Bioenergetics of simultaneous oxygen and nitrate respiration and nitric oxide production in a Pseudomonas aeruginosa agar colony biofilm

Abstract

Pseudomonas aeruginosa is a biofilm forming pathogen commonly associated with infection of the cystic fibrosis (CF) lung, chronic wounds and indwelling medical devices. P. aeruginosa is a facultative aerobe that can use nitrate (NO₃⁻) found in healthy and infected tissues and body fluids to generate energy through denitrification. Further, P. aeruginosa the expression of denitrification genes has been found in specimens from people with CF. The main aim of this study was to determine the relative energy contribution of oxygen (O₂) respiration and denitrification in single Pseudomonas aeruginosa PAO1 biofilm colonies under different O₂ concentrations to estimate the possible relative importance of these metabolic processes in the context of biofilm infections. We showed that the used strain PAO1 in biofilms denitrified with nitrous oxide (N₂O), and not nitrogen (N₂), as the end product in our incubations. From simultaneous O₂ and N₂O microprofiles measured with high spatial resolution by microsensors in agar colony biofilms under air, N₂ and pure O₂, the rates of aerobic respiration and denitrification were calculated and converted to ATP production rates. Denitrification occurred both in the oxic and anoxic zones, and became increasingly dominant with decreasing O₂ concentrations. At O₂ concentrations characteristic for tissues and wounds (20–60 μM), denitrification was responsible for 50% of the total energy conservation in the biofilm. In addition the formation of nitric oxide (NO), a precursor of N₂O and an important regulator of many cellular processes, was strongly influenced by the local O₂ concentrations. NO production was inhibited under pure O₂, present under anoxia (∼1 μM) and remarkably high (up to 6 μM) under intermediate O₂ levels, which can be found in infected tissues. Possible impacts of such NO levels on both the host and the biofilm bacteria are discussed.

1. Introduction

Pseudomonas aeruginosa is a Gram-negative opportunistic pathogen and is a well characterized biofilm-former responsible for a number of chronic infections [1,2]. Although originally considered generally an aerobic organism, P. aeruginosa is also cap(NO₃⁻) as an alternative electron acceptor for growth. In an oxic environment P. aeruginosa generates strong oxygen (O₂) gradients by rapid consumption of this, energetically most favorable electron acceptor, and the deeper parts of the biofilms can become anoxic[3]. Since NO₃⁻ is much more soluble in water than O₂ its higher concentrations and lower respiration rates allows deeper penetration into the biofilm than O₂. This allows metabolic activity deeper in the biofilm independent of O₂, thus denitrification may be an important factor in chronic biofilm infections by this pathogen. Both electron acceptors (O₂ and NO₃⁻) are present in tissues and fluids in wounds [4] and the diseased lung, as might be found in people with cystic fibrosis [5], which are also commonly infected with P. aeruginosa. Further, the expression of P. aeruginosa denitrification genes in CF isolates and sputum recovered from chronically infected patients suggesting that denitrification was occurring in vivo and might be relevant in the pathogenesis [[6], [7]]. Previously it was shown that O₂ respiration and denitrification could occur at the same depth in a P. aeruginosa agar colony biofilm [8]. The aim of the present study was to determine the relative contribution of simultaneous oxygen respiration and denitrification in single Pseudomonas aeruginosa agar biofilm colonies under different O₂ concentrations to estimate the possible relative importance of these metabolic processes in the context of biofilm infections. We investigated their relative importance for growth under different levels of O₂ saturation ranging from hypoxic to hyperoxic by switching headspaces gasses (air, pure O₂ and pure nitrogen (N₂)). Due to aerobic respiration, mass transfer resistance, and limited O₂ solubility, anaerobic micro-niches develop in biofilms [9]. Furthermore, in environments where O₂ levels can rapidly change between anoxia and air saturation, denitrification was reported to be insensitive to O₂ [10]. Therefore, we also investigated whether P. aeruginosa PAO1 agar colony biofilms were equally insensitive to O₂, and the dynamics of switching between aerobic respiration and denitrification. We used an in vitro model to simulate biofilms growing on soft surfaces where colonies of P. aeruginosa PAO1 strains grown on agar plates were submerged within a growth medium [8]. Here, we report the changes in local fluxes of O₂ and NO₃⁻ (from stoichiometry from nitrous oxide (N₂O) measurements) into the colonies under different O₂ concentrations and determined the consequences for bioenergetics. Since an infecting biofilm may be exposed to varying O₂ regimes, we hypothesized that denitrification contributes to energy conservation at varying O₂ levels. Secondly, we expect that due to well adapted e-transfer NO remains low (1 nM), independently from the O₂ supply. Bioenergetic calculations for denitrifying communities were based on data from Strous [11] to estimate the stoichiometry of ATP formed per rate of electron flow with O₂ and NO₃⁻ as e-acceptor, respectively. Although PAO1 can perform aerobic denitrification we expected effects of local O₂ concentrations on the electron flow to NO₃⁻ and thus on the denitrification pathway. The effect of different O₂ concentrations on the denitrification intermediate NO was determined by a NO microsensor.

2. Materials and methods

2.1. Bacterial strains, growth conditions and experimental set-up

Details on the growth conditions, the measurement set-up, and the microelectrodes used can be found elsewhere [8]. Briefly, frozen stocks of P. aeruginosa PAO1 were streaked onto brain heart infusion (BHI Difco 241,830) and incubated for 24 h at 37 °C. The plates were then transferred to a square glass sided (14 × 14 × 7 cm in dimensions) aquarium reservoir which was gently filled with 0.75 L of minimal salts media (MSM, in g/L): 0.7 K₂HPO₄, 0.3 KH₂PO₄, 0.01 NH₄SO₄, 0.05 MgSO₄*H₂O supplemented with 0.4 g/L glucose (MSM + G, Sigma-Aldrich) to submerge the plate with the biofilm. The pH was 7.3. The Petri plate was either weighed down to the aquarium bottom by wrapping with a flexible plastic-coated lead flask weight or attaching it using dental glue. Air, pure O₂, or pure N₂ (equating to O₂ concentrations of ∼21%, 100%, and 0% by volume in the headspace gas respectively) was introduced through an air stone to maintain gas saturation in the water phase and to provide water movement to ensure good mixing. Concentrations of O₂ in the media assuming saturation of the various gases were calculated from O₂ solubility as a function of temperature, pressure and salinity [12] exchange with the room atmosphere was reduced by using plastic balls (Allplas, Capricorn Chemicals Corp., Secaucus, NY) floating on the water surface.

2.2. Microelectrodes and calibration

O₂ and N₂O microelectrodes with tip diameters of approximately 10–15 μm were fabricated and calibrated as previously described [13,14].). The dissolved O₂ electrodes were calibrated using a 2 point linear calibration of fully saturated and anoxic sterile media obtained by sparging with pure O₂ and N₂ respectively. For the N₂O electrodes a saturated solution of N₂O was prepared. A dilution series was made to obtain a linear calibration between 0 and 400 μM N₂O. NO microelectrodes with a tip diameter of 15 μm were obtained from Unisense A/S (Aarhus, Denmark). They were calibrated in N₂ flushed water using a NO saturated stock solution to obtain a linear calibration between 0 and 7.3 μM.

The microelectrodes were mounted on micromanipulators positioned opposite of the incubation chamber, connected to a picoampere meters and the signals in picoamperes (pA) were recorded by a data acquisition system (model DAQCard AI16XE50; National Instruments, Austin, TX). Data acquisition and micromanipulator control was performed by laptop using a dedicated program.

2.3. Depth profile measurements and calculations of theoretical nitrate consumption

The microelectrodes were positioned at the surface of the same PAO1 colony using a dissecting scope (Fig. 1). They were positioned at an angle of 45° to the surface, thus the depth was corrected using the Pythagorean theorem. NO profiles were taken in the same colony while sequentially sparging the reservoir water with either air, O₂, or N₂.

Fig. 1.

A) Overview of the experimental set up showing the electrodes entering the reservoir containing the agar plate with the colony biofilm. B) Representative image of dissolved O₂ and N₂O microelectrodes positioned within a P. aeruginosa PAO1 colony at as close to the same position as possible. The biofilm colony surface was determined where the tips and shadows of the tips converged and from the signal response.

2.4. Confirmation of N₂O as a denitrification endpoint in our system

Although PAO1 has the genetic potential to denitrify completely to N₂ [15] Pseudomonas aeruginosa has been shown to produce mainly N₂O as an endpoint [16]. We tested whether our biofilms, under the given experimental conditions produce N₂ or N₂O as the primary end product by inhibiting N₂O reductase with acetylene (Yoshinari & Knowles, 1976). The incubation medium was sequentially bubbled with air, N₂ or pure O₂ and alternatingly approximately 20% acetylene was mixed in. During this procedure the biofilm was constantly profiled by an N₂O microsensor. We noted that acetylene addition led to instant swarming and dissolution of the biofilm and to stabilize the biofilm during measurements we coated them with an approximately 50 μm thick layer of 1% agar, by gently flowing 0.5 ml cooled molten agar (40 °C) over the colonies on the agar plate. Multiple profiles were taken in two independent experiment (Fig. S1) and the maximum flux of N₂O production, which occurred approximately 200 μm depth in the biofilm was calculated from the profiles [13]. Adding acetylene either reduced or did not significantly increase the N₂O production in the biofilms (P > 0.05), and also did not change the shape of the N₂O profiles (Fig. 2 and S1). This demonstrated that PAO1 under our experimental conditions performed denitrification to N₂O, and not further to N₂.

Fig. 2.

Maximum N₂O production flux in from multiple profiles of two independent biofilms under different levels of O₂ saturation (N₂, air and O₂) with and without acetylene, a nitrous oxide reductase inhibitor. There was no significant difference between independent replicates or with the addition of acetylene (P > 0.05, 2 tailed t-test assuming equal variance, d.f. > 8). The flux was significantly decreased under pure O₂ compared to air and N₂ (P < 0.05, indicated by *). The surface of the biofilm is at 0 depth. Data are mean and 1 S E., n = 18.13 and 10 for N₂, O₂ and air respectively.

2.5. O₂ and NO₃⁻ flux calculations

The distribution of O₂ and NO₃⁻ consumption by the colonies at steady state were calculated according to Fick's first law of diffusion: J = D_c (∂c/∂x), where J is the flux (mol m⁻² s⁻¹), D_c is the molecular diffusion coefficient (m² s⁻¹), and ∂c/∂x, is the local concentration gradient [17]. The N₂O profiles were converted to NO₃⁻ profiles using

$${{(NO}_3^{-})}_x={{(NO}_3^{-})}_b-\frac{D_{N2O}}{D_{NO3}}\times {\left(N_{2}O\right)}_x\times Y$$

with ${{(NO}_3^{-})}_x$ as the local NO₃⁻ concentration, ${{(NO}_3^{-})}_b$ the NO₃⁻ concentration in the medium, $D_{N2O}$ the diffusion coefficient of N₂O, $D_{NO3}$ the diffusion coefficient of NO₃⁻, ${\left(N_{2}O\right)}_x$ the measured local N₂O concentration, and Y the stoichiometry of NO₃⁻/N₂O conversion (mole per mole). Assuming a D_c in water of 2.36 × 10⁻⁹ m² s⁻¹ for O₂, 2.1 × 10⁹ m²s⁻¹ for N₂O and 1.84 × 10⁻⁹ m² s⁻¹ for NO₃⁻ at 24 °C [12,18], and correcting for an estimated reduced effective D_c (D_eff) in the biofilm of approximately 58% and 60% respectively [19]. We used a concentration of 1.2 mM NO₃⁻, approximating reported concentrations in CF sputum [20]. The local consumption rates of O₂ and NO₃⁻ were then calculated from the microprofiles as described previously [13]. Subsequently, the local volumetric ATP formation rates were calculated from the local O₂ and NO₃⁻ reduction rates, with 6.7 ATP/O₂ and 4 ATP/NO₃⁻, as reported previously [11].

2.6. Replicates and statistical analysis

We conducted a series of simultaneous measurements on five independent replicate biofilms on separate petri plates over multiple days. In our previous work we noted under air and N₂ at steady state that although profiles were similar for each location they were not the same, likely due to differences in colony size and morphology, local nutrient and mass transfer conditions and activity of the bacteria within the biofilms [8]. A representative additional data set is provided in the supplement.

3. Results

3.1. O₂ flux into the biofilm colony

The O₂ concentrations, controlled by sparging with air, pure O₂ or N₂, varied between 0.04 and 0.9 mM for the first measured biofilm (Fig. 3, S1), corresponding to approximately 23–530 mm Hg. N₂O was developed in the biofilm at all imposed O₂ levels (Fig. 3, S2), although it decreased when the bulk medium was sparged by pure O₂. Immediately upon decreasing the O₂ level in the bulk medium, the N₂O levels, driven by NO₃⁻ reduction, increased again. Similar results were obtained for replicate experiments (Figs. S2 and S3). We did not detect a threshold concentration of O₂ below which N₂O was produced, rather there was a steady increase in N₂O with a steady decline in O₂.

Fig. 3.

Time course of the O₂ (red) and N₂O (black) concentrations at approximately the same location within an agar colony biofilm. The first 4 profiles were measured under air saturation, then 3 profiles were recorded while the water column was sparged with pure O₂, followed by 4 profiles under air saturation, and 3 profiles under N₂ sparging. While pure O₂ sparging suppressed N₂O formation, it did not completely stop it. (For interpretation of the references to colour in this figure legend, the reader is referred to the Web version of this article.)

Examples of O₂ and NO₃⁻ microprofiles measured inside P. aeruginosa agar colony biofilms under different O₂ levels (i.e., O₂ oversaturation, air saturation and O₂ undersaturation), are shown in Fig. 4 and S2. The calculated O₂ and NO₃⁻ reduction rates and ensuing ATP production show NO₃⁻ consumption in the oxic and in the anoxic zone. The O₂ penetration depth was controlled by the O₂ concentration in the bulk medium. At the highest imposed O₂ levels, the NO₃⁻ consumption, and thus the coupled ATP production, decreased in the oxic zone. O₂ penetrated even at elevated medium levels maximally 0.4 mm, leaving a large part of the biofilm anoxic. At an O₂ concentration of ∼40 μM O₂, approximately 50 μM NO₃⁻ was produced. This O₂ concentration is in the range for tissues (∼13–90 μM; [21]), and approximately 20% lower than defined hypoxia (<40 mm Hg). This is common for chronic wounds [22]. O₂ penetrated only 0.05 mm, leaving most of the biofilm anoxic.

Fig. 4.

Representative measured O₂ (blue) and calculated NO₃⁻ profiles (red) under (a) O₂ oversaturation with sparging pure O₂, (b) air saturation and (c) O₂ undersaturation with sparging pure N₂ (c) in terms of concentration (left panels) and local ATP production rates calculated from O₂ and NO₃⁻ reduction, respectively (right panels). (For interpretation of the references to colour in this figure legend, the reader is referred to the Web version of this article.)

At decreased O₂ concentrations in the bulk medium, denitrification became proportionally more important, as expected (Fig. 5). During sparging with pure O₂, over 90% of the ATP in the biofilm was produced by O₂ reduction, and when the O₂ level dropped below 60 μM approximately 50% of the energy conservation was driven by NO₃⁻ reduction.

Fig. 5.

ATP production rates integrated over the depth of the biofilm, with ATP production by O₂ respiration (black), denitrification in the oxic zone (red), and denitrification in the anoxic zone (green). (For interpretation of the references to colour in this figure legend, the reader is referred to the Web version of this article.)

The NO concentrations in the biofilm varied under different gas conditions (Fig. 6). The highest NO concentrations (∼5 μM) were measured under air saturation. Under anoxia the NO concentrations were lower but still relatively high compared to reported physiological concentrations (∼1 μM), and NO almost went to zero under pure O₂ sparging.

Fig. 6.

A) Averaged NO profiles in biofilm incubated under air saturation (red, n = 4), pure O₂ (blue, n = 3) and anoxia under N₂ (black, n = 3) sparged medium over PAO1 cultures, and B) the corresponding O₂ profiles. Mean and 1 standard deviation error bars. (For interpretation of the references to colour in this figure legend, the reader is referred to the Web version of this article.)

4. Discussion

It was recently argued that a detailed description of the microenvironment of living microbial communities is essential for effective treatments of infections [23]. More generally, this holds for understanding the actual behavior of all natural microbial communities. In the present study we used an agar plate colony biofilm model to investigate spatially and temporally resolved O₂ and NO₃⁻ reduction rates, as well as NO production capability, and inferred energy conservation rates under varying oxygenation conditions. In our biofilms N₂O was the sole product of denitrification and accumulated in the biofilm, despite the strain having the genetic capacity to reduce to N₂ (Fig. 2) [24]. This suggests it is necessary to confirm inferences on physiological processes from genomic information by direct measurements. The accumulation of N₂O during denitrification was shown to occur in P. aeruginosa cultures because of inhibition of nitrous oxide reductase by low levels of O₂ (Betlach & Tiedje, 1981) as well as through quorum sensing under high cell densities (Kellermann et al., 2023). However, in our case we saw no effect of acetylene inhibition under anoxic conditions under N₂, ruling this possibility out. The incomplete denitrification with N₂O as the end product allowed us to base NO₃⁻ fluxes into the biofilm and bioenergetics of denitrification and O₂ respiration on N₂O concentrations and profiles.

The oxic status of the headspace influenced the absolute and relative rates of aerobic respiration and denitrification, and the response to changing headspace gases was relatively rapid occurring at the same location within minutes. When sparged with pure O₂, aerobic respiration was greatly stimulated by a factor of 5; but some denitrification continued. Similar to previous work [8], we found that even under O₂ supersaturation the aerobic respiration rate was so high that anoxic conditions were still produced at the base of the colony, thus allowing denitrification. Since antibiotic susceptibility assays for facultative bacteria, such as P. aeruginosa are typically conducted in a humidified 5% CO₂ incubator ([O₂] ∼ 180 μM) they may not accurately reflect the susceptibility in vivo of bacteria if present within a biofilm, as concluded previously [7]. Yet, in the oxic zone, denitrification was suppressed. When the bulk fluid was sparged with pure N₂, denitrification was stimulated due to part of the community switching from aerobic to anaerobic respiration as the conditions became anoxic. The almost instantaneous increase of denitrification upon lowering the O₂ level confirmed that the enzyme system responsible for denitrification was not damaged through oxygenation, but was only temporally inhibited [25]. The bacteria within the biofilm were therefore able to switch their metabolism in response to changing oxic-anoxic conditions rapidly, in the range of seconds to minutes. However, another explanation is that under air and O₂ supersaturation NO, the intermediate precursor to N₂O, is oxidized by O₂ ([26]), and the removal of O₂ contributes to the accumulation of NO and N₂O. Thus under these high O₂ concentration NO₃⁻ and NO₂⁻ reduction and the calculated ATP generation from N₂O production may be underestimated.

Biological energy conservation in the form of ATP is not proportional to the thermodynamics of metabolic process, but depends on the number of translocated protons per electron pair shuttled through the electron transport chains [11]. With O₂ as terminal e-acceptor, 20 protons are translocated per electron pair, while with NO₃⁻ 12 protons. As the ATP/H⁺ stoichiometry is 3 per O₂ respired 6.7 ATP is formed, while per NO₃⁻ only 4 ATP are formed. Consequently, although O₂ and NO₃⁻ respiration are thermodynamically close, at equal e-donor supply denitrification will result in lower growth rates than aerobic respiration, as indeed observed [27]. The lower growth rate by denitrification of P. aeruginosa PAO1 [5] compared to 1.5/hr (27 min doubling time) from aerobic respiration [27] may lead to reduced susceptibility to antibiotics [28,29]. The doubling time of P. aeruginosa biofilm aggregates in CF sputum was estimated to range between 100 and 200 min [27]. In chronic wounds hypoxia (<40 mm Hg O₂) is induced by damage to the vasculature, coagulation and consumption by inflammatory cells [22]. Hypoxia not only limits neutrophil activity against infecting bacteria but also fibroblast and many other host-mediated wound healing mechanisms [22]. Compounding the reduction of local oxygenation due to the inflict of a wound and the presence of bacterial biofilms that compete with host cells for O₂ further depletes the local O₂ concentration. Hyperbaric O₂ is used to correct hypoxia in order to enhance wound healing [30] and has been clinically shown to reduce bacterial load by a small amount [31] and in vitro studies show an increased susceptibility to antibiotics, presumably through increasing metabolic activity [32]. However, our data suggests that even if small aggregates of cells are present as seen clinically [33], even under hyperbaric therapy localized regions of hypoxia may occur due to the high rate of respiration, even under O₂ saturation. Biofilms may also induce local regions of hypoxia in the human CF lung as speculated from the expression of P. aeruginosa denitrification genes and supporting nitrate concentrations in samples collected from people with CF[[5], [6], [7]]

Denitrification has other implications in the context of biofilm infections. It is documented that denitrification can lead to the accumulation of the physiologically relevant concentrations of NO (120–150 nM) in the anaerobic zones of biofilms [17]. This signaling molecule induces vasodilation, inhibits platelet adhesion, inhibits leukocyte adhesion to endothelia and scavenges the superoxide anions [34,35]. We found unexpectedly high levels of NO, especially at O₂ levels typical for wounds and tissues (20–60 μM). NO levels peaked to 5 μM which is above the threshold for causing mammalian physiological effects [36,37], and interestingly, over the 1 uM concentration accepted as being toxic to bacteria [38]. Also, under anoxia NO levels were significant, but lower than at low oxygenation. NO was suppressed at high O₂ levels. The NO maximum production we observed at moderate O₂ levels (air sparged) may be explained by the balance between NO production by denitrification (intermediate accumulation enhanced by O₂ as e-transfer to NO₃⁻ is reduced) and consumption by oxidation (first order to O₂ levels) [26]. Thus at low O₂ concentration NO is low due to efficient reduction to N₂O, at high O₂ concentrations NO is rapidly oxidized chemically, while at intermediate O₂ concentration the production is optimal and degradation minimal. Interestingly, low dose NO is a known mediator of biofilm dispersal [39]. We did observe by eye “clouds”, which we presumed to be a high density of released cells, “billowing” out from the biofilm after a few hours of nitrate addition. In vivo this could lead to spreading of an infection. However, the role of NO in both health and disease is complex [40] and the impact of bacterially produced NO in localized infections remains to be elucidated. In future experiments it would be interesting to better correlate this observed phenomenon with NO concentrations within the biofilm.

Our novel approach on making simultaneous measurements at the same micro-location in a single colony allows measuring real-time responses to imposed system perturbations. Due to the technical complexity in setting up such experiments and colony-to-colony variability this technique does not lend itself well to obtain fully optimal replicate data for statistical analysis. However, we believe that our data on the transient response and relative contribution of denitrification and O₂ respiration with respect to spatial distribution and bioenergetics provides insights into metabolic processes possible within facultative aerobic biofilms. Our main conclusion is that metabolic activity within a biofilm colony is strongly influenced by dissolved gases, and that the response in the microenvironment to a change in conditions is rapid. Although under air-saturating conditions the energy yield of aerobic respiration was much higher than that of anaerobic respiration, the relative contribution of energy yield from anaerobic respiration will increase with reducing O₂ levels, as occurs in wounds and infections [22]. Increased mass transfer limitations, e.g. by thicker biofilms or reduced blood supply, will affect O₂ reduction more than NO₃⁻ reduction, as NO₃⁻ concentrations are much higher than O₂, and thus its penetration can be much deeper. Denitrification is thus important for energy generation of this pathogen. Secondly, the effect of O₂ on the NO production by these pathogens may well explain the severity of PAO1 infections and understanding this effect will help finding an optimal treatment against infections.

5. Conclusions

• 1)Although P. aeruginosa PAO1 has the genetic potential to denitrify to N₂ in our biofilm system N₂O was allowed to accumulate in the biofilm as the endpoint product, likely due to cell signaling because of high cell density.
• 2)P. aeruginosa PAO1 biofilms can rapidly switch between aerobic oxidation and denitrification in response to changes in O₂ concentrations allowing it to quickly adapt as dissolved gas conditions change.
• 3)P. aeruginosa PAO1 biofilms can rapidly consume O₂, which in vivo may afford it the ability to compete with host mediated wound repair and impair the immune response, even under hyperbaric therapy.
• 4)The expected lower growth rate by denitrifying cells in the oxic and anoxic zones in the biofilm may render these populations less susceptible to antibiotics.
• 5)At moderate O₂ concentrations as found in tissue and wounds, NO production is the highest and exceeding critical thresholds for physiological effects on host and biofilm, possibly influencing the pathogenicity of P. aeruginosa.

FUNDING information

Hanse-Wissenschaftskolleg Institute for Advanced Study (DB and PS). NIH grant R01 R01GM124436 (PS).

Competing interests statement

The authors declare no competing interests.

CRediT authorship contribution statement

Paul Stoodley: Formal analysis, Funding acquisition, Investigation, Methodology, Project administration, Supervision, Writing – original draft, Writing – review & editing. Nina Toelke: Investigation, Methodology, Writing – original draft, Writing – review & editing, Formal analysis. Carsten Schwermer: Formal analysis, Investigation, Methodology, Writing – original draft, Writing – review & editing. Dirk de Beer: Conceptualization, Formal analysis, Funding acquisition, Investigation, Methodology, Project administration, Supervision, Writing – original draft, Writing – review & editing.

Declaration of competing interest

The authors declare that there are no conflicts of interest.

Appendix A. Supplementary data

The following is the Supplementary data to this article.

Data availability

Data will be made available on request.