Supplementary Material
The Human Saliva Metabolome
Zerihun T. Dame1, Farid Aziat1, Rupasri Mandal1, Ram Krishnamurthy1, Souhaila Bouatra1,
Shima Borzouie1, An Chi Guo2, Tanvir Sajed2, Lu Deng1, Hong Lin1, Philip Liu1, Edison Dong1,
and David S. Wishart 1,2,3
Department of Biological Sciences1, University of Alberta, Edmonton, AB Canada T6G 2E8
Department of Computing Sciences2, University of Alberta, Edmonton, AB Canada T6G 2E8
National Institute for Nanotechnology3, 11421 Saskatchewan Drive, Edmonton, AB, Canada
T6G 2M9
Corresponding Author:
Dr. David S. Wishart
Departments of Biological & Computing Sciences
University of Alberta, Edmonton, AB Canada T6G 2E8
email: david.wishart@ualberta.ca
ph. 780-492-0383, fax 780-492-1071
1
Method S1: NMR Compound Identification and Quantification
NMR Sample Preparation: After thawing and centrifuging, the human saliva samples were
deproteinized using ultrafiltration. Prior to filtration, the 0.5 mL, 3 KDa cut-off centrifugal filter
units (Millipore Microcon YM-3) were rinsed five times each with 500 µL of HPLC-grade
water, then centrifuged at 10,000 rpm for 10 minutes, to remove residual glycerol bound to the
filter membranes. Two 500 μL aliquots of each saliva sample were then transferred into the
centrifuge filter devices. The samples were then spun at a rate of 10,000 rpm for 35 minutes to
remove macromolecules (primarily proteins and lipoproteins) from the sample. A 600 μL aliquot
of deproteinized saliva was placed in a 1.5 mL Eppendorf tube followed by the addition of 70 μL
of D2O and 30 μL of a standard buffer solution (25 mM DSS [disodium-2,2-dimethyl-2silapentane-5-sulphonate], 730 mM imidazole, and 0.47% NaN3 in H2O). The samples (700 μL)
were then transferred to a standard 5 mm thin-walled glass NMR tube for subsequent NMR
spectral analysis. In total, 16 saliva samples were prepared in this manner. The pH of the samples
ranged from 6.5-7.5.
NMR Spectroscopy: All 1H-NMR spectra were collected on a 500 MHz Inova (Varian Inc., Palo
Alto, CA) spectrometer equipped with a 5 mm HCN Z-gradient pulsed-field gradient (PFG)
room-temperature probe. 1H-NMR spectra were acquired at 25 °C using the first transient of the
Varian tnnoesy-presaturation pulse sequence, which was chosen for its high degree of
quantitative accuracy (Saude et al. 2006). Spectra were collected with 128 transients and 8
steady-state scans using a 4 second acquisition time and a 1 second recycle delay.
NMR analysis: Prior to spectral analysis, all FIDs were zero-filled to 64,000 data points and line
broadened 0.5 Hz. The methyl singlet produced by a known quantity of DSS was used as an
internal standard for chemical shift referencing (set to 0 ppm) and for quantification. All 1H-
2
NMR spectra were processed and analyzed using the Chenomx NMR Suite Professional software
package version 7.7 (Chenomx Inc., Edmonton, AB), as previously described (Wishart et al.
2008). Each spectrum was processed and analyzed by two experienced NMR spectroscopists to
minimize compound mis-identification and mis-quantification.
Method S2: GC-MS Compound Identification and Quantification
Extraction and derivatization of polar metabolites: Prior to analysis by GC-MS, 200 μL of 16
thawed saliva samples containing 2 μL of ribitol in water (20 µg/mL) as an internal standard
were extracted with 800 µL of cold HPLC grade methanol and double-distilled water (8:1
vol/vol) and vortexed for 1 min. The sample was kept at 4 °C for 20 min and then centrifuged at
6,000 rpm for 8 min. After centrifugation, 200 µL of the supernatant were evaporated to dryness
using a Speedvac concentrator (Savant Instruments Inc., SDC-100-H, Farmingdale, NY).
Extracted residues were reconstituted in 40 µL of 20 mg/mL methoxyamine hydrochloride
(Sigma-Aldrich) in ACS grade pyridine and incubated at room temperature for 16 hours. Then
50 µL of MSTFA (N-methyl-N-trifluoroacetamide) with 1% TMCS (trimethylchlorosilane)
derivatization agent (Pierce) was added and incubated for 1 hour at 37 °C. The samples were
vortexed twice during incubation to ensure complete dissolution. Derivatized samples were used
for GC MS analysis.
Extraction and derivatization of organic acid metabolites: 1 mL of each 16 saliva samples was
incubated with 200 μL of methoxylamine hydrochloride solution at a concentration of 75 g/L at
60 °C for 30 min. to convert ketoacids to their methoxime derivatives (MO) (Tanaka et al. 1980).
After the MO conversion, 3 drops of 6 N HCI were added to each sample and the pH was
checked to insure that it was 1.0 or below. The mixture was extracted two times, the first time
3
with 3 mL of ethyl acetate and the second time with 3 mL diethyl ether. The supernatant from
the first and the second extraction was then combined. The resulting supernatant was decanted
into a vial and put in a 37 °C heating block under nitrogen for complete evaporation. After
evaporation, the residue was dissolved in 40 μL hexane and 30 μL of N,OBis(trimethylsilyl)trifluoroacetamide (BSTFA) was added. The resulting mixture was placed in a
heating block at 60-90 °C for 10 min after which the BSTFA-supplemented extract was placed
into a GC/MS auto sampler vial using a transfer pipet.
Extraction and analysis of non-polar organic metabolites: for the non-polar organic salivary
metabolites analysis, the saliva samples were saponified with a 2.0 M solution of sodium
hydroxide and 0.5 g of sodium chloride was added to increase the solubility of metabolites in the
matrix. The saponified samples were heated at 100°C for 20 min. The sample tube was cooled,
acidified with 6N hydrochloric acid and vortexed thoroughly. The samples were extracted with
MTBE by vortex shaking for 30 min and centrifuged at 10,000 x g to separate the aqueous and
organic layers (Lochner et al. 1986). Transfer the upper organic layer to a 2 mL vial and
evaporate the solvent under a stream of nitrogen gas. The extracted components from the saliva
were derivatized with methoxyamine hydrochloride and bis (trimethylsilyl)trifluoroacetamide
(BSTFA). GC-MS method was developed on an Agilent Technologies 6890A gas
chromatograph-5973N mass spectrometer (GC-MSD, Agilent Technologies, Palo Alto, CA).
Method S3: DFI/LC MS/MS Compound Identification and Quantification
A total of 16 saliva samples were analyzed via a hybrid direct-flow-injection/liquid
chromatography mass spectrometry (DFI/LC-MS/MS) method using the AbsoluteIDQ kit and
the protocol described in the AbsoluteIDQ user manual. Briefly, saliva samples were vortexed
4
and centrifuged at 10,000 rpm for 10 minutes. 10 µL of supernatant from each saliva sample was
loaded onto the kit’s filter paper substrate and dried under nitrogen. Subsequently, 20 µL of a 5%
phenyl-isothiocyanate solution was added for derivatization and after 20 min incubation the filter
spots were dried using a nitrogen evaporator. Extraction of the metabolites was then achieved
using methanol containing 5 mM ammonium acetate. The extracts were analyzed using an ABI
4000 Q-Trap (Applied Biosystems/MDS Sciex) mass spectrometer equipped with a solvent
delivery system. A standard flow injection protocol consisting of two 20 μL injections (one for
the positive and one for the negative ion detection mode) was applied for all measurements.
MRM detection was used for quantification. The MetIDQ software was used to control the entire
assay workflow, from sample registration to automated calculation of metabolite concentrations
to the export of data into other data analysis programs.
Method S4: Characterization of Vitamins using HPLC-MS/MS
Standards and Reagents -Water soluble vitamins employed in this study (B1, B2, B3, B5, B6,
B6-phosphate, B7, B9, B12 and C) were purchased from Sigma-Aldrich. The internal
isotopically labeled standards including B3-amide* (Nicotinamide-D4), B6* (Pyridoxine-d2
HCl), B7* (Biotin(ring-6,6-d2)), B9* (Folic acid-d4) and C*( Ascorbic acid-13C) were purchased
from Cambridge Isotope Laboratories, Inc. Fat-soluble vitamins employed in this study (A, D3,
D2, 25OH-D3, E, K1, K2, and K3) were purchased from Sigma-Aldrich. The internal standard
Vit K1* (Vitamin K1-d7) was purchased from Santa Cruz. The internal standard Vit E* (Vitamin
E-d6) was purchased Cambridge Isotope Laboratories, Inc.
5
Standard Preparation - For water soluble vitamins, the aqueous stock solutions of B1, B2, B3, B5,
B6, B6-phosphate, B7, B9, B12 and C were prepared weekly. B3-amide*, B6*, B7*, B9* and C*
were dissolved in D2O. All solutions were stored in a freezer in amber vials to protect vitamins
from light-induced oxidation. 7 working standard solutions were prepared fresh daily from stock
solutions. Water soluble vitamin internal standard solution was made fresh to contain proper
amount of B3-amide*, B6*, B7*, B9* and C*. 40 uL of working standard solution was spiked
with 10uL internal standard solution for LC-MS analysis. For fat soluble vitamins, the stock
solutions of A, D3, D2, 25OH-D3, E, K1, K2, K3, E* and K1* were prepared in Methanol
weekly and stored in a freezer. 7 working standard solutions were prepared fresh daily from
stock solutions. Fat soluble vitamin internal standard solution was made fresh to contain proper
amount of E* and K1*. 40 uL of working standard solution was spiked with 10 uL internal
standard solution for LC-MS analysis.
Method S5: Characterization of Thiols, and Nucleotides using HPLC
For nucleotide analysis, saliva (1 mL) samples were extracted with 1 mL of 0.6 M perchloric
acid and then centrifuged at 10,000 rpm for 5 min at 4ºC. The supernatant were collected and
neutralized with 2 M potassium hydroxide. Neutralized samples were further centrifuged at
10,000 rpm 5 min at 0 ºC and aliquots of this extract were used for HPLC injection.
For salivary thiols analysis, 20 μL of saliva was treated with 10 μL of 4 mM NaBH4 (dissolved
in a solution of 333 mL/L dimethyl sulfoxide and 66 mM NaOH), followed by 5 μL of a solution
containing 2 mM EDTA and 2 mM DTT, then 5 μL of 1-octanol, and finally 5 μL of 2 mol/L
HCl. The mixture was incubated for 1 min at 25 °C (to allow reduction of disulfides). 25 μL of
N-ethylmorpholine buffer (2 mol/L, pH 8.0) and 10 μL of 25 mM bromobimane (in 1:1
6
acetonitrile/H2O, v/v) were added to the mixture. After incubation of the mixture for 1 min, 20
μL of the derivatized sample was injected immediately into HPLC Column connected to an
Agilent fluorometer operating at an excitation wavelength of 365 nm and emission wavelength
of 475 nm. The 150 × 4.6 mm Hypersil-ODS column was equilibrated with 30 mM ammonium
nitrate and 40 mM ammonium formate buffer, pH 3.6 (buffer A). The thiols were eluted from the
column with a linear gradient of acetonitrile (buffer B) (0-11% in 18 min at a flow rate of 1
mL/min). The column equilibration time was 5 min, and the column was run at ambient
temperature.
Method S6: Trace Element Analysis Using ICP-MS
A total of 1 mL of saliva sample was aliquoted and sonicated in an ultrasound water bath for 10
min in order to obtain a homogeneous dispersion. The sample was then diluted to 5 mL with 2%
HNO3. Trace element concentrations were determined on a Perkin-Elmer Sciex Elan 6000
quadrupole ICP-MS operating in a dual detector mode. Blank subtraction was applied after
internal standard correction. A four point calibration curve was used to quantify compounds (0,
0.025, 0.050, and 0.100 ppm for Na; 0, 0.25, 0.50, and 1.00 ppm for Ca, Mg, Fe, K; 0, 0.005,
0.010, and 0.020 ppm for the remaining elements). A typical count rate for a 10 ppb Pb solution
would be 150,000 to 200,000 cps. The sample uptake rate was approximately 1 mL/min with 35
sweeps per reading using one reading per replicate and three replicates. Dwell times were 10 to
20 ms for all elements with the exception of as (which was 100 ms). The relative standard
deviations (2σ level) for As, Ni, Pb, and Zn were between 5 and 10%.
7
References
Heitland, P., & Koster, H. D. (2006). Biomonitoring of 37 trace elements in blood samples from
inhabitants of northern Germany by ICP-MS. J Trace Elem Med Biol, 20(4), 253-262,
doi:10.1016/j.jtemb.2006.08.001.
Lochner, A., Weisner, S., Zlatkis, A., & Middleditch, B. S. (1986). Gas chromatographic-mass
spectrometric analysis of volatile constituents in saliva. J Chromatogr, 378(2), 267-282.
Saude, E. J., Slupsky, C. M., & Sykes, B. D. (2006). Optimization of NMR analysis of biological
fluids for quantitative accuracy. Metabolomics, 2(3), 113-123, doi:DOI 10.1007/s11306006-0023-5.
Tanaka, K., Hine, D. G., Westdull, A., & Lynn, T. B. (1980). Gas-chromatographic method of
analysis for urinary organic-acids. I. retention indexes of 155 metabolically important
compounds. Clinical Chemistry, 26(13), 1839-1846.
Wishart, D. S., Lewis, M. J., Morrissey, J. A., et al. (2008). The human cerebrospinal fluid
metabolome.
Journal
of
Chromatography
B,
871(2),
164-173,
doi:http://dx.doi.org/10.1016/j.jchromb.2008.05.001.
8