 24 hours after sampling, the plates are examined for Colony Forming Units (CFUs). CFUs is defined as the number of visible colonies on the agar surface that grew from microbes picked up by the contact plate during sampling (Figure 1). CFUs then become the unit used to quantify the microbial population on the surface being sampled. It should be noted that the contact plate method cannot identify specific microbes beyond simply classifying them as bacteria or yeasts and fungi.
Two experiments were performed to measure the effectiveness of different sanitation practices and disinfectants:
Experiment #1 evaluated the effectiveness of cleaning and positioning of the humidity domes we use on our clone trays (Figure 2). Two factors were evaluated in this experiment: hood position (full-on vs askew) and pre-cleaning (freshly cleaned domes vs normal domes took from the production pile). Because humidity domes are disinfected with Physan 20 during washing, this factor could also be thought of as a comparison between 10% Bleach and Physan 20.
Figure 2. Full-on humidity domes with significant condensation building up on the insides. It was thought this condensation was a major contributor to possibly pathogenic microbes growing on the inside of the humidity domes and in turn, our clones.
The first factor of the experiment compared two different methods of placing the humidity domes on new clone trays as they are placed on the table. Full-on humidity domes create a seal around the edge that traps more heat, moisture, and humidity than humidity domes placed askew to provide more ventilation (Figure 3 below).
Figure 3. Askew (left) and a full-on (right) humidity domes.
The second factor of the experiment compared the cleanliness of the humidity domes as they begin their time on the table. Previous measurements had found very high microbial populations on ‘clean’ humidity domes used in production, which had been sprayed with Physan 20 and then rinsed off. This factor tested the hypothesis that pre-cleaning with 10% bleach would be more effective than the domes coming off the ‘clean’ pile that had gone through the normal Physan cleaning process. Other data, not included here, had shown that hoods cleaned with 10% bleach remain quite sterile, even after being put away wet and left at the production tables overnight.
Four trays of DHN Heartlets were chosen from normal production towards the end of the day and assigned one of the four treatment combinations of hood position and cleaning. Each treatment was replicated six times, with two replicates run per day for three days within a seven day period. The results were evaluated for statistical significance using the student’s t-test.
Experiment #2 evaluated the effectiveness of different disinfectants in sanitizing 7- gallon pots. All pots were scrupulously cleaned with soap, warm water and a stiff- bristled brush to remove all traces of soil or debris and leave the pot as clean as practically possible. The pots were then subjected to one of three treatments:
(1) No treatment: The pots were left to dry upside down immediately after washing without using any disinfectant. (2) Physan 20 Spray: Immediately after washing the pots were sprayed with copious amounts of Physan 20 and not rinsed. After the spray, the pots were turned upside down and allowed to dry. (3) 10% Bleach Spray: Immediately after washing the pots were sprayed with copious amounts of 10% Bleach solution, left for 30 seconds and then rinsed thoroughly. After rinsing the pots were turned upside down and allowed to dry.
All pots were allowed to drain and dry for 10-15 minutes before sampling with the contact plates. Both types of contact plates were applied to each pot. Three pots were used for each replicate in the experiment, chosen randomly from the pile of dirty pots next to the sink. There were a total of six replicates, with one replicate measured per day within a 14-day period.
Data analysis of CFU counts requires a process known as log transformation to transform the data into a distribution approximating the normal or ‘Gaussian’ bell curve distribution (Figure 4). The usual descriptive statistics (mean, standard deviation, etc) and relevant statistical tests (t-tests, ANOVA) assume a normal distribution and are essentially meaningless when applied to highly skewed data sets.
Figure 4. Non-transformed vs log-transformed distributions of the TSA plates used in the pot sanitation experiment. The non-transformed values are highly skewed, ranging from 0 to 2000, while the log-transformed values fit neatly into an approximately normal Gaussian bell curve.
The steps for the log transformation are to add a small offset (in this case 1) to each value in the dataset to remove any zero values, as the log of 0 is not a real number. Then the log of each value is taken and these ‘log-transformed’ values are used for the statistical tests. Log transformation is a very common statistical practice used with many types of data that tend to be heavily skewed, in particular, biological data. All values used in the analysis of these experiments were log-transformed in this manner prior to reporting and input into statistical analysis.
The statistical tests used for these experiments included the student’s t-test, one-way Analysis of Variance (ANOVA) and Tukey’s range test. The t-test and ANOVA compare the means and distributions of two or more groups of values to test the hypothesis that they are different populations. The analyses return a p-value that estimates the probability that the groups are from the same population. A p-value below 0.05, or a less than 5% chance they are from the same population, was used as the threshold for statistical significance. Tukey’s range test is a ‘post hoc’ test performed after an ANOVA to pinpoint the significant differences between specific treatments.
Results and Discussion
Experiment #1: Hoods
Table 1. The mean log-transformed Colony Forming Units (CFUs) of each treatment in the hood position/cleaning experiment. Each treatment was replicated six times.
The hood position had a small, but clear impact on the microbial populations, but the magnitude of the hood cleaning effect was much larger (Table 1, Figure 5). The student’s t-test was used to evaluate the significance of each factors effect, as well as the interaction effect of the two factors (Figure 5). The interaction effect is the magnitude of how much the level of one factor depends on the level of the other factor or factors. The only statistically significant factor for both contact plate types was found to be hood cleaning, with the hood position and interaction effects being much smaller.
These results illustrate that the microbial populations on the insides of the humidity domes are largely a function of how clean they are going into the process. It is not necessarily dependent on how they are used once they get onto the trays. Humidity domes that are relatively sterile going onto the trays will remain clean despite the lack of airflow and condensation that tends to build up on the insides of full-on domes. We’ve found in previous studies that in addition to trapping moisture, full-on humidity domes also trap heat. This may negatively impact our clones in another way, but the impact on microbial populations is relatively modest.
Figure 5. The magnitude of the effects of hood position, hood cleaning and the 2-factor interaction effect for the hood cleaning and position experiment. The decision limit is the threshold of statistical significance at the 95% confidence level. Only the hood cleaning effect proved to be statistically significant for both agar types.
Experiment #2: Pots
10% Bleach was found to be the most effective disinfectant for controlling both types of microorganisms (Figure 6). Physan 20 was largely ineffective at controlling bacteria but slightly more effective at controlling yeasts and fungi. The Physan treatments generally did not perform much better than no disinfectant at all (Figure 7).
Figure 6. Mean log-transformed CFUs for each treatment in the pot sanitation experiment for each agar type. Each treatment was replicated six times. The gold error bars represent one standard deviation around the mean.
No Disinfectant Physan 20 10% Bleach
Figure 7. Contact plates used to measure the three treatments of the pot sanitation experiment. The plates on the top are the bacteria culture plates. The bottom plates are the yeasts and fungi culture plates. Note how Physan (middle plates) has similar numbers of microbial colonies to the untreated pots while bleach has almost none.
The one-way Analysis of Variance (ANOVA) found a significant difference for the yeasts and fungi plates (p = 0.0056), but not for the bacteria (p=0.38). The lack of a significant difference in the bacterial plates was largely due to the high variation observed (note the large error bars in figure 6 above) This created a lot of overlap between the groups in the statistical comparison. Tukey’s range test confirmed that the only statistically significant difference was between the bleach treatments and no treatment for the yeasts and fungi plates.
10% Bleach was found to be a superior disinfectant to Physan 20 on both pots and humidity domes. It can be very effective at killing microbes when sprayed onto a surface, left for as little as 30 seconds, and then rinsed off. Physan 20, when used as a ‘spray and rinse’ or ‘spray and leave’ disinfectant, does not appear to be effective as in controlling microbial populations on hoods and pots. Using bleach as our primary disinfectant will thus be the most effective and convenient option for optimizing our sanitation practices.
Physan 20 can be an effective disinfectant on pots and other surfaces, but it does require a lot of time to be used properly. The Physan 20 label recommends soaking pots in a bath of Physan 20 or leaving the disinfectant to dry on the surface after application. Because our fast-paced sink washing process does not allow for long periods of soaking or drying, we need a ‘spray and rinse’ option that will be more convenient and conducive to the high throughput of washing we do every day. 10% bleach will be a much better fit for our needs at the sink and likely other applications throughout the nursery. And as these results show, even when used in this quick spray and rinse manner, 10% bleach can be very effective in sanitizing pots and other surfaces.
The methods used to measure microbes for these studies can be adapted to evaluate virtually any disinfectant or method of sanitizing a surface. There are even ways of adapting the method described here for use on curved or complex surfaces not easily sampled with a flat contact plate. In the future, we can use this method to evaluate the effectiveness of any proposed sanitation practices and keep our plants as clean and disease-free as possible.