LC/MS Q-TOF-Based Structure

Elucidation and Profiling in Food

and Agricultural Sciences

Luigi Lucini

Università Cattolica del Sacro Cuore

luigi.lucini@unicatt.it

Presentation outline

i. General concepts on QTOF analysis:

High resolution MS

Principles of compound mining

ii. Create a dataset

Alignment and filtering

Recursive analysis

Data interpretation

iii. Some examples from recent research

Physiology

Food technology

Traceability of productions

Compound ID (Molecular Structure Correlation)

Screening of contaminants (All Ions MS/MS)

Accurate mass HRMS

High Resolution Mass Spectrometry (HRMS)

Average mass: isotope-abundance averaged

molecular weight

Accurate mass +high resolution: monoisotopic

actual mass measured with

high accuracy (ppm compared

to nominal)

Why HRMS?

Compound identification increases

at increasing mass accuracy and resolution

Isotopic mass No. compounds Resolution

Monoisotopic mass

(varies in ppm)

Isotope spacing

(varies in ppm)

Isotope distribution

(varies in %)

Scoring based on:

5

Identification / screening

Mass Match +

Abund. Match +

Spacing

Match =

Overall Score

Q-ToF Chemometrics workflow

Run Sample

& Control in

MS mode

Inclusion list / Preferred list

for tandem MS

Run Sample

in MS/MS

mode

Find Compounds

- MFE or FBF

Statistically Filter

Compounds - Recursive

Analysis

Identify Compounds

- Database Search

- MFG

Find Compounds

- Auto or Targeted MS/MS

Confirm structure

-Molecular Structure Correlation

(MSC)

Identify Compounds

- Library Search

- MFG

Identification / screening

Isotopic profile:

- Accurate mass

- Accurate isotope

spacing

- Isotopes ratio

Peak volume:

- Extracted ion current from total

signal

Naive vs forced approach

Forced: Find-by-Formula

Needs a formula or a source of formula (database)

Deeper in compounds fishing

Higher number of false positive

Naive: Find-by-Molecular-Feature

Needs an isotopic model (e.g. glycans, peptides, common organic molecules)

Often requires a following identification step

Lower number of false positives, higher number of false negative

From meaningless to

meaningfull data

Raw data

• Huge amount of information

• Redundancy? False positive?

dataset

• Data gathering

• Alignment and filtering

Interpret.

• Statistics

• Chemometrics & bioinformatics

Answer(s) to a biological problem?

Alignment and filtering by

MassHunter Profinder

1- Review compounds at a

glance:

- Check presence/absence

across treatments

- Preliminary investigate

alignment and check for

peak shape

2- Align and filter

3- Export for recursive

analysis

Choosing thresholds in Profinder

Area: min, max and

med

Mass: med and RSD

Retention time

difference, span and

max width

Alignment and filtering in

Mass Profiler Professional

1. Alignment (optionally using

internal standards) - Retention

time tolerance

2. Filter by mass tolerance

3. Remove irreproducible

compounds – Filter by

Frequency

4. Remove weak compounds –

Filter by Abundance

5. Remove highly variable

compounds – Filter by Sample

Variability

Recursive analysis

Feature Extractor in MassHunter qualitative

Mass & Rt alignment, filtering

Find by Ion in MassHunter qualitative

Mass & Rt alignment, filtering and

chemometrics

1

2

3

4

Recursion

Before recursion

After recursion

Cluster analysis of

artichoke genotypes

Differences in phytochemical profile

were correlated to different radical

scavenging activities

Comparison of different

thermal treatments on tomato sauce

Analysis of variance and fold-change

were combined to highlight the most

important compounds

PCA on Aloe gel based

commercial products

• All samples 99% Aloe gel as

minimum

• Cosmetics and food

supplements clearly

distinguished

• Products from lyophilized aloe

gel could be identified

Oral administration

Topic application lyophilized

Effect of abiotic stress on plant metabolism

(Zn vs salinity vs control)

Fold-change analysis let to perform:

• Unsupervised cluster analysis

• Venn diagrams

Effect of abiotic stress on plant metabolism

(Zn vs salinity vs control)

K-means clustering helps to

identify compounds co-varying

across treatments

Assure POD Italian cheese authenticity

via PLS-DA

Abroad

POD Italy

bactofugation

100% accuracy in PLS-DA class prediction

Assure POD Italian cheese authenticity

via PLS-DA

I. Compounds weight in PLS-DA

model formula can be seen «at a

glance» in the loading view

II. The most important compounds

can be exported as a list, to be

further verified as markers

Pathways analysis of differential

compounds after Aloe administration

Pathways analysis

Molecular Structure Correlator

• “Systematic bond-

breaking” approach

[Hill and Mortishire-

Smith, 2005] to

correlate tandem MS

data with chemical

structures

• Scores are generated from each product ion,

the mass accuracy of the fragments, and the

overall percentage of ions intensity being

plausibly explained with sub-structures

Structural

elucidation to

discriminate

isobaric

compounds

Structure elucidation &

metabolite ID

Structural confirmation

Confirm structure of

selected differential

metabolites

From totally unknown to structures

Identification of

compounds in inks

for food packaging

material has been

possible using MSC

interfaced to

ChemSpider

All Ions MS/MS acquisition:

combining MS and MS/MS data

• Different experiments in the same

acquisition using

o No collision energy

o Low, middle and high collision energy

• Get MS/MS information using scan rates of

MS-only analysis

• Needs a library for comparison (e.g. PCDL)

• Uses a modified find-by-formula algorithm enabled for

fragment confirmation

• Correlates parent and daughter ions, then calculates a co-

elution score for the latter ones

• Enhanced specificity = false positives are removed

All Ions MS/MS acquisition:

combining MS and MS/MS data

THANKS FOR ATTENTION!

The work referred herein was carried out at the

Faculty of Environmental, Food and Agricultural

Sciences: Institute of Environmental and Agricultural Chemistry

PRONUTRIGEN – Research Centre for Nutrigenomics and

Proteomics