The role of compatible solutes in desiccation resistance of Acinetobacter baumannii

Abstract Acinetobacter baumannii is a nosocomial pathogen which can persist in the hospital environment not only due to the acquirement of multiple antibiotic resistances, but also because of its exceptional resistance against disinfectants and desiccation. A suitable desiccation assay was established in which A. baumannii ATCC 19606T survived for ca. 1 month. The growth medium slightly influenced survival after subsequent desiccation. A significant effect could be attributed to the growth phase in which bacteria were dried: In exponential phase, cells were much more desiccation sensitive. The main focus of the present study was the elucidation of the role of compatible solutes, which are known to protect many bacteria under low water activity conditions, in desiccation survival of A. baumannii. Exogenous trehalose was shown to efficiently protect A. baumannii on dry surfaces, in contrast to other compatible solutes tested such as mannitol or glycine betaine. To analyze the importance of intracellularly accumulated solutes, a double mutant lacking biosynthesis pathways for mannitol and trehalose was generated. This mutant accumulated glutamate as sole solute in the presence of high NaCl concentrations and showed severe growth defects under osmotic stress conditions. However, no effect on desiccation tolerance could be seen, neither when cells were dried in water nor in the presence of NaCl.

, but it has also has been isolated from various places in hospitals during outbreaks, for example from furniture, door knobs, or equipment (van den Broek et al., 2006) and can survive in desiccated infant formula for 2 years (Juma, Manning, & Forsythe, 2016).
To date, few factors contributing to this extraordinary desiccation resistance are known. Besides the fact that biofilm forming strains survive longer on dry surfaces (Chiang et al., 2017;Espinal, Martí, & Vila, 2012;Orsinger-Jacobsen et al., 2013), RecA (a protein involved in DNA repair) (Aranda et al., 2011) as well as the acylation of lipid A (Boll et al., 2015) have been reported to be involved in desiccation resistance. A proteomics study performed by Gayoso et al. (2014) revealed mainly general features associated with desiccation resistance, such as the downregulation of genes involved in transcription, translation, and cell division, and the upregulation of genes for efflux pumps and antimicrobial resistance. Combined with observed changes in membrane composition, the authors propose a so-called "bust-and-boom" strategy.
In the present study, we aimed to investigate a possible role of compatible solutes in desiccation resistance of A. baumannii.
Compatible solutes are small, organic molecules which can be accumulated in the cell in up to molar concentrations without interfering with the central metabolism (Kempf & Bremer, 1998;Roeßler & Müller, 2001). They not only provide protection from osmotic stress by counterbalancing the osmolarity, but also by stabilizing membranes and proteins. Due to these stabilizing properties, compatible solutes can protect from many other environmental stresses, including desiccation. Besides that, the beginning of drought stress is usually accompanied by an additional osmotic stress, as the concentration of soluble substances increases when the liquid evaporates (Potts, 1994). Indeed, transcriptomic analyses under desiccation conditions in several bacteria, for example Anabaena, Rhodococcus, or Salmonella, revealed the upregulation of genes involved in biosynthesis or transport of compatible solutes (Katoh, Asthana, & Ohmori, 2004;Leblanc, Gonçalves, & Mohn, 2008;Li, Bhaskara, Megalis, & Tortorello, 2012). In other organisms such as the cyanobacterium Nostoc or E. coli, the compatible solute trehalose is accumulated in response to drought stress (Sakamoto et al., 2009;Zhang & Yan, 2012).
In response to salt stress, A. baumannii accumulates glutamate, mannitol, and trehalose or, if present, takes up glycine betaine from the environment (Zeidler et al., 2017), but so far nothing is known about a possible involvement of these solutes in desiccation tolerance. In particular, trehalose is a very potent protector against desiccation used, amongst others, by anhydrobiotes (Crowe, Oliver, & Tablin, 2002), and the unusual solute mannitol, which is a radical scavenger, could be involved in protection against oxidative stress that occurs upon rehydration (Efiuvwevwere, Gorris, Smid, & Kets, 1999). Herein, we have addressed the role of compatible solutes in desiccation resistance of A. baumannii.

| Bacterial strains and culture conditions
All bacterial strains used in this study are given in Table 1.
Acinetobacter baumannii strain ATCC 19606 T , Escherichia coli DH5α, and Bacillus subtilis JH642 were grown at 37°C and 130 rpm, while growth conditions for Micrococcus luteus were 30°C and 130 rpm.
Growth media were Luria Bertani broth (LB) (Bertani, 1951) Tschech and Pfennig (1984)), 20 mM sodium succinate as a carbon source, and 50 mM phosphate buffer. Stock solutions of all components were autoclaved separately. For growth under osmotic stress conditions, NaCl was added to the medium in the concentrations indicated (200-500 mM).
Growth rates were determined using the "exponential growth equation" analysis of GraphPad Prism for the exponential growth phase.

| Desiccation assays
Bacteria were grown in 5 ml cultures overnight and harvested in stationary phase, unless stated otherwise. 1 ml was harvested and washed twice in the same volume of H 2 O, with the addition of salt or compatible solutes where indicated. The same liquid was used to adjust the sample to an OD 600 of 2.0 ± 0.1, which corresponds to 1.2 × 10 7 -3.2 × 10 9 colony forming units (CFU) per ml. Aliquots of 20 μl of sample were applied to small polycarbonate filters (Nuclepore Track-Etch Membrane, 13 mm, 0.4 μm), which had been sterilized by autoclaving. Where indicated, saline (0.9% NaCl), 200 mM NaCl, or 10 mM of different compatible solutes were used for washing and resuspending instead of H 2 O. The membrane filters were put in petri dishes and incubated with slightly opened lids in a climate chamber, To test for statistical significance, unpaired t tests were performed at defined time points. P-values were calculated using the software GraphPad Prism. Statistical significance was assigned when p < 0.05.

| Extraction and quantification of solutes
Mannitol, trehalose, and glutamate were extracted from cells and quantified as described earlier (Zeidler et al., 2017). Briefly, bacteria were harvested in late exponential growth phase and lyophilized, followed by extraction of intracellular solutes with methanol and chloroform by a modified Bligh-and-Dyer method (Bligh & Dyer, 1959;Galinski & Herzog, 1990

| Establishment of a desiccation assay
In order to establish a desiccation assay suitable to investigate A. baumannii desiccation tolerance, we initially compared the survival of A. baumannii ATCC 19606 T with other bacteria. Bacteria were grown in LB medium overnight and washed before drying to remove all nutrients. In most desiccation assays described in literature, either water or saline is used to resuspend bacteria prior to drying. We decided to use water as was done for example in the study by Jawad et al. (1996)

| Desiccation resistance of A. baumannii depending on growth conditions
Many factors can contribute to bacterial desiccation tolerance.
To analyze the influence of the growth medium, A. baumannii was grown in mineral medium instead of LB. This resulted in a slight decrease in survival rate (Figure 2), which could be due to protective substances present in LB which can be taken up during growth. Desiccation resistance was therefore tested after growth in mineral medium supplemented with 1 mM of the compatible solute glycine betaine, which is contained in LB, but this did not increase survival compared to growth in non-supplemented mineral medium (data not shown). Another factor influencing bacterial physiology is growth temperature. However, bacteria grown in mineral medium at room temperature (22°C) did not exhibit a different desiccation tolerance (data not shown). Although growth temperature and growth media did not induce appreciable changes in the desiccation resistance of A. baumannii, the growth phase from which bacteria were harvested was found to significantly (p = 0.036 after 1 day) impact survivability of cells exposed to desiccation. In comparison with stationary phase cells, exponential cells of the pathogen proved to be extremely sensitive to the stress. For example, viable stationary phase cells were still detectable after ca. 3 weeks of drying while nearly no viable exponential phase cells could be detected after just 1 day of desiccation ( Figure 2).

ΔmtlD-otsB
Many bacteria are known to produce compatible solutes in response to drought stress. To elucidate whether intracellularly accumulated solutes are needed for survival of A. baumannii on dry surfaces, we established a deletion mutant defective in the biosynthesis pathways for both mannitol and trehalose. As expected, this mutant did produce glutamate as sole compatible solute during growth at high salinities (200-400 mM NaCl). Up to 0.3 μmol glutamate/mg of protein were accumulated, which is slightly less as determined earlier for the wild type ( Figure 5) (Zeidler et al., 2017). Growth of the ΔmtlD-otsB double mutant was significantly impaired at high NaCl concentrations, comparable to the ΔmtlD single mutant described in Zeidler et al. (2018) ( Figure 6). Without additional NaCl, the growth rate of the mutant was 0.60 ± 0.04 hr −1 , which is nearly identical to the wild type (92%) (Zeidler et al., 2017). In the mineral medium supplemented with 200, 300, or 400 mM NaCl, growth rates were 0.50 ± 0.04 hr −1 (86% of the wild type), 0.34 ± 0.08 (67%), and 0.06 ± 0.01 (19%), respectively, and at 500 mM NaCl no growth was observed. The addition of glycine betaine restored growth in the presence of 500 mM NaCl to a level comparable to non-osmotic stress conditions (data not shown).
However, when the mutant was grown in mineral medium and dried in water, survival rates were similar to the wild type (Figure 7), indicating that production of the compatible solutes mannitol and trehalose is not beneficial under these conditions. We hypothesized that the solutes could be required when bacteria were dried in a moderate salt concentration, as evaporation of the water increases the concentration, which might lead to the need for protective solutes as in other bacteria (Beblo-Vranesevic, Galinski, Rachel, Huber, & Rettberg, 2017;Bonaterra, Camps, & Montesinos, 2005;Reina-Bueno et al., 2012;Welsh & Herbert, 1999). Yet, drying in saline instead of water did not affect dying of the wild type nor of the mutant (data not shown). We assumed that the drying time might be too short to effectively induce production and accumulation of solutes.
Therefore, in a further experiment both strains were grown in mineral medium containing 200 mM NaCl, a condition inducing accumulation of mannitol and trehalose in A. baumannii, and subsequently dried in the presence of the same salt concentration (Figure 7).
Neither wild type nor the double mutant showed altered survival in the presence of salt.

| D ISCUSS I ON
Desiccation is one of the most important and severe stress factors that terrestrial bacteria must overcome in order to survive (Ramos  (Beuchat et al., 2013;Drabick, Gracely, Heidecker, & Lipuma, 1996;Jawad et al., 1998). Desiccation and subsequent rehydration lead, amongst others, to denaturation of proteins and DNA damage (Potts, 2001).
Simultaneously, microorganisms are frequently exposed to oxidative and osmotic stress as well as nutrient starvation, and, therefore, the respective stress responses overlap (Ramos et al., 2001). One ubiquitous key element in protection from various stress conditions is the so-called compatible solutes (Welsh, 2000). We previously analyzed compatible solutes and their role in osmotic stress protection of A. baumannii (Zeidler et al., 2017(Zeidler et al., , 2018 and now aimed to elucidate a possible role in desiccation resistance. In our study, A. baumannii ATCC 19606 T survived for approximately 1 month on dry surfaces, which is longer than was reported for the same strain in an older study by Jawad et al. (1996) but similar to that in another more recent study (Giannouli et al., 2013). When comparing survival times with data of other studies, it must be considered that many factors influence desiccation tolerance, for example the surface used and the inoculum volume (Hanczvikkel & Tóth, 2018;Neely, 2000;Wendt et al., 1997). In addition, it is highly strainspecific: clinical strains of A. baumannii exhibit a significantly higher desiccation resistance than laboratory strains (Giannouli et al., 2013;Jawad et al., 1998). The same holds true for other organisms; for E. coli, the reported survival times range from a few hours to several months (Kramer, Schwebke, & Kampf, 2006). Nevertheless, it is agreed that in general, Gram-positive bacteria have a higher desiccation resistance than Gram-negative (Janning & In't Veld, 1994;Nocker, Fernández, Montijn, & Schuren, 2012), which was also reflected in our assay.
To date, only a few studies have investigated the desiccation resistance of A. baumannii, and to our knowledge, we are the first to report on the sensitivity of the exponential phase cells of this pathogen to drying. Although this is a newly reported feature for

A. baumannii, similar traits have been reported in other bacteria
such as Salmonella enterica  or Sinorhizobium meliloti (Vriezen, de Bruijn, & Nüsslein, 2006). This can be attributed to the fact that stationary phase cells in general are more stress resistant (Kolter, Siegele, & Tormo, 1993). Stationary phase cells of A. baumannii have increased resistance against oxidative stress and it has been shown that various proteins involved in stress protection are upregulated (Soares et al., 2010). In a study by Jacobs et al. (2012), the transcriptomic analysis of stationary cells of A. baumannii revealed the upregulation of certain genes involved in trehalose biosynthesis, and this was more pronounced in clinical strains. The authors speculated a role for trehalose in desiccation resistance.
To test the hypothesis of an involvement of trehalose or compatible solutes in general in desiccation resistance of A. baumannii, we first analyzed the effect of extracellular solutes. Indeed, trehalose was the only solute which significantly increased survival times. The same has been reported for other bacteria such as E. coli (Louis, Trüper, & Galinski, 1994), Staphylococcus aureus (Chaibenjawong & Foster, 2011), and S. enterica . In all cases, glycine betaine did not have a positive effect, which is in accordance with our results. The outstanding effect of trehalose on desiccation resistance is attributed to its special chemical properties (Crowe et al., 1984;Elbein et al., 2003;Potts, 2001).
Despite the positive effect of exogenous trehalose on A. baumannii survival, a double mutant lacking biosynthesis genes for trehalose (otsB) and mannitol (mtlD) did not exhibit decreased survival.
We assumed that trehalose accumulation could be important for persistence under dry conditions not only because the otsBA promoter in A. baumannii is activated under osmotic and temperature stress (Zeidler et al., 2017) and otsB is important for persistence in Galleria mellonella larvae (Gebhardt et al., 2015), but also because endogenous solutes are involved in desiccation tolerance of many organisms, with trehalose playing an outstanding role (Argüelles, 2000;Elbein et al., 2003;Potts, 1994). In S. enterica, the genes for trehalose biosynthesis (otsBA) are upregulated 11-fold during desiccation (Finn et al., 2013). Transcriptomics in Bradyrhizobium japonicum, Anabaena, and Rhodococcus jostii revealed upregulation of biosynthesis pathways of compatible solutes as a common feature (Cytryn et al., 2007;Katoh et al., 2004;Leblanc et al., 2008). Also E. coli produces more trehalose, proline, and glutamine when dried (Zhang & Yan, 2012), and the high desiccation tolerance of C. sakazakii could be attributed to trehalose accumulation (Breeuwer, Lardeau, Peterz, & Joosten, 2003 Taken together, our data clearly show enhanced desiccation survival for cells in stationary phase and a protective role of exogenous trehalose, but do not point toward a connection between F I G U R E 7 Desiccation resistance of Acinetobacter baumannii ΔmtlD-otsB. Wild type (○, •) and the markerless deletion mutant ΔmtlD-otsB (Δ, ▲) were grown overnight in mineral medium and washed and dried in H 2 O (empty symbols, starting value 2-6.4 × 10 7 CFU per filter) or grown in mineral medium supplied with 200 mM NaCl, followed by washing and drying in 200 mM NaCl (filled symbols, starting value 6.4 × 10 6 -1.9 × 10 7 CFU per filter). For each experiment, mean values of at least four biological replicates are shown. Error bars represent the standard error of the mean (SEM) accumulation of trehalose or mannitol and desiccation resistance. Still it should be kept in mind that the exact experimental conditions can influence the results to a great extent (Finn et al., 2013), reflected for example by the fact that Gruzdev, McClelland, et al. (2012) could not detect upregulation of solute transporters in Salmonella, in contrast to others studying the same organism (Finn et al., 2013;Li et al., 2012).
Therefore, the fact that a proteomics analysis in A. baumannii did not reveal upregulation of proteins involved in synthesis of compatible solutes (Gayoso et al., 2014) does not definitely exclude a possible connection. Whether the effect of exogenous trehalose is physiologically relevant remains unclear. Just recently the intensified use of trehalose as a food additive has received negative publicity as it is associated with the increasing threat of Clostridium difficile as a nosocomial pathogen (Collins et al., 2018). It is important to know that trehalose in food could help A. baumannii to persist, especially as this pathogen has already been detected in several foods (Dijkshoorn et al., 2007).

ACK N OWLED G M ENTS
This work was funded by the Deutsche Forschungsgemeinschaft through DFG Research Unit FOR 2251.

CO N FLI C T O F I NTE R E S T
The authors declare no conflict of interest.

AUTH O R CO NTR I B UTI O N
SZ and VM designed the research, analyzed the data, and wrote the manuscript. SZ performed the experiments. All authors read and approved the final manuscript.

E TH I C S S TATEM ENT
This research did not involve studies with human or animal subjects, materials or data; therefore, no ethics approval is required.

DATA ACCE SS I B I LIT Y
All data are included in the main manuscript. Raw data and materials are available on request.