MALDI-TOF Characterization of Protein Expression Mutation During

Subscriber access provided by UNIV OF NEW ENGLAND ARMIDALE

Article

MALDI-TOF Characterization of Protein Expression Mutation During Morphological Changes of Bacteria Under the Impact of Antibiotics Dongxue Zhang, Yi Yang, Qin Qin, Juan Xu, Bing Wang, Jianwei Chen, Baohong Liu, Weijia Zhang, and Liang Qiao Anal. Chem., Just Accepted Manuscript • DOI: 10.1021/acs.analchem.8b05080 • Publication Date (Web): 10 Jan 2019 Downloaded from http://pubs.acs.org on January 11, 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 9 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

MALDI-TOF Characterization of Protein Expression Mutation During Morphological Changes of Bacteria Under the Impact of Antibiotics Dongxue Zhang1, Yi Yang1, Qin Qin4, Juan Xu1, Bing Wang5, Jianwei Chen5, Baohong Liu1,3, Weijia Zhang* 2,3 and Liang Qiao* 1 1. Department of Chemistry, Shanghai Stomatological Hospital, Fudan University, Shanghai 200000, China. 2. Shanghai Institute of Cardiovascular Diseases, and Institutes of biomedical sciences, Zhongshan Hospital, Shanghai Key Laboratory of Medical Imaging Computing and Computer Assisted Intervention, Shanghai Medical College of Fudan University, Shanghai 200032, China 3. State Key Laboratory of Molecular Engineering of Polymers, Fudan University, Shanghai 200433, China. 4. Changhai Hospital, The Naval Military Medical University, Shanghai 200433, China. 5. BGI-Qingdao, BGI-Shenzhen, Qingdao, 266555, China ABSTRACT: Antimicrobial resistance (AMR) is one of the most serious problems affecting public health and safety. When treated by antibiotics, bacteria usually experience changes on morphology that can lead to the development of AMR. In this work, we propose a strategy to study mutation in protein expression during morphological changes of bacteria under the impact of antibiotics. The study is focused on small proteins that can be detected by matrix assisted laser desorption/ionization time-of-flight mass spectrometry (MALDI-TOF MS), which is nowadays widely used in hospitals for bacterial identification. We used a gradient microfluidic chip to observe the morphological changes of various bacteria under the impact of antibiotics. Differential peaks related with the morphological changes were firstly figured out by MALDITOF MS, then identified by considering the molecular weight of both candidate proteins and their tryptic digested peptides, and further validated by liquid chromatography tandem mass spectrometry (LC-MS/MS) based label-free quantitative proteomic method. Specifically, carbapenem-resistant Klebsiella pneumoniae (CR-KP), super extended spectrum β-lactamases Escherichia coli, Vibrio parahemolyticus from South America prawns, and carbapenem-resistant Pseudomonas aeruginosa were used as model samples to illustrate the strategy. Eight proteins closely correlated with its morphological change were identified for CR-KP. Among the eight proteins, three, i.e. fimbrial subunit type 3, Penicillin-binding protein activator LpoB and 30S ribosomal protein S14, were further verified at the transcriptome level.

Antimicrobial resistance (AMR) is a severe global problem, closely related with human morbidity, high cost on therapy and high expenditure on preventing hospital-acquired infections.1 According to the Review on Antimicrobial Resistance, antimicrobial resistant bacteria is an increasing global issue, which can result in 10 million of deaths every year and 2%~3.5% Gross Domestic Product (GDP) reduction by 2050.2 To fight against AMR, it is crucial to understand AMR from molecular level and to find out molecular biomarkers for fast diagnosis. When treated by antibiotics, bacteria usually experience changes on morphology. For example, Escherichia coli cells treated and killed by beta-lactam experience four procedures: filamentation, bulge formation, bulge stagnation and lysis.3 Filamentation is a survival way and intended response of bacteria under the impact of antibiotic drug. When the antibiotic drug is removed, bacterial cell division can recover.4 In low concentration of antibiotic, bacterial filament cells can generate resistant daughter cells, which own stronger drugresistance than the parental cells.5 The morphological changes are important for AMR development, and must be associated with mutation in protein expression and/or protein posttranslational modifications. Therefore, identifying the differential proteins from bacteria with different morphologies is essential for the mechanism study of AMR.

Microfluidics is an emerging technique that has been applied to accomplish antimicrobial susceptibility testing (AST)6,7 and the study of drug resistant response of antimicrobial resistant strains, where the morphological changes of bacterial cells can be visually observed.8,9 Specifically, microfluidic gradient platforms have been developed and applied to multiple bioanalytical studies,10,11 e.g. neutrophil chemotaxis, cancer chemotaxis, bacteria growth chemotaxis, stem cell differentiation, etc. In the platforms, multiple streams carrying different concentrations of substances or different kinds of substances are mixed to generate a range of gradient concentrations.12 One type of the gradient microfluidic chips is the “Christmas tree” microchip, which has been applied to the genotyping of single nucleotide polymorphisms (SNPs)13 and cell culture14. MALDI-TOF MS is a prominent technology for bacteria identification at species level.15-17 The analytical process is fast, convenient and suitable for clinical diagnosis.18 It has also been considered as a technology to discriminate standard and drug resistant strains because of their different representative spectral profiles.19,20 Herein, we developed a strategy using microfluidics and MALDI-TOF MS to study mutations in protein expression during morphological changes of antimicrobial resistant bacteria under the impact of antibiotics. A “Christmas tree” microfluidic chip was used to generate gradient concentrations of

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

Page 2 of 9

antibiotics to study the morphological changes of bacteria under the stimulation of different concentrations of antibiotics. The study on protein expression mutation was focused on small proteins that can be detected by MALDI-TOF MS. By comparing the MALDI-TOF MS spectra of same strains with different morphologies, differential peaks for specific morphological states could be figured out, and then identified based on the molecular weight database searching against UniProt. The identification was further confirmed by high resolution MS with label-free quantitative proteomic approaches. Carbapenem-resistant Klebsiella pneumoniae (CR-KP), super extended spectrum β-lactamases Escherichia coli (ESBLEC), Vibrio parahemolyticus 61 from South America prawns (VP-61), and carbapenem-resistant Pseudomonas aeruginosa (CR-PA) were used as model samples to illustrate the performance of the strategy. Eight proteins closely correlated with its morphological change were identified and validated from CRKP. Among the eight proteins, three, i.e. fimbrial subunit type 3, Penicillin-binding protein activator LpoB and 30S ribosomal protein S14, were further verified at the transcriptome level.

EXPERIMENTAL SECTION Chemicals and materials. Polydimethylsiloxane (PDMS) elastomer Sylgard 184 and curing agent Sylgard 184 (A/B) were from Dow Coining (USA). SU8 2025 photoresist was from Micro Chem (USA). Silicon wafer (4 inch) and developing liquid were from RDMicro (Suzhou, China). Ampicillin sodium and PBS were from Sungon Biotech (Shanghai, China). Ceftriaxone sodium was from CNW (Germany). DLDithiothreitol (DTT) was from J&K (Beijing, China). αCyano-4-hydroxycinnamic acid (HCCA), iodoacetamide (IAA), trypsin, cytochrome C, myoglobin, trifluoroacetic acid (TFA) and formic acid (FA) were from Sigma-Aldrich (St. Louis, MO). Acetonitrile (ACN, HPLC grade) was from Merck (Darmstadt, Germany). Peptides mixer for calibration was from Bruker (Germany). Nylon syringe filter (13mm*0.22 μm) was from ANPEL (Shanghai, China). Trypticase soy broth (TSB) was from Beijing Land Bridge Technology CO., LTD. (Beijing, China). Carbapenem-resistant Klebsiella pneumoniae (CR-KP), carbapenem-resistant Pseudomonas

aeruginosa (CR-PA), and super extended spectrum βlactamases Escherichia coli (ESBL-EC) were from Chang Hai hospital in Shanghai, China. Vibrio parahaemolyticus 61 (VP61) from South America prawns was from Shanghai Ocean University. Deionized (DI) water (18.2 MΩ.cm) was purified by a Smart-Q deionized water system (Hitech pure water technology, Shanghai, China) and used in all aqueous solutions M icrofluidic chip design and fabrication. A “Christmas tree” microfluidic chip (Figure 1(a)) to generate concentration gradients of antibiotic drugs and to be used for bacteria on-chip culture was designed using AutoCAD, which contained two inlets, i.e. nutrient inlet and antibiotic inlet (2000 μm in diameter), serpentine channels (100 μm in width, 25 μm in depth), AMR test cells (160 μm in diameter, 25 μm in depth), and an outlet (2500 μm in diameter, 25 μm in depth). Following standard soft lithography and molding technology,21,22 a piece of mold on a 4-inch silicon wafer was fabricated from SU8 2025 photoresist. The mold containing the structure was spin-cast at 3500 rpm for 40 s. A piece of PDMS elastomer (3.5 cm in width, 4 cm in length) made of Sylgard 184 and curing agent Sylgard 184(A/B) was casted form the mold. The PDMS layer was assembled with a piece of glass (3.5 cm in width, 7.5 cm in length) by oxygen plasma to get the microfluidic chip. O n-chip AM R test. Stock solution of ceftriaxone sodium (100 mg mL-1) was prepared with DI water and stored at −20 °C. CR-KP obtained from Changhai hospital was cultured overnight with TSB to get a suspension of CR-KP at ~108 CFU mL-1. The concentration was confirmed with OD600 characterization. Before experiment, the microfluidic chip was put in biosafety cabinet under the UV light for 30 min and washed by PBS. Bacteria cells were injected into the “Christmas tree” microchip by infusing through the two inlets at the same time with a flow rate of 200 μL h-1 each for 30min. Afterwards, TSB and diluted ceftriaxone sodium (5 mg mL-1) were continuously injected through the nutrient inlet and antibiotic inlet at the flow rate of 15 μL h-1, respectively. Morphology characterization was performed with an optical microscope LWD200-37T (CEWEI, China). ESBL-EC, VP-61 and CR-PA were also used for on-chip AMR tests with the same method.

Figure 1. (a) AutoCAD design of the “Christmas tree” microchip, (b) and (c) microscopic images of the bacterial cell culture wells, (d) microscopic image of the gradient channel.


Page 3 of 9 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 Proteins extraction and digestion for M S analysis. 50 μL 70% FA and 50 μL ACN were added to dissolve bacterial cell pellets and extract proteins from the bacterial cells.23 The solution was centrifuged at 10,000 g for 1 min to collect the supernatant. The extracted proteins were lyophilized overnight and redissolved in 50 μL 25 mM NH4HCO3 solution. The solution was treated successively with DTT (2.5 μL, 200 mM, 25 °C for 1 h), IAA (10 μL, 200 mM, 25 °C for 1 h in dark), and DTT (10 μL, 200 mM, 25 °C for 1 h) for the reduction of disulfide bond and alkylation of thiol groups. Then, 347.5 μL 25 mM NH4HCO3 was added to dilute the solution. Afterwards, 10 μL 1 mg mL-1 trypsin (from bovine pancreas) was added for protein digestion. The solution was incubated at 37 °C for 16 h. The digested peptides were kept for further MS analysis by MALDI-TOF or LC-MS/MS. D N A and RN A extraction and sequencing. Bacterial samples were thawed on ice and DNA extraction was performed using the DNA extraction kit design by BGI, China. Extracts were treated with DNase-free RNase to eliminate RNA contamination and sheared into fragments between 50 bp and 800 bp in size using a Covaris E220 ultrasonicator (Covaris, Brighton, UK). Fragments between 150 bp and 250 bp were subjected to generate a single-stranded circular DNA library. Whole genome of CR-KP was sequenced on a BGISEQ-500 platform (BGI-Qingdao, China) and generated the pair-end 100 bp sequence data. RNA was extracted using a RNeasy® Mini Kit (Qiagen, Hilden, Germany), and the quality was determined using an Agilent 2100 bioanalyzer (Agilent, Santa Clara, USA). The 23S/16S rRNA were removed using MicrobExpress kit (Ambion) and Amplification grade DNAse 1 (Invitrogen) respectively. After constructed the reverse transcribed cDNA library, it was sequenced on the BGISEQ-500 sequencer (BGI-Wuhan, China) and generated the single-end 50 bp sequence data. M A LD I-T O F M S analysis. A MALDI-TOF mass spectrometer (Bruker MicroFlex LRF) was employed to analyze the proteins and peptides from the bacterial samples. For each protein sample, 1 μL of the solution was deposited on a well of a MALDI target plate. After solvent evaporation, 1 μL 4 mg mL-1 HCCA matrix dissolved in 50% CH3CN, 49.9% H2O, 0.1% TFA was added to the same well. The dried sample with matrix was analyzed by the mass spectrometer under linear positive mode with optimized instrumental parameters (ion source 1: 20 kV, ion source 2: 18.6 kV, lens: 9.6 kV, polarity: positive, linear detector gain: 2.7 kV, pulsed ion extraction delay: 150 ns, attenuator offset of laser: 35%, attenuator range of laser: 30%, sample rate and digitizer settings: 2.00 GS/s) and m/z range of 2,000 - 20,000. For each sample, 6 duplicated spots were prepared to collect 6 mass spectra. External calibration of the equipment in the m/z range was performed using cytochrome C and myoglobin. Peptides were analyzed with the same equipment under positive reflection mode for better resolution with optimized instrumental parameters (ion source 1: 20 kV, ion source 2: 15.75 kV, lens: 9.35 kV, reflector: 20 kV, polarity: positive, reflector detector gain: 1.82 kV, pulsed ion extraction delay: 120 ns, attenuator offset of laser: 35%, attenuator range of laser: 30%, sample rate and digitizer settings: 2.00 GS/s) and m/z range of 700 – 3000. 1 μL of peptide sample was deposited on a well of a MALDI target plate. After solvent evaporation, 1 μL of the HCCA matrix was added to the same well. For each sample, 6 duplicated spots were prepared to collect 6 mass

spectra. External calibration of the equipment in the m/z range was performed using bradykinin (1-7), angiotensin I, angiotensin II, substance P, bombesin, renin, ACTH-clip, and somatostain. LC -M S/M S analysis. The peptide samples were lyophilized overnight and redissolved in DI water. The solution was desalted by C18 Spin Columns (Thermo-Fisher, USA) and lyophilized. An LC-MS/MS (DGU-20A3R LC from Shimadzu with triple TOF 4600 MS from SCIEX) system was used to analyze the peptide samples. Mobile phase A was 97.9% H2O, 2% ACN and 0.1% TFA. Mobile phase B was 97.9% ACN, 2% H2O and 0.1% TFA. 120 μL mobile phase A was used to dissolve the desalted peptide sample. The solution was then filtered by a nylon syringe filter (13 mm*0.22 μm). 15 μL of the filtered sample was injected for LC-MS/MS analysis. The LC column was a C18-WP column (CNW Athena, 2.1x150 mm, 3 um). The mobile phase gradient program was 0-70-74-75-8080.01-90 min; 10-30-50-80-80-10-10% B phase. Mass spectrometer worked under ESI positive mode with the optimized parameters: MS range: 350-1500 m/z; MS2 range: 100-1250 m/z. Source parameters: GS1, 40 psi; GS2, 40 psi; CUR, 30 psi; TEM, 300 °C; Spray voltage, 5500 V; Collision Energy, 45 ev. For each sample, one injection was performed for data dependent acquisition (DDA) for identification, and three injections for data independent acquisition (DIA) for quantification. M A LD I-T O F M S data analysis. MALDI-TOF mass spectra were processed using flexAnalysis software (version 3.0; Bruker Daltonics, Germany). The m/z range of 3-18 kDa was chose for data analysis. Peaks with a signal-to-noise ratio (S/N) of at least 3 were extracted from each spectrum after baseline correction, and exported as text files (.txt) with m/z and intensity. Six replicated peak lists of proteins extracted from rod-like CR-KP and 6 peak lists of proteins extracted from threadlike CR-KP were merged into a matrix, where samples were in columns and features in rows, using a peak alignment algorithm based on hierarchical clustering. The distance between two peaks was defined as

d=

( m z )1 − ( m z )2 max {( m z )1 , ( m z )2 }

(1)

if the two peaks were in different spectra. Otherwise, d = 1. Complete linkage was used when calculating inter-cluster distances. The hierarchical clustering tree was cut at a specified height (tolerance of 2,000 ppm in this study), and the peaks were divided into several bins to get the peak table. The peak alignment procedure was implemented with R from the R Foundation for Statistical Computing. The biomarker discovery procedure was conducted with MetaboAnalyst (version 3.0; http://www.metaboanalyst.ca).24 The data stored in the peak table were preprocessed with the following options: normalization by sum, log transformation and Pareto scaling. Partial least squares-discriminant analysis (PLS-DA) was performed and the important features were identified by the variable importance in projection (VIP) scores. Protein identification with MALDI-TOF mass spectra was implemented with R. For both proteins and digests, 6 replicated spectra were merged into a combined spectrum using the peak alignment algorithm described above. The m/z of the peaks in the combined spectrum of extracted proteins were



Page 4 of 9

compared with the reviewed entries of Klebsiella pneumoniae from UniProt (http://www.uniprot.org/) database to obtain the candidate proteins. Charges of 1+ and 2+ were considered, and the m/z tolerance was set to 2,000 ppm. The candidate proteins were in-silico digested with trypsin, and the peaks in the combined spectrum of the digests were further compared with the 1+ ions of the peptides with the m/z tolerance of 500 ppm. Candidate proteins with at least one matched peptide were reported to be further validated by LC-MS/MS. The R scripts are available at https://github.com/lmsac/BacteriaMSdifference/. LC -M S/M S data analysis. The LC-ESI-MS/MS DDA data were analyzed with ProteinPilot (version 5.01; AB SCIEX) for protein identification. FDR was set at 0.05. The DIA data were analyzed with PeakView (version 2.2; AB SCIEX) and MS/MSALL with SWATH Acquisition MicroApp (version 2.0; AB SCIEX) for protein quantification. In the PeakView software, number of peptides per protein and transitions per peptide were both set at 6. Peptide confidence threshold was 99% and false recovery rate threshold was 1%. Extracted ion chromatogram (XIC) extraction window was 3 min and XIC width was 75 ppm. The important features were identified with MetaboAnalyst via the biomarker discovery procedure described above. D N A/RN A sequencing data analysis. After filtering low quality data, adapter contamination and duplication reads by SOA SOAPnuke (v1.5.2), the CR-KP genomic DNA clean reads were assembled into contigs and scaffolds employing the de novo assembler SPAdes (v3.11.1) with a range of kmer from 41 to 61 (step size 10 bp).25 All coding sequences (CDSs) were predicted by Glimmer (V3.02),26 and functional annotation was carried out using the BLAST (v2.2.26) search tool with the non-redundant protein database (nr, v20171010) of GenBank, KEGG and SwissProt (v201707). Genomic RNA clean reads which obtained after filtering the low-quality data and adapter contamination reads using SOAPnuke (v1.5.2) were mapped to the newly assembled genomics and contigs genes using Bowtie 2 (v2.2.5). The gene expression levels were normalized to the library and the gene length by calculating the FPKM value using RSEM (v1.2.12) software.27,28 After normalization, differentially expressed genes (DEGs) under different stimulate conditions of all transcripts were quantified using EBSeq (1.4.0) by R (v3.1.1).29 Significant DEGs were screened based on an PPEE (posterior probability of being equivalent expression) threshold of ≤ 0.05, and a |log2 Fold change| value ≥ 1, and different significant DEGs were enriched and clustered according to GO and KEGG function. The data reported in the study are available in the CNGB Nucleotides Sequence Archive (CNSA: https://db.cngb.org/cnsa/; accession number CNP0000194).

RESULTS AND DISCUSSION M orphological change of A M R bacteria under the stim ulation of antibiotics in m icrofluidic chips. To study the mutation in protein expression during morphological changes of AMR bacteria under the impact of antibiotics, we have combined the microchip based on AMR tests and MS

analysis, using a protocol as illustrated in Scheme 1. The bacterial morphological changes were firstly monitored using a gradient microchip to select proper concentrations of antibiotic drugs in high throughput. Then, the bacteria were cultured in centrifugal tubes in the presence of the antibiotic drugs with the determined concentrations. Finally, proteins were extracted from the bacteria with different morphologies, and digested for MALDI-TOF MS analysis to identify protein biomarkers related with the morphological changes.

Schem e 1. The procedure to study the mutation in protein expression during morphological changes of bacteria under the impact of antibiotics. Ten gradient concentrations can be accurately generated with the microchip. If the initial concentration of antimicrobial drug was C injected from the inlet, the concentrations of the drug in the ten bacterial cell culture wells (Figure 1) would be 0, 1/9 C, 2/9 C, 3/9 C, 4/9 C, 5/9 C, 6/9 C, 7/9 C, 8/9 C and C, respectively.12 For each well, four micropillars (25 μm in diameter) were designed for building a favourable bacterial growing environment from being influenced by nutrient flows.30 Therefore, high throughput characterization of bacteria morphological changes under the stimulation of different concentrations of antimicrobial drugs can be realized with the microchip. Figure 2 shows the morphological change of CR-KP under different concentrations of ceftriaxone sodium from microcells No. 1 to 8. Concentrations of ceftriaxone sodium in the wells were 0 mg mL-1, 0.56 mg mL-1, 1.11 mg mL-1, 1.67 mg mL-1, 2.22 mg mL-1, 2.78 mg mL-1, 3.33 mg mL-1 and 3.89 mg mL-1 for wells No. 1 to 8, respectively. As shown in the figure, filamentation and lysis of CR-KP happened with the concentration increasing of ceftriaxone sodium. When the concentration of ceftriaxone sodium increased to 1.67 mg mL-1, the morphology of the cell rapidly changed from rod-like into threadlike structure. When the concentration of ceftriaxone sodium was higher than 1.67 mg mL-1, there was little morphological change, while the density of the bacterial cells decreased because of cell lysis. Counting the number of the bacterial cells was also done to obtain the minimum inhibitory concentration (MIC, the lowest concentration of antibiotic for preventing visible growth of bacterium) of CR-KP,31,32 which was between 1.67 mg mL-1 and 2.22 mg mL-1. Based on the result, 2 mg mL-1 ceftriaxone sodium was then used to stimulate the bacteria in centrifugal tubes to obtain sufficient CR-KP cells with threadlike structure for MS characterization.


Page 5 of 9 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


Figure 2. The morphological changes of CR-KP under different concentrations of ceftriaxone sodium: 1. 0 mg mL-1; 2. 0.56 mg mL-1; 3. 1.11 mg mL-1; 4. 1.67 mg mL-1; 5. 2.22 mg mL-1; 6. 2.78 mg mL-1; 7. 3.33 mg mL-1; 8. 3.89 mg mL-1. The bacterial cells are stained in red color, expect those not in the same focal plane.

The morphological changes of CR-PA, VP-61, and ESBLEC under the stimulation of ceftriaxone sodium or ampicillin sodium were also studied with the microchip as shown in supporting information Section S1. The filamentation of ESBLEC started when ampicillin sodium reached 200 μg mL-1. When ampicillin sodium concentration was as high as approximately 300 μg mL-1, the ESBL-EC elongated obviously. The filamentation of CR-PA began at 500 μg mL-1 of ceftriaxone sodium. Its resistant ability was extremely strong and didn’t lyse at 4 mg mL-1 of ceftriaxone sodium. In contrast, VP-61 was weak in terms of AMR. The whole process of VP morphological changes from filamentation to lysis could be clearly observed with 0 to 500 μg mL-1 of ampicillin sodium.

CR-KP proteins before and after stimulation by 2 mg mLceftriaxone are shown in supporting information Section S3.

1

M A LD I-T O F m ass spectra of bacteria w ith different m orphologies. Based on the characterization results using microchips, CR-KP was cultured with or without the stimulation of 2 mg mL-1 ceftriaxone sodium in centrifugal tubes. The morphological change was confirmed with microscope. Proteins were extracted from approximately 105 bacterial cells with rod-like or threadlike structures. The concentration was confirmed with OD600 characterization. The cells were quickly lysed by formic acid and acetonitrile to get the proteins with respect to the corresponding morphological states of the bacterial cells. MALDI-TOF MS was used to analyze the protein extracts. As shown in Figure 3, up-regulated peaks could be figured out at m/z = 7062, 11260, etc. for the rod-like KP, and at m/z = 7569, 10287, 11565, 11766, 15124, etc. for the threadlike KP. Similarly, CR-PA, VP-61 and ESBLEC with or without antimicrobial drug stimulation were also analyzed by MALDI-TOF MS to figure out their corresponding differential peaks, as shown in supporting information Section S2. Identification and validation of the differential peaks from M A LD I-T O F M S. The differential peaks on MALDI-TOF mass spectra were identified by matching the m/z values of the peaks for proteins and the m/z values of the peaks for peptides to the molecular weights of reviewed proteins in the UniProt database for Klebsiella pneumonia and the peptides in-silico digested from the proteins by following the rule of tryptic digestion with the maximum miscleavage sites of 2. The MALDI-TOF MS spectra for peptides digested from

Figure 3. (a) The MALDI-TOF mass spectra with the m/z range of 3,000 to 18,000 after normalizing the total peak area to 100% for proteins extracted from CR-KP in rod-like (blue) and threadlike (green) shapes; (b-g) zoomed mass spectra of the dif-



ferential peaks with the m/z of 7062, 7569, 10287, 11260, 11565, 11766, 15124.

type3,36 Penicillin-binding protein activator LpoB37, have already been reported to be related to bacterial AMR.

Figure 4 shows the identification workflow of all the differential peaks from the MALDI-TOF results shown in Figure 3. There are in-total 62 differential peaks with VIP>1 between rod-like and threadlike CR-KP identified by MALDI-TOF MS. Among them, 20 could not match any intact protein, which could be from protein fragments or proteins with unrecorded post-translational modifications. 42 matched to 160 proteins by considering 1+ or 2+ charges, and were supported by at least 1 peptide match according to MALDI-TOF results of peptides. Among the 160 proteins, 51 were also identified by LC-MS/MS in DDA mode, related with 27 differential peaks on the MALDI-TOF mass spectra. Among the 51 proteins, 10 related with 13 differential peaks on the MALDI-TOF mass spectra were also quantified by DIA LC-MS/MS and showed the same regulation in signal intensity in DIA LC-MS/MS and MALDI-TOF. Among the 10 proteins, 8 related with 11 differential peaks on the MALDI-TOF mass spectra were identified by LC-MS/MS with > 1 unique peptide, which could be considered as potential biomarkers, shown in supporting information Section S4.

We have further studied the quantity changes in mRNA during the morphological changes of CR-KP from rod-like to threadlike. The mRNAs for three proteins among the eight showed same regulation in quantity changes as the proteins, which were fimbrial subunit type 3, penicillin-binding protein activator LpoB and 30S ribosomal protein S14 as shown in supporting information Section S4. It should be noticed that the correlation between transcriptome and proteome is originally weak, i.e. with R2 = 0.41 between the protein copies per cell and mRNA copies per cell.38

Figure 4. The identification workflow of protein biomarkers correlated to the morphological changes of CR-KP from the MALDI-TOF results in Figure 3.

The eight protein biomarkers are 30S ribosomal protein S10 (ID in UniProt, B5XN93), 30S ribosomal protein S4 (ID in UniProt, B5XNB5), autonomous glycyl radical cofactor (ID in UniProt, B5XNF9), 30S ribosomal protein S6 (ID in UniProt, A6THB1), 50S ribosomal protein L32 (ID in UniProt, B5XXI1), 30S ribosomal protein S14 (ID in UniProt, B5XNA6), fimbrial subunit type 3 (ID in UniProt, P12267) and penicillin-binding protein activator LpoB (ID in UniProt, A6T7G4), respectively. Among these proteins, 30S ribosomal protein S1033,34, 30S ribosomal protein S435, fimbrial subunit

Page 6 of 9

The quantity changes of fimbrial subunit type 3 (m/z=10283) and Penicillin-binding protein activator LpoB (m/z=11278) in CR-KP cells with different drug stimulation time were further studied, since the two proteins among the three have previously been reported to closely relate with bacterial AMR. The relative peak areas normalized to total peak area for the two proteins were compared among the bacteria after 0 h, 1.5 h, 3 h and 4 h of stimulation using 2 mg mL-1 ceftriaxone sodium, as shown in Figure 5.

Figure 5. Variations in protein expression level in CR-KP cells with time under the stimulation of 2 mg mL-1 ceftriaxone sodium measured by MALDI-TOF MS. Relative peak area of (a) Fimbrial subunit type 3 and (b) Penicillin-binding protein activator LpoB changing with time. The stars marked the peaks on mass spectra corresponding to the proteins. The mass spectra were acquired from ~105 bacterial cells. The error bar represents the standard deviation from three replications.

The expression of Fimbrial subunit type 3 was increased with time as shown in Figure 5 (a). The protein is encoded by the gene MrkA, and is the major composition of type 3 fimbriae structure. MrkA facilitates biofilm formation.36,39 The pro-


Page 7 of 9 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


duction of biofilm matrix makes bacteria tolerant to antibiotics by preventing the drug molecules to pass through the biofilm.40 Consequently, up-regulation of the protein expression is reasonable when bacterial cells are stimulated by antibiotics. In Figure 5 (b), penicillin-binding protein activator LpoB exhibited an apparent descending trend with an ascending in small scale from 1.5 h to 3 h. In some bacteria, e.g. E. coli, penicillin-binding protein activator LpoB was found to activate PBP1b which is essential in peptidoglycan (PG) bio synthesis.41 PG synthesis forms a sacculus to maintain cell shape from lysis.37 Thus, more LpoB should normally be produced by bacteria to activate PBP1b for PG production from being burst in the presence of antibiotic drugs. However, it was also reported that in some pneumococci PBP1b is not involved in penicillin resistance.42 According to our findings, in CR-KP the situation is different from E. coli. The possible reason is that PBPs other than PBP1b can take effect for bacterial resistance developing in this strain.

CONCLUSIONS A strategy combining microfluidic gradient platforms and MS for protein profiling has been developed to study the mutation in protein expression during the morphological changes of bacteria stimulated by antibiotic drugs. Four drug resistant bacteria, i.e. CR-KP, ESBL-EC, CR-PA and VP-61, were studied by the strategy. Specifically, 8 protein biomarkers were identified for the morphological change of CR-KP under the stimulation of ceftriaxone. The biomarkers were also confirmed with quantitative proteomic approach by LC-MS/MS. Among the 8 biomarkers, 3, i.e. fimbrial subunit type 3, Penicillin-binding protein activator LpoB and 30S ribosomal protein S14, were verified by transcriptome analysis. Identification of protein biomarkers were focused on small proteins that could be detected by MALDI-TOF MS, since the technique holds the advantages of fast analysis, easy operation and is nowadays widely used in hospitals for clinical identification of pathogenic bacteria. For future improvements, direct conjunction of microfluidic chip and MALDI-TOF MS may be achieved by using microarrays for mass spectrometry (MAMS) 43 to analyze samples in extremely low volume at high sensitivity. MALDI-Fourier transform ion cyclotron resonance (FTICR) with ultra high resolving power 44 and MALDI Top-Down sequencing methods 45,46 can be employed to avoid protease treatment and LCMS/MS based proteomic analysis for biomarker identification. The microfluidic chip with MALDI mass spectrometry system can be valuable to the mechanism study of AMR, and is of potential value in clinical usages.

ASSOCIATED CONTENT Supporting Information: The Supporting Information is available free of charge on the ACS Publications website: morphological change of three species of bacteria under the stimulation of antibiotics in microfluidic chips; MALDI-TOF MS characterization of three species of bacterial cells with or without the stimulation of antimicrobial drugs; MALDI-TOF MS characterization of the proteins extracted from CR-KP after digestion with or with-

out the stimulation of antimicrobial drugs; information on the eight protein biomarkers identified by MALDI-TOF and LC-MS/MS from CR-KP.

AUTHOR INFORMATION Corresponding Author * Correspondence should be addressed to Dr. Weijia Zhang ([email protected]) or Dr. Liang Qiao ([email protected])

ACKNOWLEDGEMENTS This work is supported by National Natural Science Foundation of China (NSFC, 81671849), Ministry of Science and Technology of China (MOST, 2016YFE0132400, 2017YFC0906700), Science and Technology Commission of Shanghai Municipality (17JC1400900, 18441901000, 16391903900). W.Z. acknowledges funding from the NSFC (31501555, 81772007, and 21734003), Science and Technology Commission of Shanghai Municipality (17JC1400200), and Innovation Program of Shanghai Municipal Education Commission (2017-0107-00-07-E00027).

REFERENCES (1) Organization, W. H. WHO Global Strategy for Containment of Antimicrobial Resistance. Wkly. Epidemiol. Rec. 2001, 76, 298-299. (2) O'Neill, J. Antimicrobial Resistance : Tackling a crisis for the health and wealth of nations. Tackling drug-resistant infections globally: final report and recommendations; Review on Antimicrobial Resistance 2014. (3) Yao, Z.; Kahne, D.; Kishony, R. Distinct Single-Cell Morphological Dynamics under Beta-Lactam Antibiotics. Mol. Cell 2012, 48, 705-712. (4) Justice, S. S.; Hunstad, D. A.; Cegelski, L.; Hultgren, S. J. Morphological plasticity as a bacterial survival strategy. Nat. Rev. Microbiol. 2008, 6, 162168. (5) Bos, J.; Zhang, Q.; Vyawahare, S.; Rogers, E.; Rosenberg, S. M.; Austin, R. H. Emergence of antibiotic resistance from multinucleated bacterial filaments. Proc. Natl. Acad. Sci. U. S. A. 2015, 112, 178-183. (6) Zhang, D.; Bi, H.; Liu, B.; Qiao, L. Detection of Pathogenic Microorganisms by Microfluidics Based Analytical Methods. Anal. Chem. 2018, 90, 5512-5520. (7) Tang, Y.; Zhen, L.; Liu, j.; Wu, J. Rapid antibiotic susceptibility testing in microfluidic pH sensor. Anal. Chem. 2013, 85, 2787-2794. (8) Jiang, X.; Kang, Y.; Pan, X.; Yu, J.; Ouyang, Q.; Luo, C. Studies of the drug resistance response of sensitive and drug-resistant strains in a microfluidic system. Integr. Biol. 2014, 6, 143-151. (9) Li, B.; Qiu, Y.; Glidle, A.; McIlvenna, D.; Luo, Q.; Cooper, J.; Shi, H.-C.; Yin, H. Gradient Microfluidics Enables Rapid Bacterial Growth Inhibition Testing. Anal. Chem. 2014, 86, 3131-3137. (10) Chung, B. G.; Choo, J. Microfluidic gradient platforms for controlling cellular behavior. Electrophoresis 2010, 31, 3014-3027. (11) Toh, A. G. G.; Wang, Z. P.; Yang, C.; Nguyen, N.-T. Engineering microfluidic concentration gradient generators for biological applications. Microfluid. Nanofluid. 2014, 16, 1-18. (12) Dertinger, S. K. W.; Chiu, D. T.; Jeon, N. L.; Whitesides, G. M. Generation of gradients having complex shapes using microfluidic networks. Anal. Chem. 2001, 73, 1240-1246. (13) Russom, A.; Irimia, D.; Toner, M. Chemical gradient-mediated melting curve analysis for genotyping of SNPs. Electrophoresis 2009, 30, 25362543. (14) Hung, P. J.; Lee, P. J.; Sabounchi, P.; Lin, R.; Lee, L. P. Continuous perfusion microfluidic cell culture array for high-throughput cell-based assays. Biotechnol. Bioeng. 2005, 89, 1-8.



Page 8 of 9

(15) Claydon, M. A.; Davey, S. N.; EdwardsJones, V.; Gordon, D. B. The rapid identification of intact microorganisms using mass spectrometry. Nat. Biotechnol. 1996, 14, 1584-1586. (16) Yang, Y.; Lin, Y.; Chen, Z.; Gong, T.; Yang, P.; Girault, H. H.; Liu, B.; Qiao, L. Bacterial Whole Cell Typing by Mass Spectra Pattern Matching with Bootstrapping Assessment. Anal. Chem. 2017, 89, 12556-12561. (17) Yang, Y.; Lin, Y.; Qiao, L. Direct MALDI-TOF MS Identification of Bacterial Mixtures. Anal. Chem. 2018, 90, 10400-10408. (18) van Veen, S. Q.; Claas, E. C. J.; Kuijper, E. J. High-Throughput Identification of Bacteria and Yeast by Matrix-Assisted Laser Desorption Ionization-Time of Flight Mass Spectrometry in Conventional Medical Microbiology Laboratories. J. Clin. Microbiol. 2010, 48, 900-907. (19) Du, Z.; Yang, R.; Guo, Z.; Song, Y.; Wang, J. Identification of Staphylococcus aureus and determination of its methicillin resistance by matrixassisted laser desorption/ionization, time-of-flight mass spectrometry. Anal. Chem. 2002, 74, 5487-5491. (20) Zhang, R.; Qin, Q.; Liu, B.; Qiao, L. TiO2-Assisted Laser Desorption/Ionization Mass Spectrometry for Rapid Profiling of Candidate Metabolite Biomarkers from Antimicrobial-Resistant Bacteria. Anal. Chem. 2018, 90, 3863-3870. (21) Qin, D.; Xia, Y.; Whitesides, G. M. Soft lithography for micro- and nanoscale patterning. Nat. Protoc. 2010, 5, 491-502. (22) Xia, Y.; Whitesides, G. M. Soft lithography. Annu. Rev. Mater. Sci. 1998, 28, 153-184. (23) Drevinek, M.; Dresler, J.; Klimentova, J.; Pisa, L.; Hubalek, M. Evaluation of sample preparation methods for MALDI-TOF MS identification of highly dangerous bacteria. Lett. Appl. Microbiol. 2012, 55, 40-46. (24) Xia, J.; Wishart, D. S. Using MetaboAnalyst 3.0 for Comprehensive Metabolomics Data Analysis. Curr. Protoc. Bioinformatics 2016, 55, 14.10.11-14.10.91. (25) Bankevich, A.; Nurk, S.; Antipov, D.; Gurevich, A. A.; Dvorkin, M.; Kulikov, A. S.; Lesin, V. M.; Nikolenko, S. I.; Pham, S.; Prjibelski, A. D.; Pyshkin, A. V.; Sirotkin, A. V.; Vyahhi, N.; Tesler, G.; Alekseyev, M. A.; Pevzner, P. A. SPAdes: a new genome assembly algorithm and its applications to single-cell sequencing. J. Comput. Biol. 2012, 19, 455-477. (26) Delcher, A. L.; Harmon, D.; Kasif, S.; White, O.; Salzberg, S. L. Improved microbial gene identification with GLIMMER. Nucleic Acids Res. 1999, 27, 4636-4641. (27) Guo, Y.; Li, J.; Liu, J.; Wang, T.; Li, Y.; Yuan, Y.; Zhao, J.; Chang, D.; Fang, X.; Li, T. Effects of Space Environment on Genome, Transcriptome, and Proteome of Klebsiella pneumoniae. Arch. Med. Res. 2015, 46, 609618. (28) Frazee, A. C.; Pertea, G.; Jaffe, A. E.; Langmead, B.; Salzberg, S. L.; Leek, J. T. Ballgown bridges the gap between transcriptome assembly and expression analysis. Nat. Biotechnol. 2015, 33, 243-246. (29) Leng, N.; Dawson, J. A.; Thomson, J. A.; Ruotti, V.; Rissman, A. I.; Smits, B. M. G.; Haag, J. D.; Gould, M. N.; Stewart, R. M.; Kendziorski, C. EBSeq: an empirical Bayes hierarchical model for inference in RNA-seq experiments. Bioinformatics 2013, 29, 1035-1043. (30) Sun, P.; Liu, Y.; Sha, J.; Zhang, Z.; Tu, Q.; Chen, P.; Wang, J. Highthroughput microfluidic system for long-term bacterial colony monitoring and antibiotic testing in zero-flow environments. Biosens. Bioelectron. 2011, 26, 1993-1999.

(31) Kim, S. C.; Cestellos-Blanco, S.; Inoue, K.; Zare, R. N. Miniaturized Antimicrobial Susceptibility Test by Combining Concentration Gradient Generation and Rapid Cell Culturing. Antibiotics 2015, 4, 455-466. (32) Choi, J.; Yoo, J.; Lee, M.; Kim, E.-G.; Lee, J. S.; Lee, S.; Joo, S.; Song, S. H.; Kim, E.-C.; Lee, J. C.; Kim, H. C.; Jung, Y.-G.; Kwon, S. A rapid antimicrobial susceptibility test based on single-cell morphological analysis. Sci. Transl. Med. 2014, 6, 267ra174. (33) Niebel, M.; Quick, J.; Prieto, A. M.; Hill, R. L.; Pike, R.; Huber, D.; David, M.; Hornsey, M.; Wareham, D.; Oppenheim, B. Deletions in a ribosomal protein-coding gene are associated with tigecycline resistance in Enterococcus faecium. Int. J. Antimicrob. Agents 2015, 46, 572-575. (34) Hu, M.; Nandi, S.; Davies, C.; Nicholas, R. A. High-Level Chromosomally Mediated Tetracycline Resistance in Neisseria gonorrhoeae Results from a Point Mutation in the rpsJ Gene Encoding Ribosomal Protein S10 in Combination with the mtrR and penB Resistance Determinants. Antimicrob. Agents Chemother. 2005, 49, 4327-4334. (35) Maisnier-Patin, S.; Berg, O. G.; Liljas, L.; Dan, I. A. Compensatory adaptation to the deleterious effect of antibiotic resistance in Salmonella typhimurium. Mol. Microbiol. 2002, 46, 355-366. (36) Langstraat, J.; Bohse, M.; Clegg, S. Type 3 fimbrial shaft (MrkA) of Klebsiella pneumoniae, but not the fimbrial adhesin (MrkD), facilitates biofilm formation. Infect. Immun. 2001, 69, 5805-5812. (37) Jean, N. L.; Bougault, C. M.; Egan, A. J. F.; Vollmer, W.; Simorre, J. P. Solution NMR assignment of LpoB, an outer-membrane anchored Penicillin-Binding Protein activator from Escherichia coli. Biomol. Nmr Assigm. 2015, 9, 123-127. (38) Schwanhäusser, B.; Busse, D.; Li, N.; Dittmar, G.; Schuchhardt, J.; Wolf, J.; Chen, W.; Selbach, M. Global quantification of mammalian gene expression control. Nature 2013, 495, 126-127. (39) Vuotto, C.; Longo, F.; Pascolini, C.; Donelli, G.; Balice, M. P.; Libori, M. F.; Tiracchia, V.; Salvia, A.; Varaldo, P. E. Biofilm formation and antibiotic resistance in Klebsiella pneumoniae urinary strains. J. Appl. Microbiol. 2017, 123, 1003-1018. (40) Gupta, P.; Sarkar, S.; Das, B.; Bhattacharjee, S.; Tribedi, P. Biofilm, pathogenesis and prevention-a journey to break the wall: a review. Arch. Microbiol. 2016, 198, 1-15. (41) Greene, N. G.; Fumeaux, C.; Bernhardt, T. G. Conserved mechanism of cell-wall synthase regulation revealed by the identification of a new PBP activator in Pseudomonas aeruginosa. Proc. Natl. Acad. Sci. U. S. A. 2018, 115, 3150-3155. (42) Plessis, M. D.; Smith, A. M.; Klugman, K. P. Analysis of penicillinbinding protein 1b and 2a genes from Streptococcus pneumoniae. Microb. Drug Resist. 2000, 6, 127-131. (43) Urban, P. L.; Jefimovs, K.; Amantonico, A.; Fagerer, S. R.; Schmid, T.; Maedler, S.; Puigmarti-Luis, J.; Goedecke, N.; Zenobi, R. High-density micro-arrays for mass spectrometry. Lab Chip 2010, 10, 3206-3209. (44) Nicolardi, S.; Switzar, L.; Deelder, A. M.; Palmblad, M.; van der Burgt, Y. E. M. Top-Down MALDI-In-Source Decay-FTICR Mass Spectrometry of Isotopically Resolved Proteins. Anal. Chem. 2015, 87, 3429-3437. (45) András, K.; Smith, D. F.; Reschke, B. R.; Powell, M. J.; Heeren, R. M. A. Top-down mass spectrometry imaging of intact proteins by laser ablation ESI FT-ICR MS. Proteomics 2014, 14, 1283-1289. (46) Suckau, D.; Resemann, A. MALDI Top-Down sequencing: calling Nand C-terminal protein sequences with high confidence and speed. J. Biomol. Tech. 2009, 20, 258-262.


Page 9 of 9 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


Insert Table of Contents artwork here

9 ACS Paragon Plus Environment

MALDI-TOF Characterization of Protein Expression Mutation During

Recommend Documents