Assessing Gibberellins Oxidase Activity by Anion Exchange/Hydrophobic Polymer Monolithic Capillary Liquid Chromatography-Mass Spectrometry

Bioactive gibberellins (GAs) play a key regulatory role in plant growth and development. In the biosynthesis of GAs, GA3-oxidase catalyzes the final step to produce bioactive GAs. Thus, the evaluation of GA3-oxidase activity is critical for elucidating the regulation mechanism of plant growth controlled by GAs. However, assessing catalytic activity of endogenous GA3-oxidase remains challenging. In the current study, we developed a capillary liquid chromatography – mass spectrometry (cLC-MS) method for the sensitive assay of in-vitro recombinant or endogenous GA3-oxidase by analyzing the catalytic substrates and products of GA3-oxidase (GA1, GA4, GA9, GA20). An anion exchange/hydrophobic poly([2-(methacryloyloxy)ethyl]trimethylammonium-co-divinylbenzene-co-ethylene glycol dimethacrylate)(META-co-DVB-co-EDMA) monolithic column was successfully prepared for the separation of all target GAs. The limits of detection (LODs, Signal/Noise = 3) of GAs were in the range of 0.62–0.90 fmol. We determined the kinetic parameters (K m) of recombinant GA3-oxidase in Escherichia coli (E. coli) cell lysates, which is consistent with previous reports. Furthermore, by using isotope labeled substrates, we successfully evaluated the activity of endogenous GA3-oxidase that converts GA9 to GA4 in four types of plant samples, which is, to the best of our knowledge, the first report for the quantification of the activity of endogenous GA3-oxidase in plant. Taken together, the method developed here provides a good solution for the evaluation of endogenous GA3-oxidase activity in plant, which may promote the in-depth study of the growth regulation mechanism governed by GAs in plant physiology.


Introduction
Gibberellins (GAs), one family of acidic phytohormones, play crucial roles in a number of plant growth and developmental processes, such as seed germination, stem elongation, leaf expansion, and flower development [1,2]. So far, although the number of GAs identified from plants has exceed a hundred, only several GAs, including GA 1 , GA 3 , GA 4 and GA 7 , are considered to function as bioactive plant hormones [3,4]. Previous reports suggest that endogenous developmental cues and environmental signals can affect GAs-controlled plant growth by modulating GAs biosynthesis [5,6]. Therefore, investigation of the regulation of bioactive GAs biosynthesis is crucial for revealing the mechanism of how these hormones control plant developmental processes.
GAs are biosynthesized through an intricate pathway that involves several kinds of enzymes. Recently, the discovery that the GA3-oxidase possesses high catalytic activity in Arabidopsis thaliana (A. thaliana) embryos during seed germination and development demonstrates that GA3-oxidase plays a key regulatory role in controlling the appropriate levels of bioactive GAs during plant growth [7,8,9,10]. GA3-oxidase is able to catalyze the final step of GAs metabolism to produce the bioactive GAs by converting GA 9 and GA 20 into bioactive GA 4 and GA 1 , respectively ( Figure 1) [11]. In some species, GA 9 and GA 20 are also converted to GA 7 and GA 3 , respectively, via 2,3-didehydroGA 9 and GA 5 , probably as side reactions of GA3-oxidase [12,13]. Although the biosynthetic pathway of GAs catalyzed by GA3-oxidase is established and the genes encoding GA3-oxidase have been isolated from A. thaliana, the detection of catalytic activity of endogenous GA3oxidase remains challenging and no method has been reported for the evaluation of endogenous GA3-oxidase so far.
Mass spectrometry (MS)-based methods have been developed for the analysis of some enzymes activities [14,15,16,17]. The advantage of MS-based enzyme assays has the capacity for simultaneously monitoring the multiple pathways catalyzed by the same enzyme. Gas chromatography-MS (GC-MS) method has been employed for the assessment of in-vitro recombinant GA3oxidase activity expressed from Escherichia coli cells [18,19,20]. But the measurement of endogenous activity of GA3-oxidase in plant hasn't been reported, which probably is due to the lack of highly sensitive method to analyze the substrates and products of limited endogenous GA3-oxidase. Therefore, researchers normally examined the gene expression levels of the GAs oxidases to estimate the endogenous amount of GAs oxidases [21,22,23,24]. Liquid chromatography-MS (LC-MS) method has been employed to analyze GAs [21,25,26], however, determination of GAs with particulate-packed columns of 2.1 mm internal diameter (i.d.) normally requires relatively large amounts of GAs, which may not be sufficient for the analysis of endogenous GA3-oxidase activity in plant samples.
Capillary LC-MS (cLC-MS) has been widely used for the ultrasensitive analysis of small molecules [27]. By sensitively analyzing the enzymatic substrates and products, cLC-MS may service as a promising method for the accurate assay of endogenous GA3oxidase activity in small amount plant samples. The porous polymer monolithic columns have several unique properties, such as good pH stability, ease of preparation, and numerous commercial monomers, which can provide good separation resolution towards target analytes with mixed-mode chromatographic retention under isocratic elution condition [28,29,30]. Recently, we prepared a cation exchange/reversed-phase (CX/ RP) mixed mode polymer monolithic column and successfully profiled the acidic phytohormones [31], which suggests that evaluation of endogenous GA3-oxidase may be achieved with cLC-MS by sensitively analyzing its substrates and products.
Herein, we developed a highly sensitive and simple method for the determination of endogenous GA3-oxidase activity using a poly([2-(methacryloyloxy)ethyl]trimethylammonium-co-divinylbenzene-co-ethylene glycol dimethacrylate)(META-co-DVB-co-EDMA) monolithic column coupled with mass spectrometry. With the existence of quaternary ammonium and phenyl groups on the monolithic surface, five GAs (GA 1 , GA 4 , GA 9 , GA 20 and GA 53 ) were baseline separated by anion exchange/reversed-phase (AX/RP) mixed-mode retention. The in-source collision activated dissociation (CAD) of GAs was suppressed in ESI source by optimizing the ESI conditions. Additionally, a large injection volume (1200 nL) was employed to improve the detection sensitivity. Under optimized conditions, LODs for target GAs range from 0.62 to 0.90 fmol, which is, to the best of our knowledge, the lowest LODs of GAs ever reported without derivatization. Kinetic parameters (K m ) of the recombinant GA3oxidase were determined to be 1.7 mM and 14.0 mM for GA 9 and GA 20 , respectively. The developed method was further applied in the evaluation of endogenous GA3-oxidase activity in different types of plant tissues, including rice embryos, rice seedlings, A. thaliana seedlings, and A. thaliana flowers based on the quantification of isotope labeled catalytic products ([ 2 H 2 ]GA 1 and [ 2 H 2 ]GA 4 ) of GA3-oxidase. The results show that endogenous GA3-oxidase activity can be distinctly determined in all the four types of plant tissues examined.

Materials and Methods
Chemicals META (80 wt % in H 2 O, containing 600 ppm monomethyl ether hydroquinone as the inhibitor), DVB (80 wt %, 1000 ppm ptert-butylcatechol as the inhibitor), and EDMA (98 wt % pure, containing 90-110 ppm MEHQ as inhibitor) were purchased from Acros (New Jersey, USA). To remove inhibitors, DVB and EDMA were extracted with 10% aqueous sodium hydroxide and water; after drying with MgSO 4 , they were filtered and distilled under reduced pressure. META was directly used without further purification. Azobisisobutyronitrile (AIBN) and PEG-6000 were purchased from Shanghai Chemical Reagent Corporation (Shanghai, China). AIBN was purified by recrystallization from ethanol at 40uC. 3-(Triethoxysilyl) propyl methacrylate was purchased from Wuhan University Silicone New Material (Wuhan, China). HPLC-grade methanol and acetonitrile (ACN) were obtained from TEDIA (Ohio, USA). The water used throughout all experiments was purified using a Milli-Q water purification system (Millipore, Bradford, USA  9 , and GA 20 were all purchased from Olchemim Ltd. (Olomouc, Czech Republic). Thiourea, acrylamide, formic acid (FA), N,Ndimethylformamide (DMF), ammonium formate (HCOONH 4 ), DL-Dithiothreitol (DTT), phenylmethanesulfonyl fluoride (PMSF), tris(hydroxymethyl)aminomethane, NaCl, NaF, Na 3 VO 4 , and other chemicals of analytical grade used in the experiment were purchased from Shanghai Chemical (Shanghai, China). The standard solution (100 mg/mL) for each analyte was prepared in methanol and stored at 24uC in the dark.

Preparation and Characterization of Poly(META-co-DVBco-EDMA) Monolith
The poly(META-co-DVB-co-EDMA) monolith was prepared by one-step thermally initiated in-situ polymerization. Firstly, porogen PEG-6000 was dissolved in DMF and thoroughly mixed to ensure that the solution was completely homogeneous. Subsequently, META, DVB and EDMA were added in the polymerization mixture. The mixture was then briefly shaken before adding the initiator AIBN (1 wt % of monomers). The polymerization mixture was completely mixed by vortex and ultrasonication to form a homogeneous solution after adding AIBN. Then the resulting solution was filled into the capillary. Both ends of the capillary were sealed by silicon rubber for polymerization at 60uC for 12 h. Finally, the prepared monolithic column was washed with ACN to remove residual reagents, and conditioned by the mobile phase (ACN/H 2 O/FA, 60/40/0.6, v/v/v) at 1 mL/min for 1 h.
Poly(META-co-DVB-co-EDMA) monoliths were also synthesized in Eppendorf vial for the specific surface area measurement. After polymerization, the monolith was cut into small cubic pieces and submersed in ACN/FA (90/10, v/v) for at least 2 h at 60uC to remove the PEG-6000 and non-reacted chemicals. The washing step was repeated three times followed by drying in oven at 80uC. And the resulting monolithic cubic pieces were kept in a desiccator before characterization.
The specific surface area and mesopore size distribution of the prepared poly(META-co-DVB-co-EDMA) monoliths were measured by nitrogen adsorption-desorption experiments using a JW-BK specific surface area and pore size analyzer (JWGB Sci & Tech Co., Ltd., Beijing, China). Before measurement, the monolithic cubic pieces were evacuated in vacuum and heated to 120uC for 3 h to remove the physically adsorbed substances. Specific surface area values were determined by the Brunauer-Emmett-Teller (BET) equation at P/P 0 between 0.05 and 0.35 [32]. Mesopore size distributions were determined from the desorption branches of isotherms based on the Barrett-Joyner-Halenda (BJH) model [33].
The microscopic morphology of the monolith was examined by scanning electron microscopy (SEM) using a Quanta 200 scanning electron microscope (FEI Company, Holland). Permeability measurements were performed using a Shimadzu LC-10AT pump (Kyoto, Japan) under the constant flow mode. ACN was pumped through the prepared monolithic column (30 cm-long, 100 mm i.d., 360 mm o.d.) at flow rate of 2 mL/min. The back pressure was recorded when the pressure stabilized. Permeability (K) was calculated according to Darcy's Law [34].

Expression of GA3-oxidase in E. coli Cells
The coding sequences of Arabidopsis GA3ox1 were cloned into vector pGEX containing Glutathione S-transferase (GST) tag. The E. coli cell lysate expressing GST-GA3ox1 was produced according to previously described protocol [8].

Preparation of Plant Samples for Evaluation of Endogenous GA3-oxidase Activity
The plant samples (rice and A. thaliana) were grown in a growth chamber in a 16 h light/8 h dark photoperiod with 80% humidity under 28uC. Light intensity was fixed to 120 lux/m 2 /s. All plant samples were collected, weighted, immediately frozen in liquid nitrogen, and then stored at 280uC. Plant samples (2-250 mg) were frozen in liquid nitrogen and finely ground followed by extraction with 8 mL/g cell lysis buffer (25 mM Tris-HCl, pH 7.5, 150 mM NaCl, 2 mM DTT, 1 mM NaF, 0.5 mM Na 3 VO 4 , 15 mM b-glycerophosphate, 0.5 mM PMSF) for 30 min, then centrifuged at 10,000 g under 4uC for 5 min to collect the cell lysates. The cell lysates were concentrated to 100 mL at 4uC using ultrafiltration membrane (30 kDa cutoff; Millipore) for the analysis of endogenous GA3-oxidase activity.
Assessing GA3-oxidase Activity with cLC-MS For enzyme assay, the E. coli cell lysates or plant cell lysates (50 mL) were incubated with the substrates (GA 9 and GA 20 for E. coli cell lysates, [ 2 H 2 ]GA 9 and [ 2 H 2 ]GA 20 for plant cell lysates) and the internal standard ([ 2 H 2 ]GA 53 ) in the presence of 100 mM Tris-HCl (pH 7.5) and cofactor mixture (4 mM 2-oxoglutarate, 5 mM L-ascorbate and 5 mM FeSO 4 ) in a total volume of 100 mL. The E. coli cell lysate without GA3-oxidase expression was used as negative control. The incubation was performed at 30uC for 3 h and then 150 mL of acetic acid was added to stop the reaction. The resulting solution was lyophilized to dryness and then redissolved in 510 mL H 2 O/ACN (10/500, v/v). The mixture was vortexed and centrifuged at 12,000 g under 4uC for 5 min. The supernatant was collected and evaporated under mild nitrogen stream at 25uC followed by reconstituting in 100 mL H 2 O/ACN (90/10, v/v) for cLC-MS analysis.
Determination of the substrates and products of GA3-oxidase was performed on a Shimadzu Prominence nano-flow liquid chromatography system (Kyoto, Japan) coupled with a Bruker Daltonics micrOTOFq orthogonal-accelerated time-of-flight mass spectrometer (Bremen, Germany). The Shimadzu cLC system contains two LC-20AD nano pumps, two vacuum degassers, a LC-20AB HPLC pump, a SIL-20AC HT autosampler and a FCV nano valve. The Bruker mass spectrometer is controlled by Bruker Daltonics Control 3.4. Bruker Daltonics Data analysis 3.4 software was employed for the data analysis. Transfer parameters were optimized by direct infusion of an ESI tuning mix from Agilent Technologies (Waldbronn, Germany). Spectra were collected with a time resolution of 50 Hz in the m/z range of 50-600.
The analytical column of poly(META-co-DVB-co-EDMA) monolithic column (100 mm i.d., 360 mm o.d., 30-cm long) was connected to nano-flow liquid chromatographyand conditioned with the mobile phase at a flow rate of 800 nL/min for 30 min. The poly(META-co-DVB-co-EDMA) monolithic column was coupled with a ESI emitter (7 cm625 mm, with a 861 mm tip) (PicoTip company, USA) by a stainless steel union. The fusedsilica capillaries (50 mm i.d.) with different lengths were employed as the 800,1400 nL sample loop (e.g. the volume of 60-cm long capillary was 1200 nL).

Results and Discussion
Preparation of poly(META-co-DVB-co-EDMA) monolithic column Using binary crosslinkers, DVB and EDMA, a CX/RP polymer monolith with a large specific surface area was successfully prepared in our previous work [35]. DVB can enhance the crosslinking degree of monolith [36,37], and EDMA can promote to form a homogeneous polymerization mixture when adding hydrophilic/charged functional monomers. Thus, a monolithic column with large specific surface area can be obtained with facile preparation process. In current study, the GA3-oxidase catalytic substrates and products are organic molecules with carboxyl groups; therefore, the positively charged functional monomer META and hydrophobic crosslinkers DVB and EDMA are empolyed to provide a anion-exchnage/RP (AX/RP) mixed-mode chromatographic retention towards analytes.

Characterization of poly(META-co-DVB-co-EDMA) Monolithic Column
The morphology of the prepared poly(META-co-DVB-co-EDMA) monolithic column was examined by SEM (Figure 2A-C). Figure 2A shows the coarse surface of microglobules, which indicates the existence of mesopores on the monolithic surface. In addition, the continuous monolithic matrix was obtained and attached well to the inner wall of the monolithic column ( Figure 2B) with through-pores of approximate 1.5 mm ( Figure 2C). The distribution of mesopore in the poly(META-co-DVB-co-EDMA) monolith ranges from 3.3 to 4.0 nm, and the specific surface area is 426 m 2 /g ( Figure 2D), which is comparable to the hypercrosslinked monolithic column [38,39,40].
The mechanical stability of the prepared monolithic column was examined with an increase of flow rate from 100 to 1800 nL/ min (linear velocity: 0.21-5.88 mm/s). The results show that the back pressure linearly increases with the increase of the flow rate ( Figure 2E), which demonstrates that the monolithic column bed is stable even under high back pressure (,20 MPa).
The column efficiency of the poly(META-co-DVB-co-EDMA) monolith was evaluated by changing the flow rate from 0.05 to 2.50 mm/s. The effect of the flow velocity on the plate height was examined by using benzene, acrylamide, and thiourea ( Figure 2F). The results show the lowest plate height is approximate 12.2 mm for acrylamide with a flow velocity of 0.84 mm/s (,82,000 plates/ m). Moreover, under high velocity, the theoretical plate heights of these analytes did not significantly increase, which suggests fast separation of GAs can be implemented with this monolithic column.
The reproducibility of the prepared poly(META-co-DVB-co-EDMA) monoliths was evaluated by analysis of GA 1 , benzene and acrylamide. The run-to-run (n = 10), column-to-column (n = 5) and batch-to-batch (n = 5) variations (RSDs) of these monoliths were 0.9%, 5.7%, and 11.4%, respectively, for retention time of these analytes, and 1.4%, 8.8%, and 13.7%, respectively, for the column efficiency of these three analytes. Additionally, the retention time and column efficiency of analytes and the back pressure of monolithic column did not significantly change even after three-month continuous use. These results indicate that excellent reproducibility of the monoliths can be achieved.

Separation of Target GAs on Poly(META-co-DVB-co-EDMA) Monolithic Column
We investigated the separation mechanisms and optimized the separation conditions for five GAs (GA 1 , GA 4 , GA 9 , GA 20 , and GA 53 ) with cLC-MS. Since the existent of salt in the mobile phase can lower the MS sensitivity of GAs, buffer solution was not used in the mobile phase for the subsequent experiment.
Firstly, the influence of mobile phase acidity on the retention of GAs was investigated. Without a buffer solution, the addition of formic acid (FA) will lead to the immediate decrease of the mobile phase pH to approximately 3.5, and further increase of FA content could affect the mobile phase pH slightly. Whereas, the ion strength of the mobile phase will increase by adding FA. Thus, we optimized the FA content in the mobile phase instead of the mobile phase pH. In Figure 3A, when the FA content increased from 0.2 to 1% (v/v), the retention factors (k) of GAs decrease. We reason that with the increase of FA content, the degree of ionization of GAs is suppressed and the ion strength of mobile phase increased, therefore, the electrostatic interaction between GAs and the quaternary ammonium groups is weaken, which exhibits typical anion-exchange retention mechanism towards GAs. The increase of ACN content in mobile phase from 10 to 80% (v/v) results in the linear decrease of the retention of GAs ( Figure 3B), which indicates the RP retention of the monolithic column. Taken together, the results demonstrate the SAX/RP retention mechanism of poly(META-co-DVB-co-EDMA) monolithic column towards negatively charged GAs. Considering the high ACN content can improve the MS response of GAs, ACN content was kept at 60% (v/v) for the subsequent experiment.
In addition, the separation resolution of these analytes was not apparently influenced with a high flow rate ( Figure S2). At the flow rate of 500 nL/min, the column efficiencies of 5 GAs are around 75,000 plates/m. With the increase of flow rate, the column efficiencies of all five GAs slightly decreased from 75,000 to 60,000 plates/m, which indicated that this polymer monolithic column with large specific surface area is suitable for the fast separation of acidic small molecules. Yet, the nano-valves and tubes were frequently blocked after sample injection when the back pressure exceeding 10 MPa. Thus, we chose 800 nL/min as the optimal flow rate for the further methodological investigation. Finally, a 14-min baseline separation of 5 GAs was achieved with this poly(META-co-DVB-co-EDMA) monolithic column using the isocratic elution condition of ACN/H 2 O/FA (60/40/0.6, v/v/v) at the flow rate of 800 nL/min (Figure 4).

Optimization of ESI-MS Conditions for Detection of GAs
In previous reports, the acidic phytohormones were frequently analyzed in multiple reaction monitoring (MRM) mode. Whereas, the fragmentation behaviour of GAs hasn't been rigorously investigated. According to the full-scan spectrum (ESI-MS spectrum) of GA 1 (Figure 5A To circumvent this problem, we optimized the in-source ESI-MS conditions to suppress the in-source CAD occurrence of GAs, including funnel radio frequency (RF), collision RF, hexapole RF, capillary voltage, flow rate and temperature of dry gas. The results show that, under optimized in-source ESI-MS conditions (see details in Supporting Information), the I 145 , I 223 , and I 241 versus I 345 are all lower than 1/10, which increases the detection sensitivity of GA 3 for more than one order of magnitude ( Figure 5B). Consequently, the ion of I 347 (347.1496) was used for the quantification of GA 1 . The ESI source conditions for the analysis of other 4 GAs were also optimized using the same strategy ( Figure S3).

Methodological Establishment
Since the prepared poly(META-co-DVB-co-EDMA) monolithic column possesses large specific surface area of 426 m 2 /g and fast mass transfer kinetics, a large sample injection volume can be implemented on this monolithic column with no decreased separation resolution. The target GAs can be trapped in the front of analytical monolithic column with aqueous sample matrix. Subsequently, the mobile phase with high ACN can elute GAs from the sample zone. The signal/noise (S/N) ratio for most of the GAs increases with the increase of injection volume from 800 nL to 1200 nL, while the injection volume of 1400 nL results in a decrease of S/N ratio (Table S4), which indicates the diffusion of sample zone on the monolithic column. Consequently, sample injection volume was fixed at 1200 nL for further experiments.
The LODs and LOQs of GAs are calculated as the amounts of the analyte at an S/N ratio of 3 and 10, respectively. The results show the LOD and LOQ were 0.62 and 2.00 fmol for GA 1 , and 0.90 and 3.04 fmol for GA 4 , respectively (Table 1), which was comparable to that obtained by derivatization method [31,41]. The detection linearity of the method was investigated using the internal standard of 30.0 fmol [ 2 H 2 ]GA 53 spiked with GA 1 and GA 4 at different amounts ranging from 2.00 fmol to 400 fmol. The calibration curve was constructed by plotting the mean peak The results show that good linearity within the range of 2.00-400 fmol GAs was obtained with coefficient correlation (R) higher than 0.9971 (Table 1).
We then examined the stabilities of the GA3-oxidase catalytic products ([ 2 H 2 ]GA 1 and [ 2 H 2 ]GA 4 ) and internal standard ([ 2 H 2 ]GA 53 ) under the enzyme catalytic reaction conditions. The results show that there were no apparent changes of the MS intensities for these analytes in the reaction mixture even for 24 h, which demonstrates that the method is appropriate for the evaluation of endogenous GA3-oxidase activity by measuring GA3-oxidase catalytic products of [ 2 H 2 ]GA 1 and [ 2 H 2 ]GA 4 .

Measurement of GA3-oxidase Activity by cLC-MS
The kinetic parameters (K m ) of in-vitro recombinant GA3oxidase determined by our developed cLC-MS method was 1.7 mM and 14.0 mM for GA 9 and GA 20 , respectively ( Figure 6), which are consistent with previous reports (1.0-1.5 mM for GA 9 , and 13.0-15.0 mM for GA 20 ) [13,14,15,16]. Additionally, the much lower K m value of GA3-oxidase for GA 9 than GA 20 suggests the higher catalytic activity of GA3-oxidase towards GA 9 than GA 20 .
According to previous reports, the endogenous bioactive GAs could be produced by GA3-oxidase catalysis in plant germinating, seedling and flowering [7,8,9,10,42]. However, up to date, the quantification of endogenous GA3-oxidase activity in plant species has not been established. With our developed cLC-MS method, we determined the catalytic activity of endogenous GA3-oxidase for converting GA 9 to GA 4 in rice embryos (1.0160.24 pmol/g/h, n = 3), rice seedlings (0.2160.03 pmol/g/h, n = 3), A. thaliana seedlings (1.7760.36 pmol/g/h, n = 3), and A. thaliana flowers (0.2260.04 pmol/g/h, n = 3). The extracted ion chromatogram of enzyme catalytic products [ 2 H 2 ]GA 4 in rice embryos exhibited that the products can be clearly identified and quantified with no interfere (Figure 7). [ 2 H 2 ]GA 1 , which was another GA3-oxidase catalytic product by converting [ 2 H 2 ]GA 20 , was not observed even using 1 g plant sample. This result suggests that endogenous GA3oxidase has much lower catalytic activity towards GA 20 than GA 9 , which is consistent with the results obtained using in-vitro recombinant GA3-oxidase (the K m of recombinant GA3-oxidase for GA 20 was approximate one order of magnitude larger than that for GA 9 ).
In addition, we investigated the minimal plant sample required for the quantification of endogenous GA3-oxidase activity. With the decrease of sample amount from 250 to 5 mg, the endogenous GA3-oxidase catalytic activity (pmol/h) linearly decreased ( Figure  S4). The activity of endogenous GA3-oxidase can be distinctly measured from 5 mg plant samples. Compared to the normal LC-MS method, our developed cLC-MS method can achieve much  better detection sensitivity towards GAs [43], therefore, the low activity of endogenous GA3-oxidase can be easily assessed from small amount plant sample by analyzing its substrates and products.

Conclusions
In this study, we developed a sensitive and robust method for the evaluation of recombinant or endogenous GA3-oxidase activity using poly(META-co-DVB-co-EDMA) monolithic column coupled with mass spectrometry. With AX/RP mixed-mode chromatographic retention mechanism, the catalytic substrates and products of GA3-oxidase (GA 1 , GA 4 , GA 9 , GA 20 ) and internal standard of [ 2 H 2 ]GA 53 were well separated. Using this method, we investigated the K m of recombinant GA3-oxidase, which was consistent with previous reports. Additionally, the activity of endogenous GA3-oxidase catalyzing the conversion of [ 2 H 2 ]GA 9 into [ 2 H 2 ]GA 4 was successfully determined in four different types of plant samples. These results demonstrate the endogenous GA3oxidase activity can be distinctly measured by our developed cLC-MS method, which provides appropriate analytical tool for the further study of the growth regulation mechanism controlled by GAs.

Supporting Information
Text S1 Optimization for the Preparation of Poly(META-co-DVB-co-EDMA) Monolithic Column.