LC-MS/MS-based metabolic profiling of Escherichia coli under heterologous gene expression stress

2.1 Materials

The genetically modified cells of E. coli strain BL21 (DE3) RIL were obtained from Novagen (Burlington, MA). Bacterial growth media were purchased from Sigma-Aldrich (Taufkirchen, Germany), plasmid vector pET21a (+) were obtained from Novagen (MERCK), and the gene coding for lactate dehydrogenase was polymerase chain reaction (PCR)-amplified from the genomic DNA of Geobacillus thermodenitrificans DSM-465 (purchased from the Leibniz Institute DSMZ, Braunschweig, Germany). PCR kit, bacterial transformation kit, and restriction enzymes were purchased from Thermo Fisher Scientific. All other chemicals, reagents, and solutions of molecular biology grade were obtained from local and international companies.

2.2 Preparation of bacterial lysates

Three different categories of E. coli BL21 (DE3) RIL cells were prepared for the investigation: first, the normal control E. coli cells (BCT1), second, the cells transformed with pET21a (+) plasmid to acquire ampicillin resistance (BCT3), third, the cells transformed with recombinant plasmid pET21a (+) having ampicillin resistance gene and ligated with lactate dehydrogenase gene from G. thermodenitrificans DSM-465 (BCT2). All the experiments were performed in triplicates. The normal cells were grown in LB broth, and the shaking incubator was adjusted at 37°C, 200 rpm to obtain the optical density of culture up to 1.95 at 600 nm. The second group of cells was grown in the LB broth supplemented with ampicillin (100 µg/mL of medium) to obtain the same final optical density (1.95) of bacterial culture. The third group of cells was grown under the conditions similar to second group of cells except for the addition of 0.5 mM IPTG (isopropyl β-d-1-thiogalactopyranoside), when the optical density of cells was about 1.95 at 600 nm. The cell cultures for all three groups were harvested by centrifugation at 7000g for 5 minutes at 4°C. Supernatant was discarded and the cells were washed twice in phosphate-buffered saline (PBS) saline. Finally, the cells were suspended in PBS and optical density 600 nm was adjusted to assure the equal number of cells per milliliter of sample. The suspended cells were subjected to sonication at moderate power under same conditions and lysate was placed in the ice-box for further use (Table 1).

2.3 LC-MS/MS-based metabolomics

The metabolites were extracted from E. coli lysate using a combination of methanol:acetonitrile:water in a ratio of (2:2:1, vol/vol). Ice-cold solvent was added to E. coli lysates in equal volumes, mixture was vortexed for 30 seconds and incubated in for 1 hour at −20°C, followed by centrifugation for 15 minutes at 13000g and 4°C. The supernatant was removed and sample was dried in a vacuum concentrator. The dry extracts were reconstituted in 100 μL of acetonitrile:water (1:1, vol/vol), vortexed and centrifuged as above to remove insoluble debris. The supernatants were subjected to LC-MS/MS analysis. Ten microliters of sample was subjected to HPLC column hypersail gold (150 mm × 4.6 mm, 5 μm) with a flow rate of 0.2 mL/min. The mobile phase consisted of 0.1% formic acid and 99.9% ACN formic acid (0.1%, vol/vol) using a linear gradient where the component of solution was changed from 5% B to 100% B in 90 minutes at a constant flow rate of 0.2 mL/min (95% A from 5% to 30% in 72 minutes, 30% to 100% over 10 minutes, and kept at 100% for 5 minutes at a flow rate of 200 µL/min). The column temperature was maintained at 30°C. LTQ XL linear ion trap instrument (Thermo Fisher Scientific (Waltham, MA), catalog number IQLAAEGAAVFACZMAIK) was used for the investigation. Full scan scope was chosen from 100 to 1000 m/z. The spray voltage was set to −3.0 kV. The capillary voltage was fixed at 4.0 V and the temperature at 270°C. Nitrogen was used as a sheath gas and the flow rate was 40 arbitrary units. Helium was used as the buffer gas. Data processing was carried using Xcalibur software (Waltham, MA).

2.4 Data analysis for bacterial global metabolites

The raw MS data (raw files) were processed using XCMS for feature detection, retention time correction, and alignment. The parameters in XCMS were set as follows: CentWave settings for feature detection (Δm/z = 30 ppm, minimum peak width = 10 seconds, maximum peak width = 120 seconds), mzwid = 0.25, minfrac = 0.5, and bw = 10 for chromatogram alignment. After careful evaluation, retention time alignment was shown not to be required. Isotopic peaks and adducts were detected using CAMERA. The precursor were matched to METLIN at 20 ppm accuracy.

2.5 Assessment of major regulatory pathways

The set of significantly regulated metabolites were used for the assessment of major regulatory metabolic pathways involved to counter the stress induced by ampicillin resistance and recombinant protein expression. The regulated metabolites were considered for their application in the better efficiency of E. coli cell factories in recombinant protein production.

3.1 Bacterial growth under normal and stress conditions

The bacterial cells were grown in minimal medium (LB broth), under antibiotic (ampicillin) and exogenous gene transformation stress (used for heterologous protein production) (Table 1). There was no effect of stress conditions on bacterial growth as determined with reference to normal (susceptible cells). It clearly suggests major rearrangements in the level of various metabolites and metabolic pathways in the bacterial population to counter the ampicillin stress and exogenous gene transformation.

3.2 General metabolic profile of E. coli under stress

In the present study, we have evaluated the profile of bacterial metabolites under different stress conditions, that is, ampicillin (BCT3) and ampicillin plus exogenous gene expression (BCT2), with reference to normal (susceptible) bacteria (BCT1). Metabolite profiles of BCT1, BCT2, and BCT3 were studied by LC coupled with tandem mass spectrometry to generate fragmentation patterns and provide metabolite identification. The MS/MS spectra were generated using automated data dependent acquisition, whereby the instrument software Xcaliber continuously evaluated the profile and scanned MS/MS acquisition based on a set of criteria adjusted on instrument settings including acquisition rate, number of precursor candidate ions, collision energy. The total ion current (TIC) and combined mass spectra of three sample categories were determined (Figure 1). Total 592 featured metabolites were identified under the common conditions (Table S1), 107 have shown significant differential regulation (decrease/increase) with P-value ≤ .01. The cloud plot described total 107 features were differentially expressed with P-value ≤ .01 (Figure 2). Quantitative vector scaling normalization have shown that most of the metabolite were close to high significance (Figure 3).

Level of metabolites under two investigated conditions, that is, BCT1, BCT2, and BCT3 have shown major change in the metabolic profiles. However, we did not found remarkable difference in metabolite levels between the two stress groups (Figure 4), clearly suggesting that two different stress conditions modulate the metabolites almost similarly in E. coli. Further, we performed KEGG pathway analysis through metaboanalyst and found alteration in metabolite levels, which results in significant modulation of major metabolic pathways like inositol phosphate metabolism, galactose metabolism, fatty acid metabolism, starch and glucose metabolism, and valine-leucine metabolism (Figure 5). In addition to this we also noticed significant modulation in metabolites related to streptomycin biosynthesis, suggesting that both stress types alter many pathways that are important for bacterial survival and generating environment for antibacterial resistance in exposed E. coli populations.

3.3 Accumulation of 1, 4-dihydroxy-2-naphthoyl–CoA (DHNA)

One of the most significantly increased metabolite in both stressed cell types when compared to normal cells was 1, 4-dihydroxy-2-naphthoyl-CoA (DHNA; m/z 954; retention time 10.6 minutes). DHNA, is a key intermediary rate limiting metabolite for the production of lipid soluble vitamin K. It is synthesized in many bacteria's including E. coli, Mycobacterium tuberculosis, and Staphylococcus aureus. We observed 7.1-fold increase in DHNA levels in both stressed E. coli cell types (Table 2), suggesting an accumulation of DHNA and consequent decrease in vitamin K levels.

3.4 Regulation of l-citrulline and urea cycle

In E. coli, the urea cycle is meant for storing excessive nitrogen. The mitochondrial nitrogen reacts with bicarbonates in aqueous cytoplasmic medium to generate carbamoyl phosphate. The later reacts with l-ornithine to produce l-citrulline. l-Citrulline plays a critical role in the l-arginine synthesis using an ATP molecule. Four to eight times decrease in citrulline has been found in both manipulated cell types when compared to normal E. coli bacteria (Table 2). This can be linked to reduced availability of ornithine and arginine important for urea cycle in E. coli.

3.5 Carbohydrate, terpenoids, and fatty acid metabolites

An increase in oxoglutaric acid level was found in the cells under stress (2.52 times; P-value = .00091). This metabolite is considered as a molecule at the intersection of nitrogen and carbon metabolism and acts as a master regulatory metabolite. Similarly, oleic acid plays prominent role in fatty acid pathway. Its retention time was found 11 minutes and it exhibited nearly 1.6-fold increase in E. coli bacteria under ampicillin and gene expression stress (Table 2). In addition to this, we clearly observed a general increase in the levels of carbohydrate associated metabolites under the stress conditions. The level of carbohydrate metabolites including 2,3-diphosphate glycerate, myoinositol, sucrose, galactinol, glucose 1-phosphate, and fucitol was increased. However, decreased level of N-acetyl-α-d-glucosamine was found. The level of campstrol and ketotorulene was also decreased in both E. coli cell types under stress. However, the cellular levels of neoxantine and xanthine were elevated in both stress types (Figure 6).

3.6 Miscellaneous metabolites

E. coli capable to grow under ampicillin stress indicated considerable modifications in the pattern of many sets of metabolites. Among them, sugar nucleotides like thymidine 5-monophosphate, uridine 5′-diphosphate, adenosine 3 monophosphate, and cyclic 3′5′-(hydrogen phosphate) adenosine were downregulated to variable extent. 2-Deoxyadenosine 5-monophosphate was not affected by ampicillin stress, and it was downregulated during forced gene expression stress (heterologous gene expression). Further, cytidine 5′-diphosphocholine was remained unaffected in the E. coli having ampicillin stress; however, it was considerably up regulated in the E. coli having both stress together. Downregulation of a number of phenolic compounds was found in the ampicillin stress conditions. An increased level of kaempferol 3-galactoside and quercetin 3-glucoside was found. The level of energy metabolites including NADH, AMP, and ADP was decreased in E. coli under ampicillin stress (Figure 6). E. coli having ampicillin stress shows significant up regulation of p-aminobenzoic acid with a retention time is 5.6 minutes and P-value of .002. p-Aminobenzoic acid (pABA) is an intermediate metabolite in the synthesis of folic acid and coenzyme Q (Table 2).