Introduction

A central point of investigation in our project was to determine how our bacterium can survive and thrive in its designated application place. From literature, we know that P. fluorescens natively lives in the root region, also known as the rhizosphere, of plants [1]. But this information alone is not enough to base our whole product on it.

From interviews with farmers we learned that the roots of numerous fruit trees are deliberately pruned to enhance fruit yield, resulting in roots not extending beyond a depth of 30-50 cm into the soil. Since those roots are situated near soil surface, our research aims to explore the colonisation patterns of bacteria in this region.

We want to apply our product in early spring, before the flowering starts. This implies that we need to investigate if our bacterium is able to survive and produce proteins in the cold spring soil, where temperatures near the surface can drop down to 4-6 °C [2]. Furthermore, it remained unknown to which extent bacteria attach to the roots and to which extent they colonise the surrounding rhizosphere without direct attachment.

Finally, we intend to apply our product via the already existing irrigation system of the farmer. Therefore, we modeled the colonisation behaviour of our bacterium after being applied through the irrigation system, its migration to the roots and its interaction with other bacteria along the way.

Microscopic studies on root-colonisation

We used a microscopic approach to study the colonisation behaviour of our bacterium in vivo and check the results of our agent-based model. As plant model, we used Arabidopsis thaliana to study the behaviour of P. fluorescens in the root system of A. thaliana (Figure 1). Though our final product will be applied to the root of fruit species, A. thaliana has the advantage of being well characterised, fast growing and easy in handling which makes it ideal to test interactions of P. fluorescens to plant roots in the proof of concept stage.

Fluorescence microscopy pictures were taken with 1 second exposure time using the objective for 100- and 1000-time magnification and later processed and analysed with Fiji (imageJ).

Does P. fluorescens attach to the root system of A. thaliana?

Two days after exposure to Pseudomonas fluorescens, the bacteria were uniformly adhering to both the roots and root hairs of Arabidopsis thaliana. When examining this under a 10x magnification, it appeared that nearly all of the bacteria had affixed themselves to the root, with only a few bacteria existing independently in the rhizosphere (Figure 1). However, when employing a 1000x objective with a higher magnification strength, we can discern that a minority of bacteria (ca. 10 %) were not bound to the plant's surface; but actively swimming within the surrounding medium (Figure 2). In conclusion, attachment of bacteria to the plant surface occurs, allowing the subsequent injection of the antiflorigen protein into the roots.

Figure 2: A: Fluorescence microscopy picture with GFP-expressingP. fluorescens andA. thaliana root upon 28 hours of incubation. A: 100x magnification, bacteria are evenly attached to root epidermis and root hairs. B: 1000x magnification, single cells attached to the root and freely moving in the rhizosphere.

Is attachment limited to certain tissues?

To further study the colonisation of P. fluorescens we wanted to determine where exactly bacteria attach to the roots. To gain insights, we consulted plant physiologists like Wouter Kohlen and learned that rhizospheric bacteria often exhibit a preference for specific regions of the roots to form colonies.

However, our experimental observations showed a different pattern. We found that bacteria evenly distribute across the root epidermis and root hairs, without displaying a specific preference for particular root structures. In the 1000x magnification, one can see that bacteria do not attach on top, but in the groove in between the epidermal cells (Figure 3). This surprising result is favorable in the context of our project, as it shows that attachment is not limited to certain tissues. However, we are aware that these results can be influenced by laboratory method of culturing in the laboratory and attachment patterns might substantially differ in a non-sterile soil environment.

Figure 3: Fluorescence microscopy image (magnification, conditions) showing attachment of bacteria in between epidermic cells ofA. thaliana roots (1000x magnification). A: fluorescence microscopy picture; B: Merged Channel picture.

Are modelling and experimental results coinciding?

Knowing that the new applied bacteria cannot go deeper than 20-30 cm in the soil within few days, we developed an agent-based model that predicts the distribution of colonisation of P. fluorescens on plant roots (Figure 4, A and B). From this model we learned that main colonisation is expected in the basal part of the root with decreasing quantities of bacteria towards the apical root region. For detailed information on our agent-based model, visit the Model page!

Figure 4: Agent based model and experimental data for spatial distribution of colonisation ofP. fluorescens on roots. Based on the model, high quantities of bacteria are expected in the first few centimetres below the surface, with decreasing numbers of bacteria in deeper soil layers, both in small (a) and big (b) roots. c: Part of experimental results from previous research on root colonisation ofP. fluorescens on the first 14 cm of a wheat root: main colonisation is detected in the first 14 mm below ground (Turnbull et al., 2006). d: Mean brightness in different parts of the microscopic picture of a one-week-old longA. thaliana root, 1 day after inoculation withP. fluorescens. The brightness is positively correlating with bacterial quantities. Even though a different range for colonisation is showed in model and experimental results (19 cm vs 11 mm), similar colonisation patterns are recognisable.

The decrease of bacteria could best be seen, when looking at the whole root system from basal to apical site. Under the phase contrast microscope equipped with a pH3 lens, bacteria appear as discernible white spots, while the roots themselves are mostly imperceptible, except for a white outline in the stele - a core tissue of the root, containing the root vascular tissue. Employing these settings (phase contrast, pH3), we captured six individual images and combined them to create an overarching representation of the entire root (see Figure 4). The degree of brightness in the image correlates with the abundance of bacteria in the rhizosphere. We processed the overview picture using image J and analysed the brightness of the microscopic pictures every millimeter from basal to apical root part. The results are shown in a heat map with the according values on the left site of Figure 4.

Conclusion

In the context of our project, the microscopic studies on the colonisation behaviour of P. fluorescens can the concluded as follows:

1. We showed that P. fluorescens attaches in high quantities to the root but is also able to thrive in the rhizophere of the roots. In the context of our project, attachment to the root epidermis is crucial as it is a prerequisite for the subsequent injection of antiflorigens into the plant.

2. We observed that P. fluorescens does not exhibit a preference for specific root structures such as root hairs or lateral roots for colonisation. This lack of specificity can be viewed as beneficial property, as it allows Pseudomonas to attach to various parts of the root, without being restricted to particular regions.

3. We obtained initial insights that support the cross-validation of our agent-based bacterial root colonisation model. Similar colonisation patterns are recognized for both the model and our microscopic studies. In summary, the results reveal that bacteria predominantly colonise the basal part of the root and decrease in number in deeper layers. Further research should be conducted with larger roots in a nonsterile environment to show if these colonisation patterns remain the same in non-laboratory conditions.

We evaluated several potential cold-responding promoters to ensure that our bacterial chassis, Pseudomonas fluorescens SBW25, has an increased expression outside of its optimal metabolic growth temperature of 25-30 ºC [1]. Such regulatory elements would induce high transcription at low temperatures in which the deployed bacterium would need to function. In addition to its main use for enhancing the heterologous expression of antiflorigen proteins, the promoter would also be used to up-regulate the other necessary constructed genetic circuits.

Moreover, the introduction of cold promoters in the engineered pathways would decrease the burden of heterologous production outside of cold temperatures. This would facilitate e.g. the growth of the modified bacterium in bioreactors at 30ºC.

We selected the upstream region of the cold-shock genes cspA, cspB, cspD, deaD and hutU, as described by Bartolo-Aguillar et. al (2022) and Craig et al. (2021), as potential cold-responding promoters [2,3]. Found in several Pseudomonas spp., these promoters are linked to the production of cold-shock proteins that act to inhibit structural damage at cold temperatures [4].

Experimental design

To assess the expression levels induced by different promoters of interest, we amplified them from their hosts and cloned them upstream of a sfGFP gene in a pSEVA83 backbone. The sfGFP gene was flanked by an LAA degradation tag, to eliminate promoters that are only induced shortly by a cold shock, and not constitutive in cold conditions. Next, these plasmids were transformed into P. fluorescens SBW25 and positive colonies were confirmed through Sanger sequencing. Afterward, the different cold promoter-bearing strains were incubated at a range of low temperatures for a period of seven days and evaluated daily on their growth and GFP expression.

For the construction of cold promoter plasmids, fragments of 300 basepairs from the upstream region of the selected genes were amplified. Primers used for amplification of the fragments were designed to introduce BsaI sites to both prime ends of the fragments, in preparation for Golden Gate assembly with the backbone. During amplification of the latter, a strong ribosomal binding site (RBS, BBa_B0034) was added to the sfGFP gene’s 5’ end, whereas in the 3’ end, a degradation tag was added (Figure 1).

Figure 1: GFP-producing P. fluorsecens on roots of A. thaliana (t=0 days after incubation). Cells are uniformly attached to the root including the apical site of the root with the root tip. In contrast to subsequent days, many bacteria are freely moving in the rhizosphere.

In addition to the construction of a variety of cold promoter-bearing strains, a strain with the absence of any promoter (‘empty vector’ strain) and another with p100 promoter (BBa_J23100) - a strong constitutive promoter - were constructed as negative and positive GFP expression controls. The cells were grown in minimal medium (M9 + apramycin) overnight at 30ºC and diluted to an initial OD₆₀₀ concentration of 0.1. The strains were then incubated at the following temperatures: 4 ºC, 10 ºC, 20 ºC and 30 ºC. Growth and fluorescence were measured daily with a Biotek Synergy Mx plate reader. Evaluation of the performance of strains was held with two technical and three biological replicates each.

Results

All incubated strains were able to grow in the tested incubation temperatures, but different promoter-bearing strains showcased differences in GFP fluorescence (Figures 2 and 3). As expected, we observed that a decrease in temperature resulted in slower bacterial growth of P. fluorescens. However, even at 4 degrees, growth was still occurring at a satisfactory pace for our PseuPomona project. From all assayed cold promoters, the upstream region of the cold-shock CspA2 gene portrayed significant sfGFP expression at different incubation temperatures.

Besides CspA2s, no other cold promoter was revealed to enhance expression in the conducted experiment. The most plausible reason is the possibility of the promoter elements being outside of the amplified regions, despite the attempt to verify the presence of sigma 70 promoter recognition sites in the promoter sequences, via Softberry’s bacterial promoter predictor [5] . Additionally, the promoters might be regulated by other regulatory factors that were not included in the amplified fragments.

The CspA2 promoter fragment displays enhanced expression correlated with the decrease in incubation temperature. After one day of incubation at 20 ºC and 10 ºC, the ‘pCspA2’ strain displayed a higher relative fluorescenc than the strong p100 strain. Whereas in the 4ºC incubation plate, a gradual increase in GFP was observed until the fourth day on incubation, where the strain reaches its stationary growth phase. After this day the ‘pCspA2’ strain maintained a GFP expression lead over the ‘p100’ strong constitutive strain. Counterweighing the enhanced sfGFP expression observed in ‘pCspA2’ strain a significant growth burden is noticeable within 4 ºC.

Figure 2: Optical density (ODλ = 600 nm) curves of cold promoter-bearing strains, incubated at temperatures of 30 ºC, 20 ºC, 10 ºC and 4 ºC, during a period of 6 incubation days.

Figure 3: sfGFP fluorescence (λ = 505 nm, gain 50) normalised to measured OD (λ = 600nm) of cold promoter bearing strains incubated at temperatures of 30 ºC, 20 ºC, 10 ºC and 4 ºC, during a period of 6 incubation days.

Conclusion

We discovered that the pCspA2 cold promoter enhances sfGFP expression throughout the days that followed incubation at different temperatures below 30 ºC. Moreover, it showcased a low GFP expression in the absence of cold induction (seen at day 0). Despite the found related burden, heterologous expression of sfGFP was very positive towards the PseuPomona goal that implies producing and delivering antiflorigen proteins to colonised fruit tree roots. Thus, the studied pCspA2 cold promoter would be a good fit to ensure enhanced protein expression in P. fluorescens SBW25 at the cold temperatures found in the soil during frost events.