Global Profiling of Metabolite and Lipid Soluble Microbial Products in

Subscriber access provided by GAZI UNIV

Article

Global profiling of metabolite and lipid soluble microbial products in anaerobic wastewater reactor supernatant using UPLC/MSe Phornpimon Tipthara, Chinagarn Kunacheva, Yan Ni Annie Soh, Stephen C. C. Wong, Ng Sean Pin, David C Stuckey, and Bernhard O Boehm J. Proteome Res., Just Accepted Manuscript • DOI: 10.1021/acs.jproteome.6b00681 • Publication Date (Web): 09 Jan 2017 Downloaded from http://pubs.acs.org on January 11, 2017

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 free 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 accessible to all readers and 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.

Journal of Proteome Research 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 43

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

Journal of Proteome Research

Global profiling of metabolite and lipid soluble microbial products in anaerobic wastewater reactor supernatant using UPLC/MSE

Phornpimon Tipthara,1 Chinagarn Kunacheva,2 Yan Ni Annie Soh,2 Stephen C.C. Wong,3 Ng Sean Pin,1 David C. Stuckey,2,4* and Bernhard O. Boehm1*

1 Lee Kong Chian School of Medicine, Nanyang Technological University, Singapore 636921 2 Advanced Environmental Biotechnology Centre, Nanyang Environment & Water Research Institute, Nanyang Technological University, Singapore 637141 3 Waters Pacific Pte. Ltd., Singapore Science Park 2, Singapore 117528 4 Department of Chemical Engineering, Imperial College London, UK, SW7 2AZ

KEYWORDS:

metabolite profiling, Q-TOF, MS/MS, soluble microbial products,

wastewater, anaerobic bioreactors

1 ACS Paragon Plus Environment


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 43

ABSTRACT

Identification of soluble microbial products (SMPs) released during bacterial metabolism in mixed cultures in bioreactors is essential to understanding fundamental mechanisms of their biological production. SMPs constitute one of the main foulants (together with colloids and bacterial flocs) in membrane bioreactors widely used to treat, and ultimately recycle wastewater. More importantly, the composition and origin of potentially toxic, carcinogenic, or mutagenic SMPs in renewable/reused water supplies must be determined and controlled. Certain classes of SMPs have previously been studied by GC-MS, LC-MS, and MALDI-ToF MS; however, a more comprehensive LC-MS based method for SMP identification is currently lacking. Here, we develop a UPLC-MS approach to profile and identify metabolite SMPs in the supernatant of an anaerobic batch bioreactor. The small biomolecules were extracted into two fractions based on their polarity, and separate methods were then used for the polar and nonpolar metabolites in the aqueous and lipid fractions, respectively. SMPs that increased in the supernatant after feed addition were identified primarily as phospholipids, ceramides, with cardiolipins in the highest relative abundance, and these lipids have not previously been reported in wastewater effluent.

INTRODUCTION

In wastewater treatment it is known that the effluent from bioreactors contains a complex mixture of soluble microbial products (SMPs).1 Past work has shown that most SMPs are not present in the influent, but are produced during microbial metabolism and cell lysis.2-4 2 ACS Paragon Plus Environment

Page 3 of 43

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


It is also well established that SMPs cause membrane fouling in bioreactors,5,6 and act as chlorination precursors in post treatment. Hence, it is important to identify the composition of SMPs, determine their origin, and understand what system parameters influence their production.

Most studies on SMPs focus on their overall properties such as concentration,3 molecular weight distribution,7,8 aromaticity,9,10 biodegradability,11 and toxicity.12 Also, the characterization of SMPs has previously been performed using a number of colorimetric methods,13-15 and HPLC with UV detection.2,8,10,16 However, these methods only characterize SMPs at a macro level, but do not provide any molecular identification.

Gas chromatography-mass spectrometry (GC-MS) has been used to detect long chain alkanes, alkenes, aromatic carbons, esters, phenols, amides, and long chain carbohydrates.4,16,17 GC-MS is a powerful technique but is limited primarily to small nonpolar thermally stable volatile molecules, and requires chemical derivatization of many of the metabolite species prior to MS analysis.18 For this reason, the number and type of organics accessible to GC-MS is limited. Untargeted metabolomics methods using liquid chromatography mass spectrometry (LC-MS) have been applied to many research fields, including environmental research. Although LC-MS is well suited for profiling metabolites in complex biological samples,19,20 to the best of our knowledge there have been no published applications of LC-ESI-MS for identification of metabolite components in SMPs. LC-MS has only been used for discrimination of metabolite components in SMPs of aerobic and anaerobic bioreactor, while MALDI-MS has only been used for analyzing



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 4 of 43

the protein composition of SMPs, although no protein profiling was reported.2,16 The aim of ‘untargeted’ global profiling is an unbiased comprehensive analysis of all metabolites in a biological system. However, a single generic LC-MS method may not efficiently separate and detect all classes of metabolites together. Metabolites describe a very wide range of naturally occurring small molecules with very different chemical properties; however, from an analytical point of view they can be broadly separated into two classes based on their polarity; polar metabolites (e.g. amino acids, sugars, nucleotides) and lipid metabolites (e.g. phospholipids, ceramides, sterols, triacylglycerol). In order to identify all classes of metabolites, separate polar and non-polar metabolites specific LC-MS methods may be needed, and a more comprehensive profiling of SMPs will enhance our understanding of wastewater treatment systems. In this paper an untargeted analysis of SMPs was performed using UPLC-ESI-Q-ToF MS to identify polar and lipid metabolites; paralleled fragmentation acquisition of low and high collision energy, or so called MSE was applied.

EXPERIMENTAL SECTION

Chemicals

Acetonitrile, methanol, isopropanol, water (all Optima LC-MS Grade), and chloroform (Analytical Reagent Grade) were purchased from Fisher Scientific (Fair Lawn, NJ). Formic acid and ammonium formate (eluent additive for LCMS) were purchased from Sigma-


Page 5 of 43

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


Aldrich (Dorset, UK). Total recovery sample vials (TruView LC-MS Certified, Product Number 186005663CV) were supplied by Waters Corp. (Elstree, UK).

Supernatant collection

The configuration of the reactor was similar to one described previously.16 Briefly, an anaerobic continuous stirred-tank reactor used in this study was initially seeded with an inoculum of screened sludge (bacteria) from a conventional domestic wastewater treatment plant. The anaerobic bioreactor was fed with a synthetic feed comprised of glucose, peptone, meat extract and essential nutrients at a concentration of 4 gCOD/L. To ensure that steady state conditions had been achieved, the reactors were run for at least 60 days at an effective hydraulic retention time (HRT) of 7 days before the final samples for analysis were collected. Samples were collected from the anaerobic reactor at 0, 4, and 48 h after the final batch feed, and filtered through a 0.45 µm glass fiber filter to remove suspended solids. The 0 h sample was collected after the feed was fed and fully mixed with the sludge; this was used as a control sample. The 4 h and 48 h samples were selected as being representative of SMPs from the fermentation and methanogenesis stages, respectively.

Metabolite and lipid extraction

Supernatant samples including feed, and SMPs collected at 0 h, 4 h, and 48 h (50 mL) were lyophilized followed by resuspension and vortex mixing in LC-MS grade water (1 mL) to produce a homogenous solution. Liquid-liquid extraction was then performed by adding



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 6 of 43

2:1 (v/v) chloroform/methanol (3 mL) and vortex mixing at room temperature for 10 min. The mixture was then centrifuged at 1,610 g, 20 °C for a further 10 min. The upper aqueous (polar metabolite) fraction (0.4 mL) and lower organic (lipid) fraction (0.4 mL) were carefully transferred to separate 1.5 mL Eppendorf tubes and dried in a vacuum concentrator (Eppendorf).

UPLC/MSE

UPLC-MS analyses were conducted using an ACQUITY UPLC with a 2777C autosampler (Waters Corp, USA) coupled to a Xevo G2-XS QTof mass spectrometer (Waters, Manchester, UK) with an electrospray ionization (ESI) source. The intact molecular masses and their associated dissociation products were recorded for all compounds using MSE, whereby retention time aligned data were collected in two channels simultaneously; low collision energy (6 V) for precursor ions and a high collision energy ramp (15–45 V) for product ions, with no precursor ion selection in the quadrupole. A mass range of m/z 100 to 1,000 was acquired for polar metabolites, and m/z 100 to 1,500 for lipids. High mass accuracy was maintained by applying lock mass correction using leucine enkephalin as a reference mass (m/z 556.2771 in ESI+ and m/z 554.2615 in ESI-).

Metabolite profiling (aqueous fraction)

The polar metabolite extracts were reconstituted in water (200 μL) and vortex mixed for 10 min to ensure complete dissolution. The solution was then centrifuged at 9,600 g and


Page 7 of 43

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


4 °C for 10 min to remove undissolved particulates and transferred to a LC-MS sample vial. The metabolite samples were stored in the autosampler at 4°C, and 5 μL injections were separated using an ACQUITY UPLC HSS T3 column (2.1 × 100 mm, 1.8 μm; Waters) at 40 °C. Three replicate injections were made for each ESI positive and ESI negative mode. Mobile phase A consisted of water with 0.1% formic acid, while mobile phase B consisted of acetonitrile with 0.1% formic acid. The elution gradient was set as follows: 0−40% B (1.0−11.0 min), 40−80% B (11.0−11.1 min), 80% B (11.1−13.0 min), 80−0% B (13.0−13.1 min), 0% B (13.1−16 min) with a ﬂow rate of 0.4 mL/min. For the MS ionization source, capillary voltages of 3.0 kV and 1.8 kV were employed for positive and negative electrospray ionization, respectively, with 40 V cone voltage applied in both modes. The source and desolvation temperatures were set at 120 °C and 400 °C respectively, and the desolvation gas flow at 850 L/h.

Lipid profiling (organic fraction)

The lipid extracts were reconstituted in 2:1:1(v/v/v) isopropanol/acetonitrile/water (100 μL) and vortex mixed for 10 min. The solution was then centrifuged at 9,600 g and 4°C for 10 min to remove undissolved particulates and transferred to a LC-MS sample vial. The samples were kept at 4°C in the autosampler, and 5 μL injections were separated using an ACQUITY UPLC CSH C18 column (2.1 × 100 mm, 1.7 μm; Waters), with a column temperature of 55 °C. Three replicate injections were made for each ESI positive and ESI negative mode. Mobile phase A consisted of 60:40 (v/v) acetonitrile/water with 10 mM ammonium formate and 0.1% formic acid. Mobile phase B consisted of 90:10 (v/v) isopropanol/acetonitrile with 10 mM ammonium formate and 0.1% formic acid. The 7 ACS Paragon Plus Environment


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 8 of 43

elution gradient was set as follows: 40−43% B (0.0−2.0 min), 43−50% B (2.0−2.1 min), 50−54% B (2.1−12.0 min), 54−70% B (12.0−12.1 min), 70−99% B (12.1−18 min), 99−40% B (18.0−18.1 min), 40% B (18.1− 20.0 min), with a ﬂow rate of 0.4 mL/min. For the MS, capillary voltages of 2.8 kV and 2.2 kV were employed for positive and negative electrospray ionization, respectively, with 30 V cone voltage applied in both modes. The source and desolvation temperatures were set at 120 °C and 400 °C, respectively, and the desolvation gas flow at 850 L/h.

Data analysis and metabolite identification

The samples were analyzed in randomized order to avoid the effects of time dependent changes such as falling sensitivity (for example due to the ionization source becoming contaminated), on the statistical analysis. It is often observed that, for the first few injections, both retention times and peak intensities drift. To ensure that the retention times and peak intensities have stabilized before injecting the samples, the LCMS system was ‘conditioned’ using around 10 replicate injections of a pooled “Quality Control” sample consisting of equal aliquots of each sample. The system is adjudged to be properly conditioned only when the conditioning QC injections no longer ‘drift’ in the PCA. Aliquots of the same pooled QC were also injected at regular intervals between and after the samples for monitoring analytical variation. Since the pooled QC contained all compounds present in the SMPs collected at all the time points (0, 4, and 48 hours), this was also used as a reference for both peak alignment and normalization.


Page 9 of 43

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


The RAW data files for the feed, and SMPs collected at 0, 4 and 48 h were imported directly into Progenesis QI (version 2.1, Nonlinear Dynamics, Newcastle UK). During data import, a wavelet-based approach is used by Progenesis to create a “model” of the data, removing much of the background noise prior to peak picking and analysis. A noise estimation algorithm which examines the intensities of groups of MS peaks in the aggregated data (combined ion map of all samples added together) is used to determine the noise level in the data. Retention time alignment, peak picking, and deconvolution of adducts were performed by Progenesis prior to statistical analysis and identification.

Non-filtered data were exported to EZinfo (version 3.0.3.0, Umetrics) for multivariate analysis.21 Principal component analysis (PCA) with Pareto scaling, was used to provide an overview of the separation of sample groups (different time points). From the PCA, orthogonal projection to latent structures discriminant analysis (OPLS-DA) was used to determine the most significant components separating the sample groups. These significant features were extracted from the S-plot and exported back to Progenesis QI for correlation analysis and identification. Correlation analysis is performed on arcsinh-normalized compound abundances.

Polar and lipid metabolites were identified using i) MetaScope by searching compound features against structural databases (SDFs) for E.coli (http://ecmdb.ca/),22 HMDB (http://www.hmdb.ca/),23 LIPID MAPS (http://www.lipidmaps.org/),24 and ChEBI (https://www.ebi.ac.uk/chebi/),25 and ii) Chemspider by searching compound features against the KEGG database (http://www.genome.jp/kegg/). The search tolerance used for



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 10 of 43

precursor ion masses was 5 ppm, and for theoretical fragment ions masses was 10 ppm. Reported compound identifications were scored and filtered by mass accuracy, isotopic distribution similarity >90%, and fragment ion matches based on the putative chemical structures.

RESULTS AND DISCUSSION

Analytical method and data analysis for SMPs

The aim of this study was to improve the coverage of compound classes identifiable in complex SMPs by combining LLE and UPLC-TOF MS. Polar metabolites in the aqueous fraction and non-polar (lipid) metabolites in the organic fraction, were separated by high resolution chromatography (UPLC) using different stationary and mobile phases, more suited for the separation and ionization of the different types of analyte groups found in the different fractions. Both positive and negative ions were acquired to maximize the coverage of metabolites profiled. All metabolite precursor ions and their dissociation products ions were acquired simultaneously using MSE. Various SMPs extraction methods have been carried out including extraction by methanol solution (30% v/v in aqueous solution) followed by LCMS profiling analysis,2 extraction by nonpolar solvents using dichloromethane and hexane followed by GCMS analysis.16 However, based on our knowledge this study will be the first time where researchers have analysed a nonpolar (lipid) extract of SMPs by LCMS.


Page 11 of 43

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


In this study, the identification of SMPs was carried out at different times during a batch anaerobic metabolic process. Digestion via anaerobic bacteria consists of a sequence of steps, as shown in Figure 1. Firstly, they degrade feed polymers (sugar + peptone + meat extract + essential nutrients) by acid fermentation (acidogenesis) producing shorter fatty acids, monosaccharides, amino acids, purines/pyrimidines, acetate, hydrogen, carbon dioxide, alcohols, and ammonia. These products are used as substrates for the second step, methanogenesis, carried out by anaerobic methane forming bacteria. Reactor supernatant samples were collected after 4 h, at the highest concentration of volatile fatty acids (1,740 mg/L), representing SMPs from the acid fermentation step. Another supernatant sample was collected at 48 h when the volatile fatty acids were all consumed (< 5 mg/L), representing the SMPs from methanogenesis.

Figure 1. 11 ACS Paragon Plus Environment


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 12 of 43

The base peak intensity chromatograms (Figure 2) suggest marked differences in the metabolite and lipid profiles of the SMPs at each time point. The chromatographic profiles were highly reproducible and essentially identical for the triplicate injections, and this is essential for the statistical analysis that follows (coefficient of variation (CV) was shown in supplementary Table S-1).


Page 13 of 43

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.



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 14 of 43

At 0 h, the components identified are expected to be mainly from the feed, and it can be seen that the feed compounds were primarily negatively ionized polar metabolites (panel B). As the feed is metabolized, positive ion lipids become increasingly prevalent (panel C). For the 18 min gradient (22 min total analytical run time) used for profiling lipids, typically around 1.0-1.4 million aggregated positive ion peak features can be detected. Following peak picking and deconvolution of the charge states, dimers/trimers, adducts, and isotopologues, the number of features are reduced to about 10-20,000 compounds. The polar metabolite fractions are more complex than the lipid fraction, which is to be expected because the SMP collected at 0 h consists mainly of feed compounds, which are predominantly small polar molecules.

Statistical analysis of the data shows clear differences between the SMPs collected at the different time points. The tight clustering of the run order QC samples (and replicate measurements for the different time points) in the PCA scores plots (Figure 3A) indicates minimal analytical variation, conferring greater confidence that the spread of the samples in the scores plot are due to sample variation rather than analytical variation (PCA of other datasets are in supplement Figure S-1). Two general approaches were taken to determine what compounds changed over time, namely discriminant analysis using orthogonal projections to latent structures (OPLS-DA), and correlation analysis. The statistical model created by OPLS-DA is represented by an S-Plot (Figure 3B and 3C). The S-Plot shows the covariance and correlation loadings from the OPLS model, with each point an exact mass/retention time (EMRT) feature. Only those features elevated in 4 h (3B) and 48 h (3C) SMP samples (appearing in the upper right quadrant), which both contribute most to


Page 15 of 43

the separation between the sample groups and of highest confidence (most correlated), were selected for further analysis and identification.

Metabolite

A

Li

Scores Comp[1] vs. Comp[2]

Scores Comp[1] vs. Comp[2] colored by Condition

5000

1500

4000 1000

3000 2000

ESI+

500

t[2]

t[2]

1000 0

0 -1000

-500

-2000 -3000

-1000

-4000 -1500 -3000

-2000

-1000

0

1000

2000

-5000 -8000 -7000 -6000 -5000 -4000 -3000 -2000 -1000

3000

t[1] 0h

4h

48h

0h

QC

4h

EZInf o controlled b

EZInf o controlled by Progenesis QI 3.0.3 - 20160509PT_Met_Pos_3points.usp (M1: PCA-X) - 2016-05-10 12:32:37 (UTC+8)

B

Scores Comp[1] vs. Comp[2] colored by Condition

Scores Comp[1] vs. Comp[2]

S-Plot (0h = -1, 4h = 1)

2000

1500

1.0

1500

0.8

1000

0.6

500

500

t[2]

0.4

t[2]

p(corr)[1] (Correlation)

1000

ESI-

0

0

0.2 -500

-500

-0.0 -1000 -1000

-0.2

-1500

-0.4

-1500 -2000 -3000

-0.6

-2500

-2000

-1500

-1000

-500

0

500

1000

1500

2000

2500

3000

-3000

-2500

-2000

-1500

-1000

-500

t[1]

-0.8

0h

4h

48h

QC

0h

EZInf o controlled by Progenesis QI 3.0.3 - 20160510PT_Met_Neg_3points.usp (M1: PCA-X) - 2016-05-10 18:27:49 (UTC+8)

-0.14

-0.12

-0.10

-0.08

-0.06

-0.04

-0.02

-0.00

0.02

0.04

0.06

0.08

0.10

0.12

0.14

p[1] (Loadings) EZInf o controlled by Progenesis QI 3.0.3 - 20160509PT_Met_Pos_3points.usp (M2: OPLS-DA) - 2016-05-10 12:47:35 (UTC+8)

C

S-Plot (0h = -1, 48h = 1) 1.0 0.8 0.6 0.4 0.2 -0.0 -0.2 -0.4 -0.6 -0.8 -1.0 -0.12

-0.10

-0.08

-0.06

-0.04

-0.02

-0.00

0.02

0.04

0.06

0.08

0.10

p[1] (Loadings) EZInf o controlled by Progenesis QI 3.0.3 - 20160509PT_Met_Pos_3points.usp (M3: OPLS-DA) - 2016-05-10 13:07:11 (UTC+8)


4h

EZInf o controlled by P

-1.0

p(corr)[1] (Correlation)

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



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 16 of 43

Figure 3.

Correlation analysis clusters compounds according to how closely correlated their relative abundance profiles are; and was used to characterize the abundance profiles of those compounds found using OPLS-DA to increase in SMPs at 4 h and 48 h. Compounds whose behavior show a high degree of correlation are clustered together, allowing compounds with similar time dependent relative abundance profiles to be classified into three groups. The first group consists of compounds that appear at high relative abundance at 4 h, but then decrease back to near start levels again by 48 h (Figure 4A). These are likely to be ‘intermediate’ metabolic products generated by acid fermentation in the bioreactor, which are then consumed or transformed by methanogenic bacteria to form other compounds, such as those in Figure 4C, after a long period. If these ‘intermediate’ compounds were to be consumed quickly, then their end products might be expected to behave like those in Figure 4B, in which compounds increase in relative abundance from 0 h to 4 h, and again from 4 h to 48 h. These different profiles may reflect dynamic changes in the bioreactor bacterial composition, and could be correlated to changes in microbial ecology. The resulting list of deconvoluted metabolite/lipid compound features increasing in the SMPs at 4 h and 48 h were searched against multiple metabolite and lipid structural databases. Preliminary compound identification was based on a combination of selection criteria; exact mass error (90%), and theoretical fragment ion matching (

Global Profiling of Metabolite and Lipid Soluble Microbial Products in

Recommend Documents