SUPPLEMENTAL DATA

Direct Derivatization and Quantitation of Ultra-trace Gibberellins in

Sub-milligram Fresh Plant Organs

Dongmei Li1, Zhenpeng Guo1,*, and Yi Chen1,2,*

1Key Laboratory of Analytical Chemistry for Living Biosystems, Institute of Chemistry, Chinese Academy of Sciences,

Beijing 100190, China

2Beijing National Laboratory for Molecular Sciences, Beijing 100190, China

*To whom correspondence should be addressed.

Prof. Yi Chen

E-mail: chenyi@iccas.ac.cn

Tel.: +86 10 62618240, Fax: +86 10 62559373.

Dr. Zhenpeng Guo

E-mail: guozhenp@iccas.ac.cn

Tel.: +86 10 82615622, Fax: +86 10 62559373.

Supplemental Methods

Chemicals and reagents

Gibberellin (GA) and isotope deuterium-labeled GA (GAi and dGAi, i=1, 3, 4, 5, 6, 7, 8, 9, 15,

20, 29, 34, 44, 51) were all purchased from Olchemim Ltd. (Olomouc, Czech Republic).

Formic acid (FA) and HPLC grade methanol (MeOH), acetonitrile (ACN) and ethyl acetate

were from Fisher Scientific (Waltham, MA, USA). Analytical grade hydrochloric acid (HCl)

was obtained from Beijing Chemical Works (Beijing, China), and N-(3-

dimethylaminopropyl)-N’-ethylcarbodiimide hydrochloride (EDC) from Sigma (St. Louis,

Mo, USA). C18 SPE cartridges (3 mL, 500 mg) were obtained from Thermo Scientific

(Bellefonte, PA, USA). Pure water produced by a Millipore Milli-Q academic system

(Billerica, MA, USA) was used throughout the study.

Preparation of solutions

Stock solutions of each GAi and dGAi were prepared in MeOH at the concentration of 10 μM,

and stored at -20 °C. To prepare a working mixture of all 14 GAs, each at 100 nM, combined

100 μL of each GAi stock solution and added 8.6 mL MeOH to form a 10 mL solution of

GAs. A working mixture of all 14 dGAs was prepared in the same way. The working solutions

of GAs and dGAs at different concentrations were then obtained by dilution of the 100 nM

solution with MeOH just before use. EDC solutions at different concentrations were prepared

by dissolving EDC powder in water which was acidified to pH 4.5 with 1 M HCl. For

investigating the effects of solvents on the labeling reaction, EDC solutions were prepared by

dissolving EDC powder in different solvents, including MeOH, ACN, methylene chloride,

water-MeOH and water-ACN mixtures.

Extraction of GAs

Frozen plant samples were placed in a 600 µL eppendorf tube and 2 fmol of each dGA were

added as internal standard (IS). The resultant was manually ground to fine powder using a

glass rod and extracted with 100 µL solvent composed of 75% MeOH, 20% H2O and 5% FA.

The mixture was shaken on a shaker (MS 3 digital, IKA®-Webke GmbH & Co.KG, Staufen,

Germany) at 4 °C and 500 rpm for 2h, plus 10 h unshaking extraction at 4 °C. After

centrifugation at 10000×g for 10 min at 4 °C (Allegra 64R Centrifuge, Beckman Coulter Inc.,

Brea, USA), the supernatant was collected, and the lower pellet was re-extracted twice with

50 µL MeOH for 10 min each. The combined supernatant was evaporated under a nitrogen

gas stream using a nitrogen evaporator (MD 200 Sample Concentrator, Allsheng Instruments

CO., Ltd., Hangzhou, China) and re-dissolved in 20 µL H2O which was acidified to pH 2.5

with 1 M HCl. It was further extracted with ethyl acetate (3×60 µL). The ethyl acetate phase

was combined, and dried under nitrogen gas for the further derivatization.

Labeling procedures

To label trace analytes, an aliquot of 10 μL GA standards was transferred to a 600 μL

eppendorf tube. After dried under a nitrogen stream, 50 μL of 20 mM EDC aqueous solution

was added and the tube was shaken at 35 °C and 750 rpm for 1.5 h. In case of labeling a

prepared plant sample, the resultant mixture was ultrasonicated for 5 min prior to further

shaking-based reaction. The resulting solutions, after centrifugation, were ready for liquid

chromatography-electrospray ionization-tandem mass spectrometry (LC-ESI-MS/MS) or

otherwise stored at -20 °C for later analysis.

To prepare EDC-GA3 for FT-IR characterization, GA3 was weighed at 20.8 mg (0.060 mmol)

and dissolved in 10 mL H2O at pH 4.5, followed by addition of 1.1082 g (5.8 mmol) EDC.

After reaction at 35 °C for 4 h, the solution was loaded onto a C18 SPE cartridge which was

pre-conditioned with 9 mL MeOH, 9 mL 80% MeOH and 9 mL H2O, and washed with 6 mL

H2O to remove the excessive EDC. The target product of EDC-GA3 was collected by elution

with 30 mL of 80% MeOH and dried first in a rotary evaporator (RE-52AA, Zhenjie

Instruments CO., Ltd., Shanghai, China) at room temperature and then in a freeze dryer (FD-

1D-50, Biocool Instruments CO., Ltd., Beijing, China). The resulting residue was collected

and ready for later use.

Characterization of EDC-GA3 by FT-IR

The characterization was performed on an FT-IR instrument, model Tensor 37 from Bruker

Optics (Ettlingen, Germany). The freeze-dried EDC-GA3 was mixed with dry KBr powder

and then pressed into a thin KBr disk for FT-IR measurement. Samples analyzing were

acquired at wavelength between 400–4000 cm-1 with a resolution of 4 cm-1 and the scanning

interval of 1.929 cm-1.

Separation-based quantification

All separation-based assays and identifications were performed on LCMS-8040 and LCMS-

8050 (Shimadzu, Kyoto, Japan). The latter has an LC unit consisting of CBM-20A system

controller, LC-30AD pump, DG U-20A5R degasser, SIL-30AC autosampler, and CTO-30A

column oven, coupled to a quadrupled type tandem mass spectrometer (LCMS-8050) via an

ESI interface. The former is much the same as the latter with one grade lowered components

such as LC-20AD pump, DGU-20A3R degasser, SIL-20A autosampler and CTO-20AC

column oven, coupled to a quadrupled type tandem mass spectrometer (LCMS-8040) via also

an ESI interface. The LCMS-8050 and LCMS-8040 system are both controlled by

LabSolutions LCMS Ver.5.6 (Shimadzu, Kyoto, Japan).

A sample (10 μL) was injected into a reversed-phase packed column (XR-ODS, 50 mm × 3.0

mm I.D., 2.2 μm, Shimadzu) installed for either LCMS-8050 or LCMS-8040. It was eluted at

a column temperature of 40 °C and flow rate of 0.3 mL/min, with a binary solvents of 0.1%

FA in water (A) and 0.1% FA in ACN (B) at programed gradient from 99:1 A:B (v/v) to 50:50

A:B (v/v) over 31 min, and positive ion species were detected by MS via MRM mode. The

MS parameters for LCMS-8040 were set as follows: nebulizing gas flow at 3 L/min, drying

gas flow at 15 L/min, DL temperature at 300 °C, heat block temperature at 450 °C and CID

gas at 230 kPa; while those for LCMS-8050 were nearly the same except for two new factors

of heating gas flow at 10 L/min and interface temperature at 300 °C.

Figure 1. MRM chromatograms (transition of 504.30 → 388.20) of EDC-labeled 1 pM GA1

(black line) and condensed EDC-labeled 0.1 pM GA1 (red line) obtained with LCMS-8050.

Investigation of the lowest labeling concentration of GAs was limited by the sensitivity of MS

instrument. By a moderately sensitive instrument of LCMS-8040 (Shimadzu Co, Kyoto,

Japan), the reaction could reach at least 10 pM GA1 standard. By using a more sensitive

instrument such as Shimadzu LCMS-8050, the direct labeling concentration could further be

lowered to 1 pM GA1. This looks not the lowest reactable concentration but an even advanced

instrument was required to directly prove it. At present, the deduction was confirmed by the

following experiment:

GA1 at 0.1 pM reacted first with EDC, then was frozen with liquid nitrogen and completely

dried in a freeze dryer (extremely low temperature stops the reaction), and finally re-dissolved

in proton-free solvent of acetonitrile (causing no reaction) to form 1 pM EDC-GA1 and

subjected to MS detection. The result shows that the peak intensity and retention time are

comparable with that from 1 pM GA1.

Supplemental Figure 2. Possible reaction pathways and bypass for labeling GA with EDC. 1

GA3; 2 EDC; 3 O-acylurea intermediate and O→N migration; 4 N-acylurea product; 5-7

isomers of GA3 under base or high temperature conditions; 8 amidation of GA3.

Supplemental Figure 3. Representative MS and MS/MS spectra and the cleavage mode of

EDC-GA3. The MS/MS spectrum shows fragmentation patterns for EDC-GA3.

The N-acylurea structures of EDC-GAs were confirmed through their fragmentation rules in

ESI-MS/MS. For example, [(EDC-GA3)+H]+ loses two fragments of m/z 71 and 45 attributing

to CH3CH2NHC=O and CH3NHCH3 from the tertiary amino unit of EDC, respectively. The

abundant fragment at m/z 129 is attributed to the N-acylurea unit. The same rules were

followed when GA3 was replaced by other GAs.

Supplemental Figure 4. Fourier Transform infrared spectra of GA3 (lower line) and EDC-

GA3 (upper line).

The stretching vibrations of hydroxyl and antisymmetric stretching of carboxyl in free GA3,

centered at 3449.06 and 1743.13 cm-1, respectively, disappear after reaction with EDC, with

two new bands appearing at 1652.82 and 1563.89 cm-1 corresponding to C=O (amide I)

stretching and N-H (amide II) vibrations, respectively.

Supplemental Figure 5. Effects of temperature (-4, 15, 25, 35, 50, 60, 70 and 90 °C) and

time (0–150 min) on the labeling reaction of GA3 with EDC. The data were obtained with

LCMS-8040. The reaction concentration of GA3 and EDC were 1 nM and 10 mM,

respectively.

(A) Reaction temperature (°C) at: (■) -4; (●) 15; (▲) 25; (▼) 35; (♦) 50.

(B) Reaction temperature (°C) at: (■) 60; (●) 70; (▲) 90. Solid lines refer to the target

product of EDC-GA3, while dash lines refer to the isomer of EDC-GA3.

The target peak area of EDC-GA3 changes in three formats: At and below 15 °C, the reaction

cannot reach its equilibration state even after 150 min, giving a continuously increasing plot

of the peak area vs. time. The plot slop becomes very small at -4 °C, and the target peak is not

more detectable if the temperature is further lowered, implying that temperature can serve as a

way to control the reaction in case required. At temperatures between 25 and 50 °C, the peak

area of EDC-GA3 increases first with time and then turns to a plateau with a turning point

shortened from ca. 90 min at 25 °C to 45 min at 50 °C. Further increase of the temperature

shows negligible effect on reducing the reaction time but results in a decrease of the target

peak area (solid lines) and emerging of an unexpected peak (dash lines). This phenomenon

should be caused by the isomerization of GA3 at a high temperature due to the lactone

rearrangement. The structural rearrangement of GA3 was further confirmed by first incubating

free GA3 at 90 °C and then reacting with EDC after cooled to 35 °C. The two ionic species

were indeed detected which was consistent with that of directly labeling GA3 at high

temperature.

Supplemental Table 1. Selected reaction monitoring conditions for EDC-GAs and EDC-

dGAs on LCMS-8050.

AnalytesQuasi-molecular

ionsFragment ions Q1 Pre Bias (V)

CE

(V)Q3 Pre Bias (V) RT (min)

EDC-GA1 504.30 388.20-26.0 -34.0 -18.0 12.3

EDC-dGA1 506.30 390.20

EDC-GA3 502.25 431.30-26.0 -24.0 -21.0 12.1

EDC-dGA3 504.25 433.30

EDC-GA4 488 .30 372.25-25.0 -35.0 -26.0 19.6

EDC-dGA4 490.30 374.25

EDC-GA5 486.30 370.20-18.0 -35.0 -26.0 16.1

EDC-dGA5 488.30 372.20

EDC-GA6 502.30 386.15-26.0 -36.0 -26.0 14.5

EDC-dGA6 504.30 388.15

EDC-GA7 486.30 370.20-25.0 -32.0 -25.0 19.4

EDC-dGA7 488.30 372.20

EDC-GA8 520.30 404.15-38.0 -37.0 -29.0 7.7

EDC-dGA8 522.30 406.15

EDC-GA9 472.30 356.15-24.0 -34.0 -24.0 21.9

EDC-dGA9 474.30 358.15

EDC-GA15 486.30 370.25-18.0 -35.0 -25.0 23.2

EDC-dGA15 488.30 372.25

EDC-GA20 488.30 372.15-25.0 -35.0 -26.0 16.4

EDC-dGA20 490.30 374.15

EDC-GA29 504.30 388.20-26.0 -35.0 -27.0 8.5

EDC-dGA29 506.30 390.20

EDC-GA34 504.30 388.20-26.0 -35.0 -27.0 17.5

EDC-dGA34 506.30 390.20

EDC-GA44 502.30 386.20-26.0 -36.0 -27.0 18.2

EDC-dGA44 504.30 388.20

EDC-GA51 488.30 372.15-18.0 -35.0 -25.0 17.7

EDC-dGA51 490.30 374.15

GA, gibberellins; EDC, N-(3-dimethylaminopropyl)-N’-ehtylcarbodiimide hydrochloride; EDC-GAi, EDC-labeled GAi; EDC-dGAi: EDC-

labeled deuterium-labeled GAi; CE, collision energy; RT, retention time.

By MRM mode, quasi-molecular ions [(EDC-GA)+H]+ and the intensive fragment ions were

selected to preform MS detection. The abundant fragment at m/z 129, which was detected

from all EDC-GAs, was not selected because of the serious interference at low m/z region. It

should be noted that some GAs, such as GA1, GA29 and GA34, form an equally weighed

product of EDC-GA, all giving a peak of [(EDC-GA)+H]+ at m/z 504.30 and fragment at m/z

388.20, respectively, but they can easily be differentiated by their different retention times.

Supplemental Table 2. LC gradient programs for both LCMS-8040 and LCMS-8050

systems.

Time (min) Gradient (percentage of mobile phase Ba by volume)

0.01 1

0.01-1.00 Increase linearly to 4

1.00-2.00 Increase linearly to 6

2.00-5.00 Increase linearly to 9

5.00-9.00 Increase linearly to 12

9.00-21.00 Increase linearly to 34

21.00-25.00 Increase linearly to 50

25.00-27.00 100

27.00-31.00 Decrease linearly to 1

LC, liquid chromatography.a Mobile phase A: distilled water with 0.1% formic acid; mobile phase B: acetonitrile with 0.1% formic acid.

LC resolution was simply adjusted by changing the gradient elution program to completely

resolve all the target EDC-GAs available at present. It is better to elute the excessive EDC

before EDC-GAs, which was achieved by a slow increase of the content of ACN or its

gradient during the initial phase of separation.

Supplemental Table 3. Linearity, LOD and LOQ of 14 GAs obtained with LCMS-8050.

Analytesy = ax + b Liner rangea

(fmol)

LODa

(fmol)

LOQa

(fmol)Slope a Intercept b R2

GA1 0.7319 0.009480 0.9989 0.01-10 0.00327 0.0109

GA3 1.129 0.01091 0.9991 0.05-10 0.0125 0.0417

GA4 1.458 0.02724 0.9957 0.01-10 0.00686 0.0229

GA5 0.7906 0.01125 0.9988 0.01-10 0.00984 0.0328

GA6 1.218 0.002130 0.9943 0.01-10 0.00950 0.0317

GA7 1.914 0.04769 0.9995 0.01-10 0.0112 0.0374

GA8 1.310 0.004700 0.9960 0.01-10 0.00727 0.0242

GA9 0.9919 0.006742 0.9993 0.05-10 0.0221 0.0735

GA15 0.6776 0.01639 0.9984 0.05-10 0.0270 0.0901

GA20 0.8206 -0.0009900 0.9984 0.01-10 0.00775 0.0258

GA29 1.241 -0.004950 0.9998 0.01-10 0.00665 0.0222

GA34 0.5386 0.06220 0.9968 0.05-10 0.0195 0.0650

GA44 0.9825 0.003390 0.9974 0.05-10 0.0142 0.0474

GA51 0.5509 0.02120 0.9976 0.01-10 0.0950 0.0317

LOD, limit of detection (S/N=3); LOQ, limit of quantification (S/N=10).a The injection volume is 10 μL.

To evaluate the linearity of the method, eight solutions containing various amounts of each

authentic GA standards (0, 1, 5, 10, 50, 100, 500 and 1000 pM) and a fixed amount of

corresponding dGA (50 pM) as internal standards were prepared, labeled with EDC and then

analyzed. The calibration curves were constructed between the adjusted peak area ratio (y) of

EDC-GA over EDC-dGA and the amount ratio (x) of GA over dGA as the following formula:

{ y=ax+b ¿{y=AGA

AdGA−AGA α d¿¿¿¿

(1)

where a is the slop and b the intersect; AGA, αd and AdGA are the measured peak area of EDC-

GA, its natural abundance with isotope deuterium (ca. 5.5%), and the peak area of EDC-dGA,

respectively; and nGA is the amount of GA and ndGA is that of dGA. Noticeably, the portion

corresponding to the natural deuterium of EDC-GA has to be subtracted from the measured

peak of EDC-dGA in order to increase the linearity of the working curves.

Supplemental Table 4. Precisions (intra- and inter-day) and recoveries for the determination

of 14 deuterium-labeled GAs spiked in Arabidopsis thaliana flowers.

Analytes

Low concentration

(Added: 1 fmol, n = 3)

Medium concentration

(Added: 3 fmol, n = 3)

High concentration

(Added: 10 fmol, n = 3)

RSDa

(%)

RSDb

(%)

Recoveryc

(%)

RSDa

(%)

RSDb

(%)

Recoveryc

(%)

RSDa

(%)

RSDb

(%)

Recoveryc

(%)

GA1 4.19 2.11 87.2 0.840 1.09 85.5 1.59 3.83 90.8

GA3 4.93 5.01 89.8 3.59 5.21 86.4 2.31 2.19 89.7

GA4 3.23 4.27 90.9 3.42 3.98 86.9 7.56 9.50 85.9

GA5 6.53 9.75 85.5 4.81 9.26 85.0 4.76 6.06 89.0

GA6 3.66 11.3 88.8 1.93 8.54 87.0 8.93 7.44 92.1

GA7 8.69 5.36 82.3 3.43 3.60 84.9 2.37 2.04 84.0

GA8 3.38 5.12 65.9 2.87 3.54 63.1 4.98 6.25 72.0

GA9 2.79 2.65 105 4.39 10.2 91.1 1.08 7.83 90.0

GA15 7.81 11.5 97.3 3.21 8.35 93.1 5.20 3.09 91.2

GA20 7.76 5.99 81.7 3.83 4.72 85.5 7.31 1.79 84.8

GA29 3.06 9.23 74.2 2.02 6.12 79.7 1.23 1.73 72.9

GA34 9.18 3.69 96.1 2.67 3.77 90.8 3.15 2.90 91.3

GA44 8.06 8.89 94.0 4.48 6.38 91.8 6.42 7.20 92.2

GA51 1.03 3.00 90.3 7.10 6.76 88.9 3.63 5.55 87.6

RSD, relative standard deviation.a The results for intra-day precisions at three concentration levels;b The results for inter-day precisions at three concentration levels;c The mean value of recoveries for both intra-day and inter-day.

To evaluate the analytical accuracy of the method, whole flower samples of A. thaliana were

spiked with three level of each dGA at 1, 3 and 10 fmol, extracted, labeled with EDC, and

then analyzed. The intra-day variation of peak area of EDC-dGA was determined by three

replicates of each level in one day, and the inter-day variation was determined by repeating

each level for three different days. The recovery of each dGA was calculated based on the

peak area ratio of EDC-dGA that obtained from the spiked sample extracts and the standard

solution containing the same amount of dGA after derivatization.