Chemical Changes On, and Through, The Bacterial Envelope in

Subscriber access provided by Nottingham Trent University

Article

Chemical changes on, and through, the bacterial envelope in E. coli mutants exhibiting impaired plasmid transfer identified using time-of-flight secondary ion mass spectrometry Kelly Dimovska Nilsson, Martin Palm, James Hood, Jake Sheriff, Anne Farewell, and John Stephen Fletcher Anal. Chem., Just Accepted Manuscript • DOI: 10.1021/acs.analchem.9b02533 • Publication Date (Web): 30 Jul 2019 Downloaded from pubs.acs.org on August 14, 2019

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 8 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Analytical Chemistry

338x186mm (150 x 150 DPI)

ACS Paragon Plus Environment

Analytical Chemistry 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Chemical changes on, and through, the bacterial envelope in E. coli mutants exhibiting impaired plasmid transfer identified using timeof-flight secondary ion mass spectrometry Kelly Dimovska Nilsson1, Martin Palm,1,2 James Hood,3 Jake Sheriff,3 Anne Farewell1,2*, John S. Fletcher1* 1. Department of Chemistry and Molecular Biology, University of Gothenburg, Gothenburg, Sweden 2. Centre for Antibiotic Resistance Research, University of Gothenburg, Gothenburg, Sweden 3. School of Engineering, Newcastle University, Newcastle upon Tyne, United Kingdom ABSTRACT: Time-of-flight secondary ion mass spectrometry (ToF-SIMS) using a (CO2)6k+ gas cluster ion beam (GCIB) was used to analyze E. coli mutants previously identified as having impaired plasmid transfer capability; related to the spread of antibiotic resistance. The sub-set of mutants were selected as the mutations were expected to result in changes in the bacterial envelope composition, through the deletion of genes encoding for FabF, DapF and Lpp, where the surface sensitivity of ToF-SIMS can be most useful. Analysis of arrays of spotted bacteria allowed changes in the lipid composition of the bacteria to be elucidated using multivariate analysis and confirmed through imaging of individual ion signals. Significant changes in chemical composition were observed, including a surprising loss of cyclopropanated fatty acids in the fabF mutant where FabF is associated with the elongation of FA(16:1) to FA(18:1) and not cyclopropane formation. The ability of the GCIB to generate increased higher mass signals from biological samples allowed intact lipid A (m/z 1796) to be detected on the bacteria and, despite a 40 keV impact energy, depth profiled through the bacterial envelope along with other high mass ions including species at m/z 1820 and 2428, attributed to ECACYC, that were only observed below the surface of the bacteria and were notably absent in the depth profile of the Lpp mutant. The analysis provides new insights into the action of the specific pathways targeted in this study and paves the way for whole new avenues for the characterization of intact molecules within the bacterial envelope.

INTRODUCTION Increasing resistance to current antibiotic treatments presents a critical challenge for future healthcare. Resistance can arise from various different routes such as inappropriate prescription and sales of antibiotics, use outside the healthcare sector and bacterial intrinsic factors1. As such there is an increasing need for new ways of analyzing bacteria to gain a more complete understanding of their biochemical properties that might lead to new approaches of preventing resistance and treating resistant bacteria. One particularly dangerous route by which bacteria can acquire multi-resistance is through the direct cell-to-cell transfer of DNA by conjugation. The conjugative process in gramnegative bacteria can be divided up into three main steps: (1) Recognition of a recipient cell by the donor cell, containing the plasmid and a conjugative pilus, and the bringing of the donorrecipient pair together by retraction of the pilus. (2) A bridge, called the mating bridge, is formed between the cells (mating pair formation) and the plasmid is transferred. (3) Dissociation of the mating pair after transfer and replication of the plasmid is finished2. Through conjugative transfer of plasmids, genes coding for adaptive traits that might prove beneficial under selective pressures, antibiotic resistance for example, can be gained3. The possibility to share DNA also leads to larger variation, increased recombination and faster trait fixation3.

A recent study at the University of Gothenburg has identified a range of bacterial mutants that exhibit impaired plasmid transfer ability4. Several of these genes are expected to influence the cellular envelope, either through alterations in fatty acid synthesis or the production of structurally important proteins and peptides, although how these changes affect conjugation is unknown. Time-of-flight secondary ion mass spectrometry (ToF-SIMS) is both surface sensitive and has been shown to be sensitive for the detection of lipids and is therefore well suited for probing the biochemical changes to the cellular envelope related to these mutations. In the area of bacterial analysis ToF-SIMS has been used to classify causal agents of urinary tract infection, classify strains of Bacillus, and has been shown to be particularly sensitive to secreted antibiotic and signaling molecules.5-9 Recent advances in ion beam technology have led to the introduction of gas cluster ion beams (GCIBs) that can significantly increase the sensitivity of analysis to larger biomolecules up to several kDa.10,11 while at the same time maintaining good surface sensitivity and depth resolution12. The efficacy of these beams for biological analysis has been demonstrated on various tissues and also bacterial samples where Wehrli et al. identified ppGpp dependent lipid changes related to the stringent response in E.coli.13-16.


Page 2 of 8

Page 3 of 8 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Analytical Chemistry

Figure 1. Schematic of sample preparation of E. coli on Si-wafers for SIMS analysis. The optical density (OD600) was measured to ensure that the cells had reached stationary phase. The cells were then washed in ammonium formate (0.15 M) three times. After the cells had been washed, the bacterial solutions were divided into two solutions, a washing solution (W) and analysis solution (A).

In this report ToF-SIMS, with a (CO2)6k+ GCIB, has been used to elucidate chemical differences in the cellular envelope in three E. coli mutants that have been identified as having reduced capacity for plasmid transfer: mutants lacking the genes encoding FabF, DapF and Lpp. Differences in the chemical composition of the bacterial surface were detected when imaging arrays of bacterial droplets. Additionally, for the first time, differences in both the outer membrane and periplasmic space were detected where the GCIB ToF-SIMS allowed high mass (>1000 Da.) signals from species such as Lipid A and the underlying cyclic enterobacterial common antigen (ECACYC) to be profiled and compared between mutants EXPERIMENTAL

E. coli mutants & controls Three strains of E. coli carrying different mutations, (fabF::kan/F’), (dapF::kan/F’) and (lpp::kan/F’), predicted to affect the cell membrane were analyzed. To study the changes in lipid composition caused by these mutations, three different controls were added. The first control (C3) was a strain (HA14) that contained a neutral mutation (argC::kan) as well as the F’ plasmid and was grown in medium with antibiotics (Kanamycin and Tetracycline to select for the plasmid) as were the experimental strains. A second control (C1) was included where the control strain was grown without tetracycline to control for any effects of tetracycline on the resulting lipid profiles. A third control (C2) did not carry the F’ (JW3930) to allow us to possibly identify effects that were due to the F’ plasmid.

Culturing of the E. coli The E. coli strains used in this study are listed in Table 1. Cultures were grown aerobically in LB medium in a rotary shaker at 37 °C overnight. The media was supplemented with kanamycin (Kan, 50 µg/ml) and, when appropriate, tetracycline (Tet, 10 µg/ml). Label C1

Strain HA14

C2 C3 fabF dapF lpp

JW3930 HA14 HA42 HA20 HA24

Genotype BW25113 ΔargC::KanR/F’TetR Cultured without Tet BW25113 ΔargC::KanR BW25113 ΔargC::KanR/F’TetR BW25113 ΔfabF::KanR/F’TetR BW25113 ΔdapF::KanR/F’TetR BW25113 Δlpp::KanR/F’TetR

Reference Ref. 4 PMID: 16738554 Ref. 4 Ref. 4 Ref. 4 Ref. 4

Table 1. E. coli strains included in this study and the abbreviated labels used in the figures and text in this paper.

Preparation of samples for ToF-SIMS The sample preparation was optimized during a pre-study to gain maximum consistency both within the droplet arrays of single cultures, and between different experiments. The final protocol was as follows: The optical density (OD600) was measured to ensure that the cells had reached stationary phase. The cells were washed in ammonium formate (0.15 M) three times. After washing with ammonium formate all the bacterial solutions were divided into two solutions, a washing solution and analysis solution. The pipette tip used to spot the bacterial solution onto the Siwafer was washed in the washing solution by pipetting the solution in and out three times and subsequently washed in the analysis solution by pipetting the solution in and out two times before spotting the cell solution onto the Si-wafer. Four 1 µl droplets of each strain were pipetted onto the Si-wafer. A new pipette tip was used for each droplet. The procedure is illustrated in Figure 1.

ToF-SIMS analysis In this study a J105 ToF-SIMS instrument (Ionoptika Ltd, UK) was used that circumvents some of the problems faced with conventional ToF-SIMS instruments. The J105 ToF-SIMS instrument has been described in detail elsewhere 17,18. The J105 uses a quasi-continuous primary ion beam and a buncher that focuses the secondary ions before they are injected into a reflectron time-of-flight analyzer. The use of a continuous primary ion beam can lead to shorter acquisition time and the buncher decouples the ionization process from the mass resolution leading to consistent mass resolution and improved mass accuracy. These advents open up the possibility of obtaining images and depth profiles with high mass resolution within a practical time frame. In this study a 40 keV GCIB of (CO2)6k+ was used as primary ion source (selected ± ca. 20% using an electromagnetic Wien filter). Data was collected over the mass range m/z 80-2500 in positive and negative ion mode. Here we will focus on the data recorded in negative ion mode. To ensure the reproducibility of the results, cells were cultured and analyzed on 3 different occasions over several months. The primary ion dose was kept well under the traditional static limit for all experiments. The primary ion dose for Experiment 1 was 1.64 × 1011 ions/cm2, for Experiment 2 the primary ion dose was 1.23 × 1011 ions/cm2 and for


2


Experiment 3 the primary ion dose density was 1.13 × 1011 ions/cm2. The depth profiling experiment had an accumulated primary ion dose density of 1.13 × 1013 ions/cm2 over an analysis area of 1 × 1 mm2. Low energy (10 eV) electron flooding was employed for charge compensation as required. For reference these ion dose densities would correspond to erosion depths of approximately 0.5-1 nm for experiments 1-3 and 50 nm for the depth profile analysis based on sputter rate measurements made previously on Irganox 1010.19

Helium Ion Microscopy (HIM) To check the overall shape of the bacteria, in order to discern if any of the chemical changes were a result of large structural changes such as elongation or increased curvature of the individual bacterium, dried bacteria samples of the control and mutant bacteria were platinum coated using an Emitech K550X sputter coater and shipped to Newcastle University for analysis. HIM measurements were made using a Zeiss Orion NanoFab (Carl Zeiss AG, Germany), utilizing a 25 keV He ion beam with beam current of 0.2 pA.

MAF Analysis

Page 4 of 8

Figure 2 shows example results from one of the biological replicate analyses. The bacterial droplets were pipetted on to the silicon substrate to produce 6 columns of 4 droplets from each bacterial suspension. MAF scores images are presented using a red/black/green color scale with positive scoring pixels colored green, negative in red, and pixels with little or no variance on the specified factor in black. The MAF result provides an initial guide as to which peaks showed the most characteristic changes between the different bacterial isolates while inspection of single ion images from the raw data were used to confirm peaks from the loadings as useful markers. MAF 1 and 2 (Figure 2 a,e) easily identified large differences between the lpp and fabF mutants with clear peaks from the loadings which enabled production of single ion images confirming the captured chemical variance. However, while MAF could classify the dapF mutant bacteria (Figure 2c) no dominant individual ion signals associated with this phenotype were found in the loadings and original data that could also show this trend indicating 1) the power of multivariate statistics such as MAF for classification and 2) that the biochemical changes to the cell envelope in this mutant are quite subtle.

Maximum autocorrelation factor analysis was performed as described elsewhere.20 Raw data was exported fom the vendors software as an hdf5 array file and loaded into Matlab. To reduce the memory requirement for the analysis data was down sample in time (mass) to 8 ns bins. MAF analysis was performed on data using the mass range m/z 100-2000 and m/z 500-1000 where the latter was selected to remove the very strong contributions from the intense fatty acid ions. Scores images for each factor are presented using a red/black/green color scale where red colored pixels score negatively, green score positively and black exhibit no variance associated with the specified factor. MAF loadings must be inverted prior to inspection and interpretation.

RESULTS AND DISCUSSION Surface Analysis of Bacterial Arrays To investigate the effects of the mutations on the cellular envelope of E. coli the cells were analyzed using ToF-SIMS. Maximum autocorrelation factor (MAF) analysis was then used to identify the largest differences within the data set and, hence the differences between the different strains. Multivariate analysis is a popular method for the analysis of complex mass spectral data, especially from biological samples. MAF offers advantages over, more widely used, PCA where optimization of data scaling is normally required. Scaling is not relevant to MAF analysis as the noise(/data) structure of the adjacent pixel is used.21 MAF is more computationally demanding then PCA however and can only be applied to image data. Additional limitations arise if the pixels are large such that adjacent pixels are not chemically similar or the number of pixels is less than the number of mass channels. MAF has been successfully applied to imaging MS data from cancer and infarcted heart tissue and cultured cells.15,16,20,22 Previously multivariate analysis of bacterial spectra from SIMS analysis has involved the generation of multiple spectra (technical) replicates from multiple drops of bacteria.5 Here we perform multivariate analysis on a large area (18 × 11 mm2) image of all the bacterial droplets. Having all the data contained in an image provides a means of identifying if any potential biomarkers arise from drying artefacts or from possible contamination on the silicon substrate instead of from the bacterial sample surface.

Figure 2. a: Scores image from MAF analysis of mass interval m/z 100-2000. c & e: Scores image from MAF analysis of mass interval m/z 500-1000. b, d & f: Single ion images of FA(cp19:0), PG(32:1) and PG(cp33:0) at m/z 295 (Δ 3.1 ppm), m/z 719 (Δ 9.0 ppm) and m/z 733 (Δ 5.2 ppm) respectively. These are species that were identified through MAF analysis. C1: WT(-TET), C2: WT(-F), C3: WT(+TET+F). Green: Positively scoring pixels, Red: Negatively scoring pixels, Black: No variance. Scale bar in e is 2 mm.

FabF FabF (β-ketoacyl-ACP synthase II) is the enzyme responsible for elongation of palmitoleic acid (FA(16:1)) to cis-vaccenic acid (FA(18:1)) in the type II fatty acid synthetic pathway, the principal route for bacterial membrane phospholipid acyl chain synthesis 23-25. Figure 3 illustrates how the relative signal of three selected fatty acid signals vary in the SIMS experiments. Images of three biological replicate bacterial arrays are shown (Figure 3a) along with the relative signals from the different FA ions in the ToF-SIMS spectra (Figure 3b, c). The palmitoleic precursor lipid FA(16:1) is displayed by imaging the m/z 253 ([M-H]- ion) along with the elongated form, FA(18:1) at m/z 281 and a cyclopropanated fatty acid, FA(cp19:0), at


3

Page 5 of 8 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Analytical Chemistry m/z 295. Knocking out FabF would be expected to result in a reduction in the FA(18:1) signal at m/z 281 due to FabF´s

involvement in elongation of FA(16:1) to FA(18:1), but this does

Figure 3. a: Single ion images of (FA(16:1), FA(18:1) and FA(cp19:0) at m/z 253 (Δ 11 ppm), 281 (Δ 0.92 ppm), and 295 (Δ 3.1 ppm), respectively, in controls and mutants, b: Bar graph showing the average relative intensity of controls and mutants for m/z 253 and 281, c: Bar graph showing the average relative intensity of controls and mutants for m/z 253, 281 and 295. Error bars are of the standard error of the mean (SEM) n=3.

not appear to be evident in the SIMS data. This is likely explained by two things. 1) the cells were grown at 37°C and FabF is most involved in production of 18:1 in response to growth at lower temperatures. 2) In addition, E. coli has a second Fatty acid synthase (FabB) which has been shown to be able to make 18:1 from 16:1 though not with the same efficiency 26. We propose that at 37°C the low level of 18:1 present in the membrane can be produced by FabB. However, there is a dramatic reduction in signal intensity of the FA(cp19:0) in the fabF mutant which suggests that there is a direct link between FabF and production of cyclopropanated fatty acids. The same trend is observed in lipids that have been putatively assigned to be partly consisting of cyclopropanated fatty acids specifically a cp19:0 chain detected at m/z 730.6, PE(16:0)/(cp19:0),

and

m/z 761.6,

PG(16:0)/(cp19:0)

(Supplementary Figure 1) 27. Cyclopropanated lipids are associated with high curvature membranes and their increase has been associated with stress responses in E. coli. In a previous study in our laboratory we observed an increase in cyclopropane containing lipids in carbon starved bacteria that was dependent on the presence of the global stress signaling compound ppGpp 13. This increase in cyclopropane formation had previously been reported in GCMS analysis but the SIMS experiments also suggested that the process occurred via the initial formation of the unsaturated fatty acids prior to cyclopropanation. These new results, showing a strong reduction in FA(cp19:0) upon loss of FabF, suggest that the FabF route to the formation of the FA(18:1) is strongly linked to the cyclopropane formation. In our SIMS analysis the position of the double bond in the FA(18:1) or the propane ring on the FA(cp19:0) could not be determined but the data suggests that the FA(cp19:0) may predominantly arise from the cyclopropanation of the omega-7 vaccenic acid as opposed to other FA(18:1) isomers such as the omega-9 oleic acid. Another observation that was made was an increase in signal intensity from the fabF and the lpp mutant relative to the three controls in some high mass species that have been putatively assigned as short chained cardiolipins, namely m/z

1348, 1362, 1376, and 1390, and m/z 1348, 1362, 1376, 1390, and 1404 respectively Supplementary Figure 2). On the other hand, a decrease in signal intensity relative to the three controls can be observed in the fabF for m/z 1404 and the dapF for m/z 1348, 1362, and 1390.

DapF dapF is the gene coding for diaminopimelate (DAP) epimerase, the enzyme responsible for conversion of LL-DAP to mesoDAP 28,29. Meso-DAP is in turn the precursor of lysine and an important component of the peptidoglycan that is located between the inner and outer membranes of the gram-negative E. coli. Meso-DAP crosslinks subunits in the peptidoglycan and thereby strengthens it 30. Mutations in dapF result in longer generation time and incorporation of LL-DAP into the peptidoglycan as a consequence of accumulation of LL-DAP 31. Further, within the peptidoglycan the meso-DAP provides the anchor point for Lpp 32. Due to the close relationship between DapF and Lpp the expectation was that knocking out the genes coding for these proteins would result in largely similar changes in the cellular envelopes of those strains. This was not observed in this study. With the exception of a decrease in cardiolipins at m/z 1348, 1362, and 1390 (Supplementary Figure 2) only subtle changes were found in the dapF mutant while large differences were found in the lpp mutant suggesting that knocking out dapF does not have the same effect on the cellular envelope as knocking out lpp, at least as far as the ToF-SIMS analysis in negative ion mode could detect.

Lpp Lpp, also called Braun´s lipoprotein and murein lipoprotein, is the most abundant (lipo)protein present in E. coli. 33-35. Lpp tethers the outer membrane to the peptidoglycan and is therefore important for membrane stability 32,36-38. As can be seen in Figure 2e & f, variations in the secondary ion signals from the lpp mutant are readily captured in Factor 1 of the MAF analysis. The variation in individual fatty acid ions is also very clear in Figure 3 where the lpp mutant samples show a


4


particularly strong increase in cyclopropane containing fatty acids. A possible explanation could be the more rounded shape that cells take when they adapt to withstand the new harsher

Page 6 of 8

environment that is associated with stationary phase. Increased cyclopropane

Figure 4. a: Signal normalized depth profiles of PG(cp33:0) at m/z 733.5 (Δ 5.2 ppm), lipid A at m/z 1796.2 (Δ 11 ppm) and EACCYC at m/z 2427.9 (41 pmm) showing the different signal profiles of these species in C3. b: Depth profile of m/z 2428 illustrating the difference in signal comparing C3, fabF and dapF to lpp. c: Excerpt of the mass spectrum from analysis of C3 showing the isotopic mass and isotopic pattern of lipid A m/z 1796 (Δ 11 ppm) highlighted with a blue circle, m/z 1820 (suspected ECACYC fragment ion) and surrounding species. d: Excerpt of the mass spectrum from the analysis of C3 showing the isotopic mass and isotopic pattern of m/z 2428 putatively assigned as ECACYC.

containing lipids has been suggested to assist in the formation of the more curved membrane associated with this physiologic state and the increase in cyclopropanated fatty acids observed in the lpp mutant compared to the controls and the other mutants (Figure 3a) could indicate that there is a change in morphology of the lpp mutant cells. In order to verify whether the lipid changes detected in the SIMS were resulted in gross morphological differences between the different mutants helium ion microscopy (HIM) was performed. In the HIM images (Supplementary Figure 3) it can be observed that the lpp mutant bacteria and the three controls appear have the same rounded shape while the fabF and dapF mutants on the other hand appears to be more elongated. This suggests that the increase in the amount of cyclopropanated fatty acids in the lpp mutant must be a result of additional biochemical changes caused by the mutation in the Lpp strain besides those related to morphological changes. An observation that was made during data analysis was an overall increase signal intensity from the lpp mutant, some of which was in the cardiolipin region of the spectrum as mentioned previously (Supplementary Figure 1). Mutations in lpp are known to cause leakage of periplasmic proteins 39-41 and the leaky character of this mutant could explain the overall higher intensity from the lpp mutant. However, unpublished results (Farewell, A., in prep.) indicate that dapF mutants are also leaky and therefore this is not a general feature of membrane leaky mutants.

Cell Envelope Depth Profiling Compared with traditional ToF-SIMS spectra using either mono-atomic, cluster or even C60 projectiles, the GCIB spectra acquired in this study contained significant signals from higher mass species up to several thousand m/z (Figure 4). In order to establish if the changes that were detected in the ToF-SIMS data were localized to the outer membrane, or representative of the bacterial mutant as a whole, depth profiling experiments were performed on the control and mutant samples. Accumulated primary ion dose density of approximately 1 × 1013 ions/cm2. While the erosion rate through these bacteria is not yet known, as a reference this would correspond to an erosion depth of 50 nm of Irganox1010, a commonly used reference sample for ToF-SIMS analysis and depth profiling 12,19. Initial inspection of the depth profile data indicated that the changes to the fatty acids were in part surface specific with the increased intensity of the FA(cp19:0) on the lpp strain rapidly dropping to a similar steady state signal level to the other mutants and control sample, a similar trend was observed for the intact phospholipids including the m/z 733.5 PG(cp33:0) shown in Figure 4a and additionally in Supplementary Figure 4. Further interrogation of the data produced additional remarkable results with several high mass ions appearing below the surface of the bacteria. One such ion at m/z 1796.2 is putatively assigned (based on mass deficit and isotopic pattern) as lipid A. In the same region of the spectrum a peak is found at m/z 1778.2, possibly the [M-H2O-H]- of Lipid A, and at m/z 1877.2 possibly from the inclusion of an additional HPO3.


5

Page 7 of 8 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Analytical Chemistry Figure 5a shows how the lipid A signal decreases, after a small increase, during the depth profile analysis. This is expected since lipid A is the anchoring point of lipopolysaccharide (LPS), an endotoxin, which makes up the outer leaflet of the outer membrane in gram negative bacteria 33,34. In addition, high mass ions at m/z 1820 and m/z 2428, not present in the surface SIMS spectra, appeared during the depth profile (Figure 4a). The new species at m/z 1820 and m/z 2428 exhibited a significantly different deviation from the nominal mass compared to the lipid A species (0.7 and 0.9 versus 0.2 respectively) indicating a substantially different elemental composition compared with the hydrocarbon rich (while also sugar containing) lipid A. In ToF-SIMS analysis it is often assumed that higher mass (< m/z 300) species are intact chemicals however matrix assisted laser desorption ionization (MALDI) MS analysis, that produces predominantly molecular/pseudo-molecular ions, produced a spectrum with only the lipid A and the m/z 2428 ion present in this region of the spectrum (data not shown) suggesting that the m/z 1820 peak may be a fragment of the m/z 2428 moiety. Figure 4c,d show excerpts of the SIMS spectrum accumulated during the depth profile that include the peaks at m/z 1820 and m/z 2428 that were absent from the surface analysis spectra. Intriguingly the depth profiles of the different controls and mutants show differences in the signals at m/z 1820 and m/z 2428. The result for m/z 2428 is shown in Figure 4b. The intensity of this species increases during the initial stages of the depth profiles of all the bacteria, with the exception of the lpp mutant where the peak does not appear. Hence, the membrane depth profiling provides additional information about chemical differences between this particular mutant and the other bacteria used in the study. Peaks at m/z 2428 have been reported in the literature in MALDI-ToF-MS analysis of cyclic enterobacterial common antigen (ECACYC) 42. Studies have shown ECACYC to be localized to the periplasm, explaining its appearance below the Lipid A signal, and has recently been reported to be intimately involved in maintaining the outer membrane permeability barrier. We are not aware of any previous link between the lpp mutation and an absence of ECACYC. As mentioned above, mutations in lpp are known to cause leakage of periplasmic proteins, this could possibly explain the lack of ECACYC illustrated in Figure 4b. However, it is intriguing that a mutant known for its leaky membrane appears to be missing a chemical associated with membrane permeability regulation. The variation in signal intensity of phospholipid, lipid A and ECACYC, shown in Figure 4a serve as potential markers as to the structure of the bacterial envelope indicating the outer lipid membrane and the periplasm and show similar trends, in the different bacteria analyzed in this study (Supplementary Figure 4).

CONCLUSIONS Due to the large quantity of data generated by the new GCIBs in these experiments it was a great challenge to make sense of and extract the most valuable information and there are most likely plenty of findings yet to be uncovered. Multivariate analysis (MAF in this manuscript) can aid in data interpretation and bacterial classification by reducing the dimensionality of the data and classifying the bacteria based on subtle changes over multiple mass channels when intense diagnostic ions are not present in the spectrum. We have focused on lipid changes that were in part predicted by the mutation of the bacteria but

have produced results that, while in some cases match the biosynthetic hypothesis based on the known action of the knocked out enzymes, have also produced surprising changes in the bacterial envelope chemistry that challenge some of the accepted microbiological understanding. This study shows that ToF-SIMS really is a powerful technique for analysis of the surface and subsurface of the E. coli cellular envelope. Of the three mutants analyzed here the largest differences were observed in the lpp mutant which overall showed an increased signal intensity compared to the controls. Interestingly, we could detect the loss of ECACYC in this leaky mutant strain which was recently shown to be involved in membrane permeability regulation but had not been associated with lpp 41. Due to the involvement of FabF in the elongation of palmitoleic acid (FA(16:1)) to cis-vaccenic acid (FA(18:1)) a decrease in signal intensity of FA(18:1) relative to the controls was expected, this was not observed in this study. Instead a decrease in relative signal intensity was observed in the cyclopropanated fatty acid FA(cp19:0) and in lipids that are believed to incorporate FA(cp19:0). Recent advents of GCIB technology coupled to the J105 ToFSIMS instrument made it possible to not only detect but also depth profile high mass species such as lipid A and EACCYC. The depth profiles revealed molecular profiles correlating to the expected profile of the E. coli cellular envelope. While molecular depth profiling has been accepted as a general capability of SIMS when etching is performed with GCIBs, and in some cases C60 beams, the ability to uncover intact molecules over 2 kDa represents a significant shift in the expectations for molecular SIMS analysis. Optimization of the experimental approach and accurate measurement of the erosions rates through bacterial membranes may open up a new analytical area for studying biochemical interactions on and in the bacterial envelope with relevance to pathogenicity and antibiotic resistance.

ASSOCIATED CONTENT Supporting Information Supporting information is provided in PDF format containing additional secondary ion images of lipids suspected to contain FA(cp19:0), bar graphs of the relative changes in intensity of cardiolipin ions, helium ion microscope (HIM) images of the different bacteria and depth profiles of m/z 733, m/z 1796 and m/z 2428 for each bacterial sample.

AUTHOR INFORMATION Corresponding Author * MS analysis: [email protected] * Microbiology: [email protected]

ACKNOWLEDGMENTS The authors gratefully acknowledge funding from the Swedish Research Council (VR), the University of Gothenburg And the Centre for Antibiotic Resistance Research (CARe) at the University of Gothenburg MALDI analysis, data not shown, was performed by Ibrahim Kaya, University of Gothenburg. We acknowledge Nguyen Duc Khanh Tho for assistance with the platinum coating of the samples for HIM.

REFERENCES


6


(1) Marston, H. D.; Dixon, D. M.; Knisely, J. M.; Palmore, T. N.; Fauci, A. S. JAMA 2016, 316, 1193-1204. (2) Arutyunov, D.; Frost, L. S. Plasmid 2013, 70, 18-32. (3) Sorensen, S. J.; Bailey, M.; Hansen, L. H.; Kroer, N.; Wuertz, S. Nat Rev Microbiol 2005, 3, 700-710. (4) Alalam, H.; Graf, F. E.; Palm, M.; Abadikhah, M.; Zackrisson, M.; Mattsson, M.; Hadjineophytou, C.; Persson, L.; Stenberg, S.; Ghiaci, P.; Sunnerhagen, P.; Warringer, J.; Farewell, A. bioRxiv 2018, 271254. (5) Fletcher, J. S.; Henderson, A.; Jarvis, R. M.; Lockyer, N. P.; Vickerman, J. C.; Goodacre, R. Appl. Surf. Sci. 2006, 252, 6869-6874. (6) Thompson, C. E.; Ellis, J.; Fletcher, J. S.; Goodacre, R.; Henderson, A.; Lockyer, N. P.; Vickerman, J. C. Appl. Surf. Sci. 2006, 252, 6719-6722. (7) Wehrli, P. M.; Lindberg, E.; Angerer, T. B.; Wold, A. E.; Gottfries, J.; Fletcher, J. S. Surf. Interface Anal. 2014, 46, 173-176. (8) Morales-Soto, N.; Dunham, S. J. B.; Baig, N. F.; Ellis, J. F.; Madukoma, C. S.; Bohn, P. W.; Sweedler, J. V.; Shrout, J. D. 2018, 293, 9544-9552. (9) Dunham, S. J. B.; Ellis, J. F.; Baig, N. F.; Morales-Soto, N.; Cao, T.; Shrout, J. D.; Bohn, P. W.; Sweedler, J. V. Anal. Chem. 2018, 90, 5654-5663. (10) Angerer, T. B.; Blenkinsopp, P.; Fletcher, J. S. Int. J. Mass Spectrom. 2015, 377, 591-598. (11) Tian, H.; Sparvero, L. J.; Amoscato, A. A.; Bloom, A.; Bayir, H.; Kagan, V. E.; Winograd, N. Anal. Chem. 2017, 89, 4611-4619. (12) Shard, A. G.; Havelund, R.; Spencer, S. J.; Gilmore, I. S.; Alexander, M. R.; Angerer, T. B.; Aoyagi, S.; Barnes, J. P.; Benayad, A.; Bernasik, A.; Ceccone, G.; Counsell, J. D. P.; Deeks, C.; Fletcher, J. S.; Graham, D. J.; Heuser, C.; Lee, T. G.; Marie, C.; Marzec, M. M.; Mishra, G.; Rading, D.; Renault, O.; Scurr, D. J.; Shon, H. K.; Spampinato, V.; Tian, H.; Wang, F. Y.; Winograd, N.; Wu, K.; Wucher, A.; Zhou, Y. F.; Zhu, Z. H.; Cristaudo, V.; Poleunis, C. J Phys Chem B 2015, 119, 14337-14337. (13) Wehrli, P. M.; Angerer, T. B.; Farewell, A.; Fletcher, J. S.; Gottfries, J. Anal. Chem. 2016, 88, 8680-8688. (14) Angerer, T. B.; Magnusson, Y.; Landberg, G.; Fletcher, J. S. Anal. Chem. 2016, 88, 11946-11954. (15) Sämfors, S.; Ståhlman, M.; Klevstig, M.; Borén, J.; Fletcher, J. S. Int. J. Mass Spectrom. 2017. (16) Munem, M.; Zaar, O.; Nilsson, K. D.; Neittaanmaki, N.; Paoli, J.; Fletcher, J. S. Biointerphases 2018, 13. (17) Fletcher, J. S.; Rabbani, S.; Henderson, A.; Blenkinsopp, P.; Thompson, S. P.; Lockyer, N. P.; Vickerman, J. C. Anal. Chem. 2008, 80, 9058-9064. (18) Hill, R.; Blenkinsopp, P.; Thompson, S.; Vickerman, J.; Fletcher, J. S. Surf. Interface Anal. 2011, 43, 506-509.

Page 8 of 8

(19) Angerer, T. B.; Blenkinsopp, P.; Fletcher, J. S. Int. J. Mass Spectrom. 2015, 337, 591-598. (20) Henderson, A.; Fletcher, J. S.; Vickerman, J. C. Surf. Interface Anal. 2009, 41, 666-674. (21) Tyler, B. J.; Rayal, G.; Castner, D. G. Biomaterials 2007, 28, 2412-2423. (22) Tyler, B. In ToF-SIMS: Materials Analysis by Mass Spectrometry, John C. Vickerman, D. B., Ed.; IM Publications LLP and Surface Spectra Ltd: Chichester, UK, 2013. (23) Garwin, J. L.; Klages, A. L.; Cronan, J. E., Jr. J Biol Chem 1980, 255, 3263-3265. (24) Janssen, H. J.; Steinbuchel, A. Biotechnol Biofuels 2014, 7, 7. (25) White, S. W.; Zheng, J.; Zhang, Y. M.; Rock. Annu Rev Biochem 2005, 74, 791-831. (26) de Mendoza, D.; Klages Ulrich, A.; Cronan, J. E., Jr. J Biol Chem 1983, 258, 2098-2101. (27) Oursel, D.; Loutelier-Bourhis, C.; Orange, N.; Chevalier, S.; Norris, V.; Lange, C. M. Rapid Commun Mass Spectrom 2007, 21, 1721-1728. (28) Antia, M.; Hoare, D. S.; Work, E. Biochem J 1957, 65, 448459. (29) Richaud, C.; Higgins, W.; Mengin-Lecreulx, D.; Stragier, P. J Bacteriol 1987, 169, 1454-1459. (30) Desmarais, S. M.; De Pedro, M. A.; Cava, F.; Huang, K. C. Mol Microbiol 2013, 89, 1-13. (31) Mengin-Lecreulx, D.; Michaud, C.; Richaud, C.; Blanot, D.; van Heijenoort, J. J Bacteriol 1988, 170, 2031-2039. (32) Braun, V.; Wolff, H. Eur J Biochem 1970, 14, 387-&. (33) Neidhardt, F. C. Escherichia coli and Salmonella : cellular and molecular biology, 2 ed.; Washington : ASM Press, 1996. (34) Nikaido, H. In Escherichia Coli and Salmonella Cellualr and Molecualr Biology, Neidhardt, F. C., Ed.; ASM Press: Washington D.C., USA, 1997, pp 29-47. (35) Joane M. Willey, L. M. S., Christopher J. Woolverton. Prescott's Microbiology, 10 ed.; McGraw-Hill Education: New York, NY, 2017. (36) Braun, V.; Rehn, K. Eur J Biochem 1969, 10, 426-+. (37) Braun, V.; Sieglin, U. Eur J Biochem 1970, 13, 336-&. (38) Braun, V.; Bosch, V. Eur J Biochem 1972, 28, 51-+. (39) Fung, J.; MacAlister, T. J.; Rothfield, L. I. J Bacteriol 1978, 133, 1467-1471. (40) Yem, D. W.; Wu, H. C. J Bacteriol 1978, 133, 1419-1426. (41) Suzuki, H.; Nishimura, Y.; Yasuda, S.; Nishimura, A.; Yamada, M.; Hirota, Y. Mol Gen Genet 1978, 167, 1-9. (42) Mitchell, A. M.; Srikumar, T.; Silhavy, T. J. 2018, 9, e0132101318.


7

Chemical Changes On, and Through, The Bacterial Envelope in

Recommend Documents