Shotgun Sphingolipid Analysis of Human Aqueous Humor

Abstract

This protocol provides a step-by-step guide to shotgun sphingolipid analysis of aqueous humor. We describe the Bligh and Dyer crude lipid extraction method and electrospray ionization tandem mass spectrometry (ESI-MS/MS) coupled with MZmine 2.21 data processing for identification and ratiometric quantitation of sphingosine, sphingosine 1-phosphate, sphingomyelin and ceramide.

1. Introduction

Being once attributed to structural roles only, it is now known that many sphingolipid species are biologically active and play a role in development and progression of neurodegenerative diseases (1–5). Due to many similarities with the central nervous system, the eye has become an object of sphingolipid research as well. However, until recently, the functions of sphingolipids in the eye (6–10) including the aqueous humor (11,12), have been explored rarely.

Here, we describe techniques for shotgun sphingolipid analyses to be used for human aqueous humor, and with modifications for other selected classes of lipids and/or other eye tissues. The procedure is simpler, easier and faster to perform in comparison to many others reported for sphingolipids. Nevertheless, the readers should be aware of its limitations i.e. Bligh and Dyer extraction method is not specific to sphingolipids, resulting in the complexity of the sample and potential matrix effect. This method also does not ensure total sphingolipid recovery, especially of more polar species e.g. sphingosine 1-phosphate. We recommend the use of both internal and external standards to improve the precision and accuracy of the method.

2. Materials

2.1. Lipid Extraction (see Note 1)

1. Methanol-Chloroform: LC-MS grade methanol and LC-MS grade chloroform in 2:1 ratio (freshly prepared in an amber glass bottle at room temperature).

2. Chloroform: LC-MS grade (stored in an amber glass bottle at room temperature).

3. Internal lipid standards (Avanti Polar Lipids Inc., Alabaster, AL, USA) (see Tab. 1).

4. Glassware: Teflon capped tubes, homogenizers and pipette tips (see Note 2).

5. Argon gas: compressed, ultra-high purity grade.

6. Protein quantitation assay kit of choice.

7. Other:

   1. Freezer: −80^oC
   2. Ice bucket or styrofoam container with ice
   3. Vortexer
   4. Centrifuge
   5. Centrifugal vacuum concentrator.
8. Optional:

   1. Antioxidant: butylated hydroxytoluene (BHT)
   2. Water: LC-MS grade.

Tab. 1.

Parameters for analyses of different sphingolipid classes. Internal standards are indicated with *. They are also available as parts of Sphingolipid Mix I/II (LM-6002/6005).

Class	Lipid Standard (Avanti Number)	Collision Energy [V]	Ionization Mode	Q2 CID Gas Pressure [psi]	Daughter Mass [m/z]	Scan Type	Mass Range [m/z]
Sphingosine (18)	D-erythro-sphingosine (860490)	18	Positive	0.5	48	NLS	200–1000
	heptadeca-sphing-4-enine (LM-2000)*
Sphingosine 1-phosphate (19)	D-erythro-sphingosine-1-phosphate (860492)	24	Negative	1.5	79.1	PIS	200–1000
	heptadeca-sphing-4-enine-1-phosphate (LM-2144)*
Sphingomyelin (20)	N-oleoyl-D-erythro-sphingosylphosphorylcholine (860587)	50	Positive	1.0	213.2	NLS	350–1000
	N-(dodecanoyl)-sphing-4-enine-1-phosphocholine (LM-2312)*
Ceramide (21)	N-oleoyl-D-erythro-sphingosine (860519)	32	Negative	1.5	256.2	NLS	300–1000
	N-(dodecanoyl)-sphing-4-enine (LM-2212)*

2.2. Mass Spectrometric Analysis

1. Acetonitrile:Isopropanol: LC-MS grade acetonitrile and LC-MS grade 2- propanol in 1:1 ratio (stored in an amber glass bottle at room temperature).

2. Lithiated Chloroform:Methanol: 5 mM LiCl in LC-MS grade methanol and LC-MS grade chloroform in 1:1 ratio (freshly prepared).

3. Internal and external lipid standards (Avanti Polar Lipids Inc., Alabaster, AL, USA; see Tab. 1).

4. Mass spectrometry:

   1. TSQ Quantum Access Max Triple Quadrupole Mass Spectrometer (Thermo Fisher Scientific Inc., Pittsburgh, PA, USA) controlled by Xcalibur 2.3 Software.
   2. TriVersa NanoMate (Advion Bioscience Ltd., Ithaca, NY, USA): chip-based electrospray ionization technology with Chipsoft 8.3.3 software for direct infusion of samples and standards.
5. Other:

   1. Argon gas: compressed, ultra-high purity grade
   2. Nitrogen: compressed, ultra-high purity grade
   3. Vortexer
   4. Ice bucket or styrofoam container with ice.

2.3. Data Analysis

1. MZmine 2.21: an open source software designed specifically for mass spectrometry data analysis. It allows for spectra processing, building of chromatograms for identification and export of a variety of vendor specific formats.

2. Sample spectra data.

3. Custom database (optional).

3. Methods

3.1. Lipid Extraction (see Notes 1–5)

Before starting lipid extraction, transfer an aliquot of each sample to a new tube for determination of total protein content using bicinchoninic acid (BCA (13) or Bradford method (14). For lower protein concentrations not measurable using BCA or Bradford method, we recommend the use of the PhastGel system and densitometry using a bovine serum albumin (BSA) standard for protein quantitation (15–17).

1. Add the same amount of each internal standard (e.g. 250 pmols) to each of your sample.

2. Add Methanol-Chloroform in a solvent/sample ratio of 3:1 depending on the volume of the aqueous humor specimen (see Note 6):

   e.g. add 300 μL of Methanol-Chloroform to 100 μL of aqueous humor sample.

3. Homogenize for 2 minutes using a glass homogenizer (see Note 2).

4. Add Chloroform in a Chloroform/sample ratio 2–1:1 depending on the volume of aqueous humor specimen in Step 3.1(2):

   e.g. add 100–200 μL of Chloroform if the volume of aqueous humor sample in Step 3.1(2) was 100 μL.

5. Homogenize as in Step 3.1(3).

6. Vortex for 30 seconds.

7. Centrifuge samples at 20,000 x g for 15 minutes at 4^oC.

8. You should have obtained phase separation into (see Notes 7–8):

   1. Superior aqueous layer (aqueous-soluble compounds in methanol-water)
   2. Inferior organic layer (lipids in chloroform).

9. Place the lipid fractions into new tubes. To prevent contamination by proteins due to repeated pipetting through the aqueous phase, use a new tip each time

10. Aliquot each lipid fraction into 4 parts.

11. To remove chloroform dry lipid extracts in a centrifugal vacuum concentrator for 90 minutes or longer if needed (see Note 9).

12. Dried lipids can be stored at −80^oC until mass spectrometric analysis (see Note 10).

3.2. Mass Spectrometric Analysis

1. To prepare lipid samples and external standard calibrators (see Notes 10–11):

   1. Reconstitute dried lipid samples obtained in Step 3.1(12) with Acetonitrile:Isopropanol, except for samples dedicated to analysis of sphingomyelin that should be reconstituted in Lithiated Chloroform:Methanol. Vortex for 1 minute.
   2. Create each external standard calibration curve (e.g. 6-point from 0 to 300 pmol) in Acetonitrile:Isopropanol except for sphingomyelin standard that should be diluted in Lithiated Chloroform:Methanol. To each calibrator add the same amount of matching internal standard as added to samples in Step 3.1(1).

2. For different sphingolipids set the following parameters: see Tab. 1, in Thermo Scientific Xcalibur and Tunes Software.

3. Set the following parameters for ESI:

   1. Flow rate infusion: 10 μL/min
   2. Spectrum recording: 2 min
   3. Scan: 0.5 sec
   4. Ionization voltage: 1.55 kV for Positive ion mode or 1.3 kV for Negative ion node
   5. Gas pressure: 0.45 psi for Positive ion mode or 0.6 psi for Negative ion mode

4. For quality control: run lipid standards separately before and after running the samples. Moreover, a blank (Acetonitrile:Isopropanol) can be run before and after each sample to exclude carryover.

3.3. Data Analysis

3.3.1. Raw Data Import

1. Raw Data Methods > Raw Data Import. The software supports such vendor-specific formats as mzXML, NetCDF, Waters Raw, Agilent CSV and Thermo CSV.

2. Select and open the desired spectra. The spectra selected should appear in the left half of the interface (Raw Data Files) once successfully imported.

3.3.2. Mass Detection

The purpose of this step is to create a list of masses for each raw file. The settings depend on how the data were acquired and the scan quality. For centroided data, the Centroid Mass Detector is applied and assumes that every signal above the specified noise level is a peak.

1. Raw Data Methods > Peak Detection > Mass Detection

2. Adjust the parameters according to data resolution and precision:

   1. MS level: depends on type of scan. 2 for NLS and 1 for PIS or full MS1 scans
   2. Polarity: depends on type of ionization
   3. Mass Detector: Centroid
   4. Noise level = 0
   5. Mass list name: masses

3. Other mass detectors may be used as the exact mass detector: local maxima, recursive threshold or wavelength transform mass detectors.

3.3.3. Chromatogram Builder (see Note 13)

After running the mass detection, the chromatogram builder makes a constant chromatogram across the scans for each mass. While ideal setting for each chromatogram may change depending on the data set, common baseline settings for several classes are given.

The min time span parameter specifies the minimum time during which a peak must be detected to be included in the chromatogram. Min height is determined by the least possible intensity of the highest data point. Any chromatogram found below the set level is rejected as noise. The m/z tolerance is the greatest allowed m/z difference between data points to be grouped into the same chromatogram.

• 1Raw Data Methods > Peak Detection > Chromatogram Builder
• 2
  Parameters:

  1. Mass list: masses
  2. Set filters (MS level and polarity should be the same as the parameters for mass detection):

Positive Ion Mode Species

1. Min time span: 0

2. Min height: 5.00E3

3. m/z tolerance 0.3

Negative Ion Mode Species

1. Min time span: 0

2. Min height: 5.00E2

3. m/z tolerance 0.3

• 3The completed chromatograms now should appear under the Peak List column on the right side of the interface. Clicking on a completed chromatogram should display a window with the ID, average m/z, retention time, peak shape, height and area in table form.
• 4Troubleshooting

1. If the chromatogram table appears blank, the parameters may have been set too high, causing all data to be discounted as noise. Manually inspect the spectra to adjust the parameters for noise level.

3.3.4. Isotopic Peak Correction

The isotopic peaks grouper leaves one representative peak for a group of isotopes.

1. Peak List Methods > Isotopes > Isotopic Peak Grouper

   1. m/z tolerance: 0.05
   2. RT tolerance: entire duration of analysis
   3. Monotopic shape : checked
   4. Maximum charge : 2
   5. Representative isotope : Lowest m/z

3.3.5. Identification

Identification of the generated peaks is done using either custom databases or those provided by the software. Below are the instructions for identifying lipids from a custom database. If no custom database is available, the software provides an adduct ion search, a peak complex search and an online database search. Note that a custom database may be created by the user (see Section 3.3.7).

1. Peak list methods > Identification > Custom Database Search > Select the correct database

2. Parameters:

   1. Field separator:, (comma)
   2. Field order: ID, m/z, Retention time, Identity, Formula
   3. Ignore First List: Checked
   4. m/z tolerance: 0.3 (see Note 14)
   5. RT tolerance: the entire duration of the direct infusion analysis

3.3.6. Export of CSV file

1. Select file (individually) > Peak list methods > Export/Import > Export to CSV

2. Parameters:

   1. Filename
   2. Field separator:, (comma)
   3. Export common elements: Export row m/z only
   4. Export identity elements: Name only
   5. Export data file elements: Export peak area only

3. After the files are exported, they can be combined and further analyzed and quantified outside of MZmine if needed (see Section 3.3.8). However, MZmine offers several analysis and visual representation methods. Other export formats can be selected- such as MzTAB, XML and SQL which may open alternatives for later analysis.

3.3.7. Creating a Class-Specific Database (see Note 15)

Each lipid class will require a corresponding database which contains the attributes to each species. These databases can be obtained from a variety of internet sources-such as the NIH’s LIPIDMAPS. Guidelines for using LIPIDMAPS are given.

1. Visit LIPIDMAPS website > Resources > Classification

2. Open the desired main class from the eight categories

3. At the bottom of the list find the Download results tab, choose CSV format and all pages.

4. Open the file in Excel and place the columns in the order of ID, m/z, Retention Time, Identity (Common name), Formula

5. Save as .csv

6. The new custom database can now be selected in the custom database search interface for identification.

3.3.8. Post-MZmine Analysis

For experimental purposes it is often useful to organize and analyze the data as presented in the MZmine-exported file. Further manipulation is often required to analyze data from large experiments or for selecting species for detailed study.

1. Ratiometric Quantification: The quantitation of different sphingolipid species is obtained by the use of constant amount of non-naturally occurring internal standard together with calibration curves obtained with different amounts of naturally-occurring species. To compensate for some limitations of the method e.g. matrix effect or machine instabilities, instead of analyte or external standard calibrator areas, their ratios to respective internal standard areas are used.

   1. Construct external standard calibration plots of concentration of external (C_ES) standard (x-axis values) versus area ratio of external (A_ES) and internal (A_IS) standard (A_ES/A_IS) (y-axis values).
   2. For each external standard calculate linear regression curves.
   3. To determine analyte concentration from obtained linear regression equation ($y=ax+b$), substitute y with area ratio of analyte (A_A) and internal standard (A_A/A_IS) in the sample of interest (see Note 16).
   4. The amount of lipid species obtained from the ratiometric quantification analyses above is then divided by the total amount of protein in the sample for normalization. The final amount of lipid is expressed in pmol of lipid/μg protein.
2. Multi-sample analysis: The nature of experimentation often requires manipulations of parameters across different sheets to obtain a single data point. For lipidomics this may include selecting the lowest m/z and manually averaging the associated peak areas for a single species in many data sheets in tandem, quickly becoming a very time-consuming task. Third party software exists but may be difficult to locate and secure. As such, it can be necessary to design simple programs (e.g. in Excel’s VBA language). Some useful program schematics follow:

   1. Programs which will average the m/z and peak area values for each occurrence of a certain named species, assigning results to a new sheet.
   2. Programs which can indicate the frequency of occurrence of identified peaks across multiple sheets from different replicates.
   3. Programs which conduct a search comparison between data sheets, indicating the species which are unique to a set or present in both sets.
   4. Programs which can carry out other statistical analyses.