Membrane phase behavior of Escherichia coli during desiccation, rehydration, and growth recovery

Biochimica et Biophysica Acta 1788 (2009) 2427–2435

Contents lists available at ScienceDirect

Biochimica et Biophysica Acta j o u r n a l h o m e p a g e : w w w. e l s e v i e r. c o m / l o c a t e / b b a m e m

Membrane phase behavior of Escherichia coli during desiccation, rehydration, and growth recovery Cally M. Scherber a,c, Janet L. Schottel b, Alptekin Aksan c,⁎ a b c

Biology Program, University of Minnesota, Minneapolis, MN 55445, USA Department of Biochemistry, Molecular Biology and Biophysics, University of Minnesota, St. Paul, MN 55108, USA Biostabilization Laboratory, Mechanical Engineering Department, University of Minnesota, Minneapolis, MN 55445, USA

a r t i c l e

i n f o

Article history: Received 22 May 2009 Received in revised form 11 August 2009 Accepted 18 August 2009 Available online 28 August 2009 Keywords: Membrane phase change FTIR Desiccation VBNC E. coli

a b s t r a c t The membrane lipid bilayer is one of the primary cellular components affected by variations in hydration level, which cause changes in lipid packing that may have detrimental effects on cell viability. In this study, Fourier transform infrared (FTIR) spectroscopy was used to quantify changes in the membrane phase behavior, as identified by membrane phase transition temperature (Tm), of Escherichia coli during desiccation and rehydration. Extensive cell desiccation (1 week at 20%–40% RH) resulted in an increase in Tm from 8.4 ± 1.7 °C (in undried control samples) to 16.5 ± 1.3 °C. Fatty acid methyl ester analysis (FAME) on desiccated samples showed an increase in the percent composition of saturated fatty acids (FAs) and a decrease in unsaturated FAs in comparison to undried control samples. However, rehydration of E. coli resulted in a gradual regression in Tm, which began approximately 1 day after initial rehydration and plateaued at 12.5 ± 1.8 °C after approximately 2 days of rehydration. FAME analysis during progressive rehydration revealed an increase in the membrane percent composition of unsaturated FAs and a decrease in saturated FAs. Cell recovery analysis during rehydration supported the previous findings that showed that E. coli enter a viable but non-culturable (VBNC) state during desiccation and recover following prolonged rehydration. In addition, we found that the delay period of approximately 1 day of rehydration prior to membrane reconfiguration (i.e. decrease in Tm and increase in membrane percent composition of unsaturated FAs) also preceded cell recovery. These results suggest that changes in membrane structure and state related to greater membrane fluidity may be associated with cell proliferation capabilities. © 2009 Elsevier B.V. All rights reserved.

1. Introduction Proper functionality of cellular membranes is crucial for the protection and stabilization of living systems. The lipid bilayer is a primary structural and functional component of all cell membranes, and numerous studies have reported that changes in the membrane lipid packing as a result of water removal have irreversible effects on bacterial cell viability [1–3]. Cell membranes adopt a variety of conformations, including lamellar liquid crystalline (Lα) and lamellar gel (Lβ) phases, which are induced by lyotropic (osmotically induced) and thermotropic (temperature-induced) factors [4–6]. The Lα phase is characterized by greater membrane fluidity as a result of increased lipid head group spacing, decreased lipid acyl chain ordering, and decreased membrane thickness [6,7]. In contrast, membrane fluidity decreases in the Lβ phase due to denser packing of the lipid head groups, increased lipid acyl chain ordering, and increased membrane thickness [6,7]. As ⁎ Corresponding author. Mechanical Engineering Department, University of Minnesota, 241 Mechanical Engineering, 111 Church St SE, Minneapolis, MN 55455, USA. Tel.: +1 612 626 6618. E-mail address: [email protected] (A. Aksan). 0005-2736/$ – see front matter © 2009 Elsevier B.V. All rights reserved. doi:10.1016/j.bbamem.2009.08.011

water is removed from the lipid bilayer during desiccation, the polar lipid head groups are forced closer together, thereby increasing lipid packing and the van der Waals interactions between hydrocarbon acyl chains of neighboring lipids [8–10]. This increased attraction between lipid molecules forces the transition from the Lα to the Lβ phase, resulting in a decrease in membrane fluidity [2]. Upon rehydration, membrane fluidity increases as the cell reabsorbs extracellular water and transitions back into the Lα phase [8]. Membrane structure and fluidity are further affected by changes in the membrane lipid composition. Increase in membrane fluidity is directly correlated to an increase in the relative amount of unsaturated fatty acids (FAs) [11,12]. Due to the greater disorder of unsaturated FAs, the highly ordered lipid packing of the Lβ phase is prevented and the cellular membrane is maintained in the Lα phase. Numerous studies in bacteria have shown that changes in membrane lipid composition result from environmental stresses such as altered temperature [13] and pH [14], starvation [15], and desiccation [15]. Membrane phase transition during cell preservation is thought to cause the separation of membrane components [16], membrane fusion [16], and the co-existence of the Lβ and Lα phases, which creates mismatch resulting in an increase in membrane permeability and cell leakage [16,17]. These alterations in membrane lipid

2428

C.M. Scherber et al. / Biochimica et Biophysica Acta 1788 (2009) 2427–2435

organization and membrane phase change are usually catastrophic for living cells [16], although some organisms including seeds, pollen, bacterial and fungal spores, and yeast cells can survive such conditions [18] through survival mechanisms that are not well understood. Some bacteria are known to enter a viable but non-culturable state (VBNC) in response to environmental conditions that would otherwise be detrimental to cell survival such as nutrient-depleted growth medium, incubation temperature outside the normal growth range, osmotic stress, altered oxygen levels, or exposure to white light [19]. As was first identified by Xu et al . [20] in Escherichia coli and Vibrio cholerae, the VBNC state is now known to exist in 60 species [21]. In the VBNC state, cell growth is restricted and metabolic activity is repressed in media that would normally facilitate growth and colony development. Upon removal from adverse conditions, however, cells do recover, and cell proliferation and normal metabolic function are regained [19]. Although numerous studies support the VBNC state, there is some debate over the actual presence of such a state [22,23]. Previous experiments (Schottel, unpublished data) indicated that E. coli entered the VBNC state as a result of desiccation, and the majority of the cells returned to the culturable state when given adequate rehydration time. When colony formation capabilities of the cells recovered from the VBNC state were compared to undried control cells, it was observed that extensively dried E. coli cells (dried more than 1 week under ambient conditions) analyzed immediately after rehydration could form very few colonies. However, after about 72 hours of rehydration, up to 80% of the E. coli cells regained their ability to divide and form colonies. These results suggest that during desiccation E. coli enter the VBNC state and approximately 80% of the cells can recover upon rehydration if allowed to recover for 72 hours. The mechanisms by which cell recovery occurs and the changes in cellular proteins, the lipid membrane, and DNA during transition in/ out of the VBNC state are not well understood. Fourier transform infrared (FTIR) spectroscopy as well as differential scanning calorimetry [1–3,24] can be used to examine the structural changes of the major biological components of cells under different thermal and osmotic conditions. Furthermore, FTIR can identify and distinguish between the membrane lipids and cytoplasmic proteins based on the analysis of spectral bands that correspond specifically to protein secondary structure, acyl chains of the lipid molecules, and phosphodiester bonds of nucleic acids (Fig. 1A) [25]. Expanding on recent findings about the ability of E. coli to transition to/from the VBNC state, this study aimed at collecting structural evidence of changes in membrane composition and fluidity during the transition in/out of the VBNC that may be associated with cell recovery from desiccation. There are several ways to induce desiccation stress under ambient environmental conditions, including air, foam, spray, and fluidized bed drying (see Morgan et al. [26] for descriptions of the latter three methods); however, in this communication, we utilized air drying. In our study, the symmetrical acyl chain vibrations of carbon–hydrogen (CH2) bonds at 2850 cm−1 [27,28] were used to quantify structural changes in membrane lipid packing by measuring the thermotropic phase transition temperature (Tm) from the liquid crystalline (Lα) to gel (Lβ) phase (Fig. 1B). Tm is characterized as the midpoint temperature of this transition and studies have shown that Tm increased when liposomes [4,29], bacteria [1–3], pollen [30,31], and mammalian cells [28,32] were exposed to desiccation stress. For decades, cryogenic storage has been the preferred method for long-term preservation of microorganisms [26]. However, with the advancements in biomedical technology, including the development of cell-based therapies, tissue engineering and regeneration, and biopharmaceutical research, the need for inexpensive storage and processing of biomolecules and organisms in the absence of cryogenic conditions is in great demand. As a result, desiccation under ambient

conditions is a process that is attracting attention in preservation research. Understanding the physiological changes in membrane structure that occur during desiccation and rehydration, and the process by which cell recovery can occur during rehydration, may make it possible to employ desiccation as a less costly alternative to cryogenic preservation. 2. Materials and methods 2.1. Bacterial culture growth E. coli strain JM109 [F′ traD36 proA+ proB+ lacIq lacZΔM15/recA1 endA1 gyrA96(Nalr) thi hsdR17 supE44 relA1 Δ (lac-proAB) mcrA] [33] transformed with the ampicillin-resistant pmerGFP plasmid [34] was grown in 120 ml of Luria Bertani (LB) liquid medium (10 g/ l tryptone, 5 g/l yeast extract, 10 g/l NaCl, pH 7.2) at 30 °C with shaking to stationary phase. This strain, as well as E. coli strain B, has been shown to enter the VBNC state upon desiccation and to recover during rehydration (Schottel, manuscript in preparation). To assess colony-forming units (CFUs), cells were dilution plated on Penassay Agar (PAA, Difco). Ampicillin was added at a final concentration of 25 μg/ml to both liquid and solid media. 2.1.1. Membrane changes during desiccation To investigate the changes in the cell membrane during desiccation, a culture of E. coli was grown to stationary phase in LB and harvested by centrifugation at 6000 rpm for 10 min at 4 °C. The pellet was resuspended in 10 ml of phosphate-buffered saline (PBS), and the cells were harvested at 6000 rpm for 10 min. This procedure was repeated one more time, and the final pellet was resuspended in 3 ml of pyruvate buffer (PB) [5 mM pyruvate, NaK–phosphate buffer pH 6.8 (34 mM sodium phosphate, 33 mM potassium phosphate) and 0.091 mM (NH4)2SO4] [35]. A 30 μl cell sample was deposited on a CaF2 window placed in an environmental chamber at 25 °C to impose controlled drying. The relative humidity in the environmental chamber was set to 70%, and selected samples were dried for 0.5 to 4 hours to study the effects of progressive desiccation on cell membranes. Additional samples were dried at 5% RH for 1 hour to investigate the effect of rapid desiccation on cell membranes. After drying, all samples were resuspended in 10 μl of PB, immediately sandwiched between two CaF2 windows, and sealed. FTIR analysis (see below) was then conducted to monitor the membrane fluidity and Tm of the samples. 2.1.2. Desiccation followed by rehydration in PB To study the relationship between cell recovery and Tm during rehydration, a 120 ml stationary phase culture of E. coli was harvested and washed with PBS as described above except the final cell pellet was resuspended in 30 ml of PBS to an average cell density of 1.2 × 1010 cells/ml. Depending on the scale of the experiment, either 0.6 ml or 3.6 ml of cells were transferred to sterile 250 ml or 1 l Erlenmeyer flasks, respectively. The cell samples were dried under a laminar flow hood for approximately 15 hours, the flasks were capped with aluminum foil, and drying was continued for 1 week at room temperature and 20%–40% RH. After desiccation, the samples were rehydrated in PB, resulting in cell densities of 0.6–8.85 × 108 CFU/ml based on dilution plating prior to desiccation, and incubated for up to 72 hours in PB at room temperature. PB supplies a limited amount of nitrogen and carbon but does not support significant cell division (less than 20% increase in optical density of a cell sample was observed over a 3-day incubation period) [34], thereby preventing the proliferation of the residual culturable cells from interfering with cell recovery data. The bacteriostatic antibiotic chloramphenicol (Cm) [36] (25 μg/ml final concentration) was added to selected rehydrated samples to determine the effect of protein synthesis inhibition on recovery

C.M. Scherber et al. / Biochimica et Biophysica Acta 1788 (2009) 2427–2435

2429

Fig. 1. Membrane behavioral analysis of E. coli cell samples. (A) Characteristic FTIR spectrum of dried E. coli in PB. (B) Change in the CH2 stretch peak with temperature (filled squares) of E. coli in PB (undried control) shows the membrane transition between the gel and liquid crystalline phases. The first derivative of the data used to calculate Tm is shown by the circles.

during rehydration. A sample of the rehydrated cells was harvested by centrifugation, resuspended in PB (20 μl), and sealed between two CaF2 windows for FTIR analysis. The optical density at 650 nm (OD650) was measured to ensure that cell proliferation did not take place during rehydration in PB. Also, recovery was assessed by dilution plating the cells to determine colony-forming units (CFUs). CFUs of rehydrated samples were compared to the CFUs of the undried control samples to determine percent post-rehydration recovery. Fatty acid analysis was performed on samples at different time points during rehydration in PB (see below). 2.1.3. Desiccation followed by rehydration in LB Recovery and Tm were measured for E. coli cells that were desiccated for 1–2 weeks and rehydrated directly in LB for up to 72 hours at densities of 2.75–3.35 × 108 CFU/ml based on dilution plating prior to desiccation. At times during rehydration, a sample of the cells was harvested by centrifugation, suspended in PB (20 μl),

pelleted, washed, and resuspended in PB (20 μl). The sample was then sealed between two CaF2 windows for FTIR analysis. In addition, the samples were subjected to OD650 measurements and dilution plating to measure total CFU. Since LB is a nutrient-rich rehydration medium that facilitates cell division, CFU counts were not used to quantify cell recovery, but rather to determine the extent of lag time before the cells regained the ability to divide in the LB. 2.2. Determination of Tm and membrane fluidity using FTIR spectroscopy The effects of desiccation and rehydration on the Tm of E. coli membranes were studied using FTIR. All samples sandwiched between CaF2 windows were placed in a controlled temperature chamber and cooled down to −5 °C. The samples were then heated to 50 °C at a rate of 2 °C/min. FTIR spectra were collected every 1.5 minutes in the 930–8000 cm−1 spectral range using a ThermoNicolet 6700 spectrometer, fitted with a DTGS detector (Thermo-

2430

C.M. Scherber et al. / Biochimica et Biophysica Acta 1788 (2009) 2427–2435

Nicolet, Madison, WI). A characteristic IR spectrum of dried E. coli cells is shown in Fig. 1A. Membrane phase transition behavior was evaluated by monitoring the vibrational modes of the acyl chain CH2 symmetric stretching band at 2850 cm−1 [27]. The exact location of the CH2 peak was calculated through second derivative analysis, mainly by obtaining the second derivative of the spectra using an 11point Savitsky–Golay algorithm and determining the location of the peak maxima. The peak location of the band was plotted as a function of temperature. Tm was determined by identifying the temperature at which the greatest change in CH2 vibrational frequency occurred. Fig. 1B depicts an actual membrane phase transition curve and calculated Tm of an undried E. coli sample. T-test (two-sample test) was used to determine the statistical significance (P) of the experimental results. Statistical significance was defined as P ≤ 0.05. 2.3. Fatty acid analysis To assess fatty acid composition of whole cells, a culture of E. coli was grown to stationary phase in LB, harvested by centrifugation, and the pellet was resuspended in PBS. The cells were dried for 1 week and rehydrated in PB as outlined in Section 2.1.2. At times during rehydration, a sample of cells was harvested by centrifugation and washed with PB, and the pellets were stored frozen. Cells that had not been subjected to desiccation were analyzed for comparison as the undried controls. In addition, some cells were rehydrated in the presence of 25 μg/ml chloramphenicol to determine the effect of protein synthesis inhibition on changes in membrane FAs. The frozen cell pellets were lyophilized and sent to Microbial ID, Inc. (Newark, DE) where fatty acid methyl esters (FAMEs) of the whole cell samples were prepared and analyzed by gas chromatography against appropriate standards using the Sherlock Microbial ID System. The abundance of individual fatty acids was reported as percent composition of total membrane FAs. Fatty acid analysis was carried out with cells from four independent desiccation/rehydration experiments. The averages and standard deviations of the measurements at each time point for specific fatty acids were determined. 3. Results 3.1. Effect of desiccation on the Tm of E. coli FTIR analysis was used to determine the change in Tm and cell membrane fluidity with desiccation (Fig. 2). Cell desiccation was accompanied by an increase in the temperature-induced transition of the CH2 peak wave numbers from the gel to liquid crystalline state. Samples dried at 70% RH and 25 °C for 0.5–4 hours showed a

Fig. 2. The effect of different desiccation conditions on the membrane phase transition behavior of E. coli in PB. Tm values were determined as the temperature by which the greatest change in vibrational frequency of the phospholipid hydrocarbon acyl chains occurred.

progressive increase in Tm values from 9.25 ± 1.0 °C after 0.5 hour of desiccation to 14.8 ± 1.1 °C following 4 hours of desiccation. Note that the Tm for the undried samples was 8.4 ± 1.7 °C. Faster drying at 5% RH and 25 °C for 1 hour resulted in Tm = 15.4 ± 1.3 °C, and samples dried for 1 week at 20%–40% RH and 25 °C had Tm = 16.5 ± 1.3 °C. The differences in Tm were not statistically different among all groups (P = 0.6320, 0.3222, 0.2636, respectively) for cells desiccated for 4 hours (70% RH), 1 hour (5% RH), and 1 week (20%–40% RH), indicating that Tm reached its maximum value under these conditions. 3.2. Effect of rehydration in PB on the Tm and cell recovery of E. coli Tm of cells after 1 week of drying remained high at Tm = 17.3 ± 2.3 °C and was unaffected by rehydration for up to 24 hours after rehydration (Fig. 3A). A gradual decrease in Tm was seen after 24 hours of rehydration towards the Tm values of undried control cells, eventually plateauing around 48 hours at Tm = 12.5 ± 1.8 °C. Membrane analysis was continued with rehydrated samples for up to 72 hours after rehydration. However, no statistically significant differences in Tm between 48 and 72 hours was observed (P= 0.1517). The Tm values of undried control cells and cells after 48 hours of rehydration were however significantly different (P = 0.0049), indicating that the membrane phase temperature or fluidity did not entirely revert back to the state of undried control standards. E. coli samples that were rehydrated in the presence of Cm had a high Tm = 18.9 ± 3.0 °C even following 72 hours of rehydration (Fig. 3A), suggesting a hindrance in Tm regression as a result of protein synthesis inhibition. Following cell rehydration in PB, cells were plated on PAA for recovery analysis. The CFUs of rehydrated cells showed that, similar to the onset of decrease in Tm at 24 hours of rehydration, cell recovery began to occur around 24 hours, with maximum recovery occurring between 34 and 48 hours (Fig. 3B). By 48 hours of rehydration, an average of 1250-fold increase in CFU recovery was observed in comparison to cells at time zero of rehydration. Sample analysis was continued through 72 hours of rehydration, which resulted in an additional 1.5- to 3-fold increase in percent cell recovery. In samples rehydrated for 72 hours in the presence of Cm, there was little to no recovery observed (Fig. 3B). As stated previously, PB does not support significant cell division [34]. To verify division restriction in PB, OD650 readings were collected for each rehydrated sample and revealed a maximum difference of only 19% in optical density during rehydration from 0 to 72 hours (Fig. 3C). No significant change in OD650 observed during rehydration in PB indicated that no significant cell division took place during rehydration that would interfere with the cell recovery analysis. The wave number of the νsCH2 band at 1 °C, a randomly selected temperature, was plotted as a function of rehydration time to compare the membrane fluidity at different times during rehydration (Fig. 3D). For rehydration times shorter than 48 hours, the νsCH2 peak maximum was at a lower frequency than undried control cells, indicating lower membrane fluidity. However, the peak location increased above undried control samples at approximately 48 hours and continued to rise through 72 hours after rehydration. Although relatively small variations in Tm and cell recovery took place after 48 hours of rehydration (Fig. 3A and B), the continued increase in membrane fluidity up to 72 hours indicated that changes in membrane composition continued to occur. For the cells rehydrated in the presence of Cm, membrane fluidity was significantly lower than any sample (control or experimental) (Fig. 3D). Fig. 3E shows post rehydration recovery as a function of Tm. When E. coli were desiccated for 1–2 weeks, the Tm increased to 17.3 ± 2.3 °C, at which point the percent recovery was near 0.01%. At Tm values above approximately 13 °C, percent recovery remained under 1%, representing a threshold temperature at which little to no cell recovery was observed. Gradual increases in CFU recovery with rehydration time were accompanied with decreases in Tm. Neither a

C.M. Scherber et al. / Biochimica et Biophysica Acta 1788 (2009) 2427–2435

2431

Fig. 3. Recovery and Tm changes of E. coli cell samples as a result of rehydration time in PB. (A) Change in Tm with rehydration time. (B) Percent recovery change with rehydration time as measured by CFU. (C) Normalized OD650 as a function of rehydration. (D) Change of CH2 peak position at 1 °C with rehydration time. (E) Correlation between Tm and percent recovery after rehydration. Each individual point represents a separate cell sample. Independent experiments are indicated by the different symbols. (●) Tm and CH2 peak position at 1 °C were found for undried control samples for comparison. (○) Cm was added during rehydration in specific samples to inhibit protein synthesis. The dashed lines are a guide for the eye.

decrease in membrane Tm nor an increase in cell recovery was observed when samples were rehydrated in the presence of Cm, which suggested the importance of protein synthesis for recovery. 3.3. Effect of rehydration in LB on the Tm and cell recovery of E. coli In order to determine the effect of rehydration in growth medium on Tm and cell recovery, dried E. coli samples (dried for 1–2 weeks at 20%–40% RH and 25 °C) were rehydrated in LB and analyzed at various time points ranging from 0 to 72 hours of rehydration. Up to 26 hours

of rehydration, the Tm was 18.5 ± 2.3 °C (Fig. 4A). Between 26 and 33 hours of rehydration, a sharp decrease in Tm was observed, after which the Tm plateaued at 9.0 ± 2.0 °C. Note that the Tm of the undried control cells was 8.4 ± 1.7 °C; there was no statistically significant difference in Tm of undried control cells and cells after 33 hours of rehydration in LB (P = 0.6200). FTIR analysis was continued for up to 72 hours after rehydration, showing only slight variations in Tm between 33 and 72 hours. CFU analysis revealed no significant increase for up to 26 hours of rehydration in LB (Fig. 4B). However, beyond 26 hours, the CFUs

2432

C.M. Scherber et al. / Biochimica et Biophysica Acta 1788 (2009) 2427–2435

Fig. 4. Recovery and Tm changes of E. coli cell samples as a result of rehydration time in LB. (A) Variation of Tm with rehydration time. (B) Percent recovery change with rehydration time as measured by CFU. (C) Normalized OD650 as a function of rehydration time. (D) Change of CH2 peak position at 1 °C with rehydration time. Each individual point represents a separate cell sample. Independent experiments are indicated by different symbols. (●) Tm and CH2 peak position at 1 °C were found for undried control samples for comparison. The dashed lines are a guide for the eye.

began to increase exponentially. By 44 hours, there was an average percent increase of approximately 4600-fold when compared to the beginning of rehydration. In one experiment, cell recovery was measured up to 72 hours after rehydration, and the CFUs increased an additional 3.5 times between 48 and 72 hours. Since LB supports culture growth, the significant increases in CFUs (i.e. greater than 100%) that were observed after about 30 hours of rehydration were due to cell proliferation during rehydration in the LB. This was clearly reflected in the increase in OD650 observed after about 30 hours of rehydration (Fig. 4C). As with the PB rehydration experiments, the wave number of the νsCH2 band at T = 1 °C, a randomly selected temperature, was plotted for each LB rehydration time point to examine changes in membrane fluidity (Fig. 4D). With progressive rehydration time, the membrane fluidity increased until approximately 33 hours, at which point the νsCH2 peak plateaued. Unlike rehydration in PB, which resulted in membrane fluidity above that of undried cells, samples rehydrated in LB achieved maximum membrane fluidity similar to that of undried cells. 3.4. Fatty acid analysis Gas chromatographic FAME analysis showed that the percent composition of saturated and unsaturated FAs, short- and long-chain FAs, and cyclopropyl and their monoenoic precursors were all affected by desiccation and rehydration (Fig. 5). In undried control samples, the most prevalent FAs were 16:0, 16:1 (and its carbon substituted 17:0 cyclopropyl form), and 18:1. Following desiccation, the proportion of 16:0, 17:0 cyclopropane, and 19:0 cyclopropane FAs increased

(time zero in Fig. 5A, C, and E). Comparatively, the proportions of the 16:1 and 18:1 monoenoic precursors to the cyclopropyl FAs decreased (time zero in Figs. 5B and D). No change was detected amidst the short-chain FAs (12:0 and 14:0) or substituted 14:0 OH during desiccation (data not shown), all of which are characteristic components of the lipid A portion of LPS in gram-negative bacteria [37,38]. With rehydration in PB, there was an increase in 16:1 and 18:1 FAs (Fig. 5B and D), along with a decrease in their 17:0 and 19:0 cyclopropyl counterparts (Fig. 5C and E). Additionally, there was a decrease in percent composition of 16:0 (Fig. 5A), while an increase was observed in the short-chain 12:0, 14:0, and 14:0 3OH FAs (data not shown). During rehydration in the presence of Cm, no increase in unsaturated FAs (Fig. 5B and D) and no decrease in saturated FAs (Fig. 5A, C, and E) occurred. This suggests that protein synthesis is needed for the change in lipid composition observed during rehydration. In summary, FAME analysis revealed that desiccation resulted in an increased percent composition of saturated FAs and a decrease in unsaturated FAs, while during rehydration there was an increased percent composition of unsaturated FAs and FAs of the lipid A portion of LPS and a decreased percent composition of saturated FAs. 4. Discussion 4.1. Desiccation As desiccation and vitrification are being increasingly utilized for preservation of biological organisms, the need to understand the

C.M. Scherber et al. / Biochimica et Biophysica Acta 1788 (2009) 2427–2435

2433

Fig. 5. FAME direct analysis of E. coli samples rehydrated in PB. The percent composition of specific FAs was determined in cell samples at different times during rehydration in PB: (A) 16:0, (B) 16:1w7c/16:1 w6c combination, (C) 17:0 cyclopropane,( D) 18:1 w7c, (E) 19:0 cyclopropane. Fatty acid composition was measured in (●) undried control cells for comparison. (○) Cm was added during rehydration in specific samples to inhibit protein synthesis, and FAME analysis was carried out after 72 hours of incubation. The dashed lines are a guide for the eye.

mechanism(s) of cellular damage due to desiccation and rehydration is increasing [2]. During desiccation, cell membranes transition from the liquid crystalline (Lα) to the gel (Lβ) phase due to an increase in the van der Waals attractions between the hydrophobic acyl chains of neighboring lipid molecules. This results in an increase in membrane lipid packing [6,7]. Lipid packing has been quantitatively assessed in pollen [30,31], bacteria [1–3], yeast [39], and mammalian cells [28,32] by using FTIR spectroscopy to find the membrane phase transition temperature (Tm) between the Lα and Lβ phases. These studies have shown that both desiccation and osmotic stress lead to an increase in Tm. In our experiments, we have also observed that Tm of E. coli increased gradually with desiccation (Fig. 2). FTIR measurement of the Tm values also showed that Tm values of the samples after 1 week of desiccation at 20%–40% RH were similar to the Tm values measured in samples after desiccation for 1 hour at 5% RH and 4 hours at 70% RH. This indicated that water removal from the membrane caused a gradual increase in Tm until a maximum value was reached, at which point no further significant change in Tm occurred. These results also showed that the rate of desiccation did not have an effect on Tm, indicating that E. coli does not have an active response mechanism to desiccation (as opposed to what is

seen in anhydrobiotic organisms). Similar observations indicating a gradual increase in Tm have been noted during desiccation of pollen grains and dipalmitoyl phosphatidylcholine model systems [16] and during osmotic stress of E. coli [1]. Examination of compositional changes in the bacterial membranes (determined by FAME analysis) showed that a correlation exists between the percent composition of saturated and unsaturated fatty acids (FAs) and physiological changes of the membrane (as determined by changes in Tm). FA analysis showed that desiccated E. coli cells had a greater percent composition of saturated membrane FAs (Fig. 5A, C, and E) and less unsaturated FAs (Fig. 5B and D) when compared to undried control samples. These changes in membrane acyl chain composition were only observed for FAs of chain lengths greater than 14 carbons. Studies have shown that bacteria regulate membrane FA composition to adapt to variations in growth temperatures [5,40–42] and hyperosmotic conditions [43]. In a particular study conducted by Morein et al. [42], E. coli were shown to increase the fraction of unsaturated FAs in their membrane acyl chains when grown at lower temperatures in order to maintain desired membrane volume and fluidity. In contrast, when our E. coli cells were exposed to desiccation stress, there was an increase in the

2434

C.M. Scherber et al. / Biochimica et Biophysica Acta 1788 (2009) 2427–2435

Fig. 6. Scanning electron microscopy (SEM) imaging was performed on E. coli cell samples in PB: (A) undried and (B) desiccated for 1 week at 20%–40% RH using a Hitachi S-900 FESEM (Hitachi Co., Lawrenceville, GA) scanning electron microscope. The E. coli samples were sputtered with tungsten at rate of 1 Å/min for 10 min and were directly mounted on the microscope.

percent composition of saturated FAs, which may be the result of membrane shedding rather than a mechanism to maintain desired membrane fluidity. SEM images (Fig. 6) taken after desiccation but before rehydration showed that membrane shedding during desiccation might account for the variations in Tm and FA composition that were observed. Note the decrease in membrane tortuosity and the presence of round vesicles surrounding the desiccated-rehydrated cell (Fig. 6B). Bleb and vesicle formation under stressful conditions such as temperature variations, nutrient shortages, and toxic substances have been observed in E. coli and Salmonella enterica [44,45], and Katsui et al. found that heatinduced bleb formation in E. coli occurred without substantial cell death. It has been proposed that, as a stress response, bacteria release portions of the outer membrane to enhance their survival by eliminating unwanted cellular substances that have been damaged and to remodel their membrane [45]. In our desiccation experiments, the greater percent composition of saturated FAs in dried cells compared to the undried control samples may be the result of the selective removal of membrane regions highly concentrated with unsaturated FAs, possibly due to facilitated bleb or vesicle formation in more fluid-like regions. Therefore, we believe that rather than an active increase in the relative amount of saturated FAs by the cells during desiccation, the change in FA membrane composition may simply be the result of bleb or vesicle release of regions high in unsaturated FAs. 4.2. Rehydration Rehydration of E. coli, in both PB and LB, resulted in a decrease in Tm as the desiccated cells reabsorbed extracellular water, thereby increasing membrane fluidity. However, these structural changes did not begin for approximately 1 day after initial rehydration (Figs. 3A and 4A), suggesting that cellular repair was necessary prior to the onset of changes in membrane lipid packing. Following this period of delay, the Tm decreased for several hours and eventually plateaued at values closer to the Tm of undried control samples. No further decrease observed in Tm with additional rehydration time indicated that a specific state in membrane fluidity had been reached, and additional water exposure had no further significant effects on lipid packing. The decrease in Tm during rehydration paralleled an increase in peak frequency of the νCH2 band (∼2850 cm−1) at 1 °C (Figs. 3D and 4D), which also indicated an increase in membrane fluidity. We observed that E. coli samples rehydrated over 48 hours in PB possessed

slightly higher wave numbers than undried control samples. Although it is unclear why the wave numbers after 48 hours of rehydration increased beyond undried control samples, the results suggest that physical differences in the membrane still exist between undried cells and cells that had been rehydrated for over 2 days. FAME analysis showed that progressive rehydration lead to an increase in the membrane percent composition of unsaturated FAs (Fig. 5B and D) and a decrease in the percent composition of saturated FAs (Fig. 5A, C, and E). The introduction of the double bonds of unsaturated FAs disrupts the cooperativity of the network interactions of lipids within the membrane bilayer, thereby decreasing membrane order [46]. These results suggest that the increased fraction of unsaturated FAs observed with increasing rehydration time may have contributed to the decrease in Tm and increase in membrane fluidity. Numerous studies support these findings and have identified that a greater percent of unsaturated FAs in the cell membrane can directly result in decreasing of Tm [5,47,48]. Additionally, increases in FAs belonging to the lipid A portion of LPS occurred with rehydration; however, the reason for these changes is unknown. When FAME analysis was performed on samples rehydrated in the presence of Cm, a protein synthesis-inhibiting agent [36], no increase in unsaturated FA composition or decrease in saturated FA composition was observed. This suggests that newly synthesized proteins may be needed for the elongation, desaturation, transportation, and incorporation of unsaturated FAs into the cell membrane during rehydration [49]. Based on studies that E. coli enter the VBNC state upon desiccation and recover during rehydration (Schottel, manuscript in preparation), we measured colony formation in parallel to membrane analysis. Our results showed consistency with these studies and are in support of cell induction into the VBNC state during desiccation and recovery following prolonged rehydration (Fig. 3B). Furthermore, we found that the delay period of approximately 1 day of rehydration, which preceded cell recovery, was equivalent to the delay observed prior to changes in membrane lipid structure and state (i.e. Tm, membrane fluidity, and membrane FA composition). These results suggest that membrane configuration and a cell's proliferation capabilities are closely related; however, the nature of this relationship is unclear. When protein synthesis was inhibited during rehydration, no increase in membrane fluidity or cell recovery, when compared to cells at time zero of rehydration, occurred (Fig. 3A, B, and D), indicating the importance of protein synthesis on both membrane reconfiguration and cell division.

C.M. Scherber et al. / Biochimica et Biophysica Acta 1788 (2009) 2427–2435

The results presented here identify the changes in membrane lipid packing and composition that occur throughout desiccation and rehydration of E. coli. Furthermore, we have been able to relate changes in membrane fluidity and FA composition to cell recovery during rehydration, although the exact relationship between them remains unclear. One clear finding, however, is that protein synthesis is critical for both membrane changes and cell recovery. Current preservation research employs the use of protective molecules, such as disaccharides, during desiccation [2]. We have established, however, that even in the absence of osmoprotectants, E. coli cells can recover given sufficient rehydration time. These data support previous studies that suggest excessive desiccation of certain organisms does not lead to cell death, but rather to the entry into the VBNC state. These findings show promise for the field of preservation, while they raise great concern for both food safety and sanitation. Acknowledgements This research was supported by an NSF REU supplement (CBET0807858) to A.A., an Undergraduate Research Opportunities Program (UROP) grant from University of Minnesota to C.S. and A.A., and a grant from the MN Agricultural Experiment Station to J.S. The authors thank Mr. Vishard Ragoonanan for careful reading of the manuscript and for his constructive comments, and Mr. Eduardo Reategui for his help with SEM imaging. References [1] L. Beney, Y. Mille, P. Gervais, Death of Escherichia coli during rapid and severe dehydration is related to lipid phase transition, Appl. Microbiol. Biotechnol. 65 (2004) 457–464. [2] S.B. Leslie, E. Israeli, B. Lighthart, J.H. Crowe, L.M. Crowe, Trehalose and sucrose protect both membranes and proteins in intact bacteria during drying, Appl. Environ. Microbiol. 61 (1995) 3592–3597. [3] V. Ragoonanan, J. Malsam, D.R. Bond, A. Aksan, Roles of membrane structure and phase transition on the hyperosmotic stress survival of Geobacter sulfurreducens, BBA-Biomembranes 1778 (2008) 2283–2290. [4] D. Chapman, The role of water in biomembrane structures, J. Food Eng. 22 (1994) 367–380. [5] D.L. Melchior, J.M. Steim, Thermotropic transitions in biomembranes, Annu. Rev. Biophys. Bioeng. 5 (1976) 205–238. [6] A. Tardieu, V. Luzzati, F.C. Reman, Structure and polymorphism of the hydrocarbon chains of lipids: a study of lecithin–water phases, J. Mol. Biol. 75 (1973) 711–718. [7] J.F. Nagle, Theory of the main lipid bilayer phase transition, Annu. Rev. Phys. Chem. 31 (1980) 157–196. [8] J.H. Crowe, L.M. Crowe, J.F. Carpenter, A.S. Rudolph, C.A. Wistrom, B.J. Spargo, T.J. Anchordoguy, Interactions of sugars with membranes, Biochimica et Biophysica Acta (BBA) - Reviews on Biomembranes 947 (1988) 367–384. [9] L.M. Crowe, J.H. Crowe, A. Rudolph, C. Womersley, L. Appel, Preservation of freezedried liposomes by trehalose, Arch. Biochem. Biophys. 242 (1985) 240–247. [10] L.M. Crowe, C. Womersley, J.H. Crowe, D. Reid, L. Appel, A. Rudolph, Prevention of fusion and leakage in freeze-dried liposomes by carbohydrates, Biochim. Biophys. Acta (BBA) - Biomembranes 861 (1986) 131–140. [11] B. Aricha, I. Fishov, Z. Cohen, N. Sikron, S. Pesakhov, I. Khozin-Goldberg, R. Dagan, N. Porat, Differences in membrane fluidity and fatty acid composition between phenotypic variants of Streptococcus pneumoniae, J. Bacteriol. 186 (2004) 4638–4644. [12] R. Meji 'a, M.C. Gómez-Eichelmann, M.S. Fernández, Escherichia coli membrane fluidity as detected by excimerization of dipyrenylpropane: sensitivity to the bacterial fatty acid profile, Arch. Biochem. Biophys. 368 (1999) 156–160. [13] J. Cullen, M.C. Phillips, G.G. Shipley, The effects of temperature on the composition and physical properties of the lipids of Pseudomonas fluorescens, Biochem. J. 125 (1971) 733–742. [14] J.L. Brown, T. Ross, T.A. McMeekin, P.D. Nichols, Acid habituation of Escherichia coli and the potential role of cyclopropane fatty acids in low pH tolerance, Int. J. Food Microbiol. 37 (1997) 163–173. [15] T.L. Kieft, D.B. Ringelberg, D.C. White, Changes in ester-linked phospholipid fatty acid profiles of subsurface bacteria during starvation and desiccation in a porous medium, Appl. Environ. Microbiol. 60 (1994) 3292–3299. [16] J.H. Crowe, L.M. Crowe, F.A. Hoekstra, Phase transitions and permeability changes in dry membranes during rehydration, J. Bioenerg. Biomembr. 21 (1989) 77–91. [17] L.M. Hays, J.H. Crowe, W. Wolkers, S. Rudenko, Factors affecting leakage of trapped solutes from phospholipid vesicles during thermotropic phase transitions, Cryobiology 42 (2001) 88–102. [18] J.H. Crowe, Anhydrobiosis: an unsolved problem, Am. Nat. (1971) 563–573.

2435

[19] J.D. Oliver, The Viable But Nonculturable State and Cellular Resuscitation, Microbial Biosystems: New Frontiers, Atlantic Canada Society for Microbial Ecology, Halifax, Canada, 2000, pp. 723–730. [20] H.S. Xu, N. Roberts, F.L. Singleton, R.W. Attwell, D.J. Grimes, R.R. Colwell, Survival and viability of nonculturable Escherichia coli and Vibrio cholerae in the estuarine and marine environment, Microb. Ecol. 8 (1982) 313–323. [21] J.D. Oliver, The viable but nonculturable state in bacteria, J. Microbiol. 43 (2005) 93–100. [22] G. Bogosian, P.J.L. Morris, J.P. O'Neil, A mixed culture recovery method indicates that enteric bacteria do not enter the viable but nonculturable state, Appl. Environ. Microbiol. 64 (1998) 1736–1742. [23] D.B. Kell, A.S. Kaprelyants, D.H. Weichart, C.R. Harwood, M.R. Barer, Viability and activity in readily culturable bacteria: a review and discussion of the practical issues, Antonie van Leeuwenhoek 73 (1998) 169–187. [24] C. Cacela, D.K. Hincha, Low amounts of sucrose are sufficient to depress the phase transition temperature of dry phosphatidylcholine, but not for lyoprotection of liposomes, Biophys. J. 90 (2006) 2831–2842. [25] H. Binder, Water near lipid membranes as seen by infrared spectroscopy, Eur. Biophys. J. 36 (2007) 265–279. [26] C.A. Morgan, N. Herman, P.A. White, G. Vesey, Preservation of micro-organisms by drying; a review, J. Microbiol. Methods 66 (2006) 183–193. [27] H.H. Mantsch, R.N. McElhaney, Phospholipid phase transitions in model and biological membranes as studied by infrared spectroscopy, Chem. Phys. Lipids 57 (1991) 213–226. [28] W.F. Wolkers, L.M. Crowe, N.M. Tsvetkova, F. Tablin, J.H. Crowe, In situ assessment of erythrocyte membrane properties during cold storage, Mol. Membr. Biol. 19 (2002) 59–65. [29] D.K. Hincha, M. Hagemann, Stabilization of model membranes during drying by compatible solutes involved in the stress tolerance of plants and microorganisms, Biochem. J. 383 (2004) 277. [30] F.A. Hoekstra, J.H. Crowe, L.M. Crowe, Effect of sucrose on phase behavior of membranes in intact pollen of Typha latifolia L., as measured with Fourier transform infrared spectroscopy 1, Plant Physiol. 97 (1991) 1073–1079. [31] W.F. Wolkers, F.A. Hoekstra, Aging of dry desiccation-tolerant pollen does not affect protein secondary structure, Am. Soc. Plant Biol. vol. 109 (1995) 907–915. [32] E.Z. Drobnis, L.M. Crowe, T. Berger, T.J. Anchordoguy, J.W. Overstreet, J.H. Crowe, Cold shock damage is due to lipid phase transitions in cell membranes: a demonstration using sperm as a model, J. Exp. Zool. 265 (1993) 432–437. [33] C. Yanisch-Perron, J. Vieira, J. Messing, Improved M 13 phage cloning vectors and host strains: nucleotide sequences of the M 13 mp 18 and pUC 19 vectors, Gene (Amsterdam) 33 (1985) 103–119. [34] J.L. Schottel, P.M. Orwin, C.R. Anderson, M.C. Flickinger, Spatial expression of a mercury-inducible green fluorescent protein within a nanoporous latex-based biosensor coating, J. Ind. Microbiol. Biotech. 35 (2008) 283–290. [35] F.M. Ausubel, R. Brent, R.E. Kingston, D.D. Moore, J.G. Seidman, J.A. Smith, K. Struhl, Current Protocols in Molecular Biology, vol. 1, John Wiley & Sons, Inc., Brooklyn, New York 3 1-3.10. [36] M.J. Weber, J.A. DeMoss, The inhibition by chloramphenicol of nascent protein formation in E. coli, Proc. Nat. Acad. Sci. 55 (1966) 1224–1230. [37] I.M. Helander, H.L. Alakomi, K. Latva-Kala, T. Mattila-Sandholm, I. Pol, E.J. Smid, L. G.M. Gorris, A. von Wright, Characterization of the action of selected essential oil components on gram-negative bacteria, J. Agric. Food Chem. 46 (1998) 3590–3595. [38] M. Süsskind, B. Lindner, T. Weimar, H. Brade, O. Holst, The structure of the lipopolysaccharide from Klebsiella oxytoca rough mutant R29 (O1−/K29−), Carbohydr. Res. 312 (1998) 91–95. [39] C. Laroche, H. Simonin, L. Beney, P. Gervais, Phase transitions as a function of osmotic pressure in Saccharomyces cerevisiae whole cells, membrane extracts and phospholipid mixtures, BBA-Biomembranes 1669 (2005) 8–16. [40] J.E. Cronan Jr, C.O. Rock, Biosynthesis of Membrane Lipids, Escherichia coli and Salmonella: Cellular and Molecular Biology, 2nd ed, ASM Press, Washington, DC, 1996, pp. 612–636. [41] A.G. Marr, J.L. Ingraham, Effect of temperature on the composition of fatty acids in Escherichia coli, J. Bacteriol. 84 (1962) 1260–1267. [42] S. Morein, A.S. Andersson, L. Rilfors, G. Lindblom, Wild-type Escherichia coli cells regulate the membrane lipid composition in a “window” 'between gel and nonlamellar structures, J. Biol. Chem. 271 (1996) 6801. [43] L.O. Ingram, Changes in lipid composition of Escherichia coli resulting from growth with organic solvents and with food additives, Appl. Environ. Microbiol. 33 (1977) 1233–1236. [44] N. Katsui, T. Tsuchido, R. Hiramatsu, S. Fujikawa, M. Takano, I. Shibasaki, Heatinduced blebbing and vesiculation of the outer membrane of Escherichia coli, J. Bacteriol. 151 (1982) 1523–1531. [45] A.J. McBroom, M.J. Kuehn, Release of outer membrane vesicles by gram-negative bacteria is a novel envelope stress response, Mol. Microbiol. 63 (2007) 545. [46] G.D. Fasman, C. Chemical Rubber, CRC Handbook of Biochemistry and Molecular Biology, 1975. [47] R.N. McElhaney, The use of differential scanning calorimetry and differential thermal analysis in studies of model and biological membranes, Chem. Phys. Lipids 30 (1982) 229. [48] M.C. Phillips, H. Hauser, F. Paltauf, The inter- and intra-molecular mixing of hydrocarbon chains in lecithin–water systems, Chem. Phys. Lipids 8 (1972) 127. [49] C.D. Stubbs, A.D. Smith, The modification of mammalian membrane polyunsaturated fatty acid composition in relation to membrane fluidity and function, BBAReviews on Biomembranes 779 (1984) 89–137.