16O Measurement of Carbohydrates. I. O

Aug 6, 2018 - The 18O/16O ratio at both molecular and positional levels in the carbohydrates of higher plants is a reliable proxy for the plant growth...
0 downloads 0 Views 472KB Size
Download PDF
Subscriber access provided by University of Sussex Library

Article

Novel Position-Specific 18O/16O Measurement of Carbohydrates. I. O-3 of Glucose and Confirmation of 18O/16O Heterogeneity at Natural Abundance Levels in Glucose from Starch in a C4 Plant Ran Ma, Zhenyu Zhu, Bo Wang, Yu Zhao, Xijie Yin, Fengyan Lu, Ying Wang, Jing Su, Charles H. Hocart, and Youping Zhou Anal. Chem., Just Accepted Manuscript • DOI: 10.1021/acs.analchem.8b02022 • Publication Date (Web): 06 Aug 2018 Downloaded from http://pubs.acs.org on August 9, 2018

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 9 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Analytical Chemistry

Novel Position-Specific 18O/16O Measurement of Carbohydrates. I. O-3 of Glucose and Confirmation of 18O/16O Heterogeneity at Natural Abundance Levels in Glucose from Starch in a C4 Plant Ran MA¶, Zhenyu ZHU¶, Bo WANG¶,Yu ZHAO¶, Xijie YINΘ, Fengyan LU§, Ying WANG¶, Jing SUΘ, χ

Charles H. HOCART¶ , Youping ZHOU¶* ¶

Isotopomics in Chemical Biology & Shaanxi Key Laboratory of Chemical Additives for Industry, School of Chemistry & § Chemical Engineering, Shaanxi University of Science & Technology, Xi’an, China; Institute of Earth Environment, ChiΘ nese Academy of Sciences, Xí an, China; SOA Third Institute of Oceanography, Xiamen, China, χResearch School of Biology, Australian National University ABSTRACT: The 18O/16O ratio at both molecular and positional levels in the carbohydrates of higher plants is a reliable proxy for the plant growth environment, and a potential indicator of the plant photosynthetic carbon assimilation mode, and its physiological, biochemical and metabolic status. The lack of exploitable nuclear resonance in 18O and 16O and the extremely low 17O abundance make the NMR-based PSIA (position-specific isotopic analysis) a significant challenge. In this paper, an alternative three-step wet chemistry based method for accessing the 18O/16O of glucose O-3 is presented. The O atoms (OH groups) at positions 1, 2, 5 and 6 were first protected by acetonation (converting glucose to 1,2;5,6-di-O-isopropylidene-glucofuranose). The protected glucose was then esterified at O-3 by thionoformylation. Subsequent Barton-McCombie deoxygenation quantitatively removed the O-3 from the protected sugar. Mass balance was then applied to calculate the 18O/16O of O-3 using the isotopic values of the protected sugar before and after the deoxygenation step. The method is innovative in that i) isolation and purification of individual compounds for 18 O by EA/Pyrolysis/IRMS analysis is unnecessary as the reaction mixture can be analyzed on a GC/Pyrolysis/IRMS; ii) sample quantity is dramatically reduced; iii) the approach to access the O-3 isotopic signal can be easily expanded to other positions within glucose and other sugars. It was shown that O-3 is enriched by 12 mUr relative to the molecular average (O-2~O-6) for a glucose of C4 photosynthetic origin. We highlighted the potential applications of the intramolecular O isotopic heterogeneity of glucose this method revealed.

There are a number of good reasons for using position-specific isotopic analysis (PSIA) of O (18O/16O or 17O/16O ratio) at natural abundance levels in carbohydrates, in particular glucose (the basic unit of sucrose, starch and cellulose in higher plants). Firstly, the sensitivity and reliability of quantitative climate-isotope correlations for a retrospective understanding of the plant growth environment (and by extension, the prevailing climatic conditions)1 may be compromised by correlating climatic signals with the averaged isotopic composition of all Os in the glycosidic unit (CSIA: compound-specific isotopic analysis) and in the mixture of all O atoms in a bulk sample (BIA: bulk isotope analysis), often accessible by analyzing starch, sucrose and cellulose as a whole.2-4 Secondly, the position-specific O is predicted to carry a photorespiration signal as does the position-specific H and C isotope compositions.5-7 Thirdly, when combined with compound-specific measurements, the position-specific isotopic signal can be used to gain a deeper understanding of plant physiology,3, 8-12 in particular the detailed process involved in the transfer of the evaporative enrichment signal of leaf water to the plant synthesized organic matter.13-14 Finally, the position-specific O isotopic signature can assist in

the elucidation of the chemical linkages between major biochemical components of plant cell walls.15-16 PSIA of the Os in sugar is, however, challenging and ideally it should be conducted by quantitative NMR. Unfortunately, unlike the individual 13C and 1H and 2H atoms of carbohydrate 7, 17-19 and lipids20, which can be accessed with NMR, neither 18 O nor 16O has an exploitable NMR signal. Although 17O can theoretically be quantified by NMR (nuclear spin I=5/2) and used as a substitute for 18O in mass-dependent isotopic fractionation processes, the extremely low natural abundance of 17O (0.037%) and the quadrupolar nature of the 17O nuclide which results in peak broadening/poor peak resolution (even after derivatization) for individual 17O signals precludes the use of NMR with current technology.21 Sternberg et al pioneered the first method of accessing the O-2 signal of the glucose unit in cellulose by converting the hydrolytically produced glucose to an osazone (hereafter referred to as Osazone).2 The δ18O2 is then calculated using a mass-balance approach with isotopic ratios determined before and after glucose conversion to osazone. Waterhouse et al went further by taking a stepwise approach to strip the O individually in the form of benzoic acid and then applied a similar mass balance calculation of individual O isotopic composition (hereafter

ACS Paragon Plus Environment

Analytical Chemistry 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

referred to as BA).4 Notwithstanding the novelty of these methods, a few drawbacks remain. Firstly, in the Osazone method a large amount of N is introduced. This can cause serious isobaric interferences for CO analysis as during pyrolysis N2 is generated in an amount-dependent manner, in addition to CO. The Osazone method is limited to the analysis of O-2 only. Secondly, although the BA method can access all individual O atoms, no evidence of the required (near) quantitative conversions necessary for the reliable calculation of the isotopic composition of a particular O atom was provided. Finally, although both methods were developed to access individual O atoms at natural abundance levels, they have only been reported as being applied to cellulose samples derived from plants grown in highly 18O-enriched waters. As the mass-balance based calculation of a particular position unavoidably involves propagation of errors which depends on the number of steps and the isotopic uncertainty involved in chemical conversion, it is uncertain if these methods can be used to detect the intramolecular O heterogeneity (if it exists and if the heterogeneity is not that pronounced), which carries important information on climate, physiology and metabolism. Glucose plays a central role in plant metabolism, and it is likely that the biochemical and physiological processes in which it is involved will be reflected in the position-specific isotopic abundances. In order to probe the valuable metabolic information locked into these position-specific oxygen isotopic compositions, we have developed a chemical strategy for accessing the 18O/16O ratios of all 6 individual O atoms in glucose. We here report a method to determine the natural abundance 18O/16O ratio of the O-3 of glucose (of starch) from a C4 photosynthetic plant. We show that O-3 in the C4 starch is relatively enriched against the molecular average (O-2~O-6) by 12 mUr (throughout this paper we use mUr, equivalent to and in place of ‰, as suggested by Brand and Coplen).26. We discuss the future direction of PSIA-O method development and its potential applications. EXPERIMENTAL SECTION Materials and Chemicals. All reagents (except veratraldehyde) and solvents used for the reported work are of analytical or chromatography grade or better. Glucose (99.9%, Aˊ in Figure 1) was purchased from SigmaAldrich (Prod Nr: 158968, carbon isotope analysis shows that it is sourced from a C4 photosynthetic plant). Veratraldehyde (3,4-dimethoxybenzaldehyde) used as a reference material in this study was of reagent grade (>99%) and commercially available from Acros (Prod Nr: 162610050) and was further purified by column chromatography. Phenyl chlorothionoformate was purchased from Alfa (Prod Nr: L00838). Solvents used in this study including N,Ndimethylformate (DMF), pyridine, cyclohexane, dichloromethane (DCM), acetone and ethylacetate (EtOAc) were obtained from CNW Technology, Sinopharm Chemical and J&K Chemical. n-Tetrabutylammonium peroxydisulfate (n-Bu4N)2S2O8 was freshly prepared by simple extraction of an aqueous solution of potassium peroxydisulfate (K2S2O8, Sinopharm Chemical, 99.5%, J0092) and two equivalents of tetrabutylammonium hydrogen sulfate (Adamas beta, 98%, 48290B) with DCM, drying, filtration, and evaporation gave white crystals of ((n-Bu4N)2S2O8) which were washed (3 times) with deionized water and then dried at 25 °C under vacuum in a desiccator for two days.22

Derivatization Procedure. A three step process was employed to remove O-3 from glucose (Scheme 1). Glucose (A, laboratory moisture-equilibrated) was first converted to DAGF (diacetone glucofuranose, B); then the -OH group at C3 was protected by thioformylation with phenyl chlorothionoformate to form DAGF thionoformate (C); O-3 was then stripped from the glucofuranose ring by BartonMcCombie deoxygenation with (n-Bu4N)2S2O8 to form 3deoxyDAGF (D). Scheme 1. The three-step approach employed in this work to remove O-3 from glucose: i) diacetonation of glucose (A); ii) chlorothionoformylation of O-3 (B); and iii) Barton-McCombie deoxyenation of O-3 with (nBu4N)2S2O8 and HCOONa.

Preparation of 1,2:5,6-di-O-isopropylidene-α-Dglucofuranose (DAGF, B). The Kakitsubata et al23 method was followed in the preparation of B. Briefly, concentrated H2SO4 (> 98%) was added dropwise to a premixed anhydrous glucose and anhydrous acetone stirred vigorously at 5~10 °C. When the acid addition is completed, the temperature was raised gradually to room temperature. After a further 8 h stirring, the solution was cooled and neutralized with NaOH and NaHCO3 to pH 7. After standing overnight, the solution was filtered, and the organic layer was taken to dryness under reduced pressure. The solid residue was redissolved in CHCl3 and washed with deionized water. Recrystallization from cyclohexane gave white crystals with a conversion rate greater than 92%. Preparation of DAGF phenoxythionoformate (C): The Roy et al24 method was followed in the preparation of C. Briefly, phenyl chlorothionoformate was added slowly to a pre-cooled solution (0 °C) of B in dry pyridine. When the yellow precipitate generated during addition was fully dissolved, the reaction temperature was raised to room temperature and stirring was continued for another 30 min. Following the completion of the reaction, the solvent was removed under reduced pressure. The oily mass thus obtained was then dissolved in DCM, washed with deionized water and dried with anhydrous Na2SO4. GC/MS analysis of the crude yellow solid obtained indicated the presence of the desired compound C, a byproduct Cˊ and unreacted B in a peak area ratio of 80:12:8. The yellow solid was then column chromatographed over silica gel using petroleum ether/ethyl acetate (4/1, v/v) to afford C as a colourless solid. Preparation of 3-deoxyDAGF (D): The preparation of D followed the protocol of Park et al.25 A mixture of purified phenyl thionocarbonate (C), HCOONa, (n-Bu4N)2S2O8, Na2CO3 and DMF (2 mL) was stirred at 65 °C for 25 min. On completion of the reaction, the mixture was then poured into water and extracted with DCM. The extract was washed with aqueous NaHCO3 and brine and dried over Na2SO4. GC/MS analysis of the reaction mixture sampled at various stages of

ACS Paragon Plus Environment

Page 2 of 9

electrical signal

1500

24.02 23.98 24.03 sample

M ass 28 M ass 29 M ass 30

electrical signal

sample

24.00 24.01

3000

23.99 24.00

reactor at a flow rate of 0.6~0.8 mL/min (corresponding to 1% H2 in He). In the reactor, compounds were pyrolyzed to CO. After being swept to the ionization chamber, CO gas was ionized to [C16O]+, [C17O]+ and [C18O]+, the 18O/16O ratio of the sample was determined by monitoring m/z 28, 29 and 30 of CO using ISODAT 3.0. To condition the reactor, injections of n-hexane (5 µL) in a split-mode (split ratio 20:1) were performed at regular intervals27 to deposit C on the Ni surface to test for the presence of the carbon surplus phenomenon as observed by Zech and Glaser 28 and to provide carbon to the reactor to exhaust residual O from the reactor to maintain stable and low background values for m/z 28, 29 and 30. Pyrolytic products were also checked for the existence of CO2 by running the IRMS in the CO2 mode i.e. by monitoring the ion currents of m/z 44, 45 and 46. For a routine GC/Py/IRMS run (see Figure 1), five 20 s contiguous CO (99.996%) reference pulses with 30 s gap in between were performed during the first 250 s, followed by the elution of the analytes peaks which were interspaced with two more CO pulses of the same duration and gap depending on the actual sample. At the end of each run, 3 more CO pulses with the same duration and gap were introduced to monitor the stability of the reference gas over the entire run, using CO from a reference CO gas cylinder. To allow evaluation of the instrumental performance and correction for drift, reference materials (veratraldehyde, V), which is a close match to the derivatised sugars) were incorporated within each analytical cycle.

electrical signal

the reaction indicated the presence of the desired product D, a byproduct Dˊ (structurally equivalent to B), and a gradually disappearing reactant C. In the final reaction mixture, only trace amounts of C (2%) were found (85% D and 13% Dˊ). For details of the derivatization procedure, GC/MS analytical conditions and structural identifications of the products, please refer to SI. EA/Py/IRMS analysis. Glucose (A and Aˊ), veratraldehye (V), DAGF and purified compounds B, C, D and Cˊ and Dˊ, ((n-Bu4N)2S2O8, HCOONa, PhOC(=S)Cl and water were analyzed for 18O/16O on a MAT 253 IRMS coupled to a Flash HT (1112 series) EA (elemental analyzer) via a Conflo III (Thermo Fisher Scientific GmbH, Bremen) as detailed in Zhou et al.16 Briefly, analytes were quantitatively and pyrolytically converted to CO in a continuous-flow mode at 1360 oC and analysed for their 18O/16O ratios after being dried in-line with a MgClO4 trap and separated from other gases with an in-line GC column (Part #: 26007900, oxygen separation column, 1 m x 5 mm id, 6 mm od, stainless steel, MolSieve, 2mm unions, Thermofisher Scientific, Bremen). Routine EA/Py/IRMS analysis was preceded by 5 pulses of reference gas (CO), followed by sample peaks and then terminated with another three pulses of CO from a gas cylinder. External precision of repeated analysis of isotopic standards were better than 0.3 mUr for benzoic acid (IAEA-601) and 0.25 mUr for our inhouse water standard (−5.5 mUr, calibrated against IAEA VSMOW-2), respectively, over a period of 3 days of analysis. All isotopic values reported here were the average of 8 replicated analyses. For glucose, the measured isotopic values were corrected for the contribution of the adsorbed moisture by using the method of Zhou et al.16 All isotopic data were recorded and processed with ISODAT 3.0 software (Thermo Fisher Scientific GmbH, Bremen). GC/MS analysis. Reaction mixtures and purified compounds dissolved in DCM were analyzed on an ISO single quadrupole GC/MS system (Thermo Fisher Scientific GmbH) interfaced to a Trace 1300 GC using helium (99.999%) as carrier gas. Separation of the analytes was carried out on an HP-5MS column (30 m x 0.25 mm x 0.25 µm, Agilent Technologies). The oven temperature was held at 80 °C for 2 min and then increased at 10 °C/min to 290 °C, where it was held for 2 min. GC/Py/IRMS analysis. All individual derivatized sugars were analyzed for δ18O on a GC/Py/IRMS system (which consists of a Trace GC 1300 GC, a MAT 253 Plus IRMS, an Isolink conversion unit, a Conflo IV universal interface and AS 1310 autosampler, Thermo Fischer Scientific GmbH, Bremen). Samples and standards dissolved in DCM were injected onto the same column used for GC/MS analysis and components separated using the same temperature program (see above). The separated reaction mixture and external standards were then thermally converted (pyrolyzed) in an oxygen-free atmosphere at 1280 oC to CO in the Isolink conversion unit. The CO and other byproduct gases were then swept to the IRMS by He (99.995%) carrier gas via the Conflo IV universal interface. The pyrolysis reactor (HTC: high temperature conversion) is a non-porous ceramic (Al2O3) tube (length 320 mm, i.d. 0.8 mm, o.d. 1.55 mm) filled with a platinum tube (0.8 mm o.d.) with braided Ni-wires as catalyst. An auxiliary gas was introduced to the reactor to maintain reductive conditions and 18 O integrity within the conversion via a MCD (microchannel device) mounted at the GC splitter holder in front of the

24.00 24.04 24.00

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Analytical Chemistry

Intensity (mV)

Page 3 of 9

0 0

500

Tim e (s)

1000

1500

Figure 1. A routine GC/Py/IRMS run showing the stability of the reference gas CO (numbers over the reference peaks are the δ18O values).

The routine run was set up so that i) isotopic drift of reference CO can be monitored and used to make correction to the (external) standard and sample peaks following the procedure for CSIA-O as outlined by Zech and Glaser28-29; ii) the isotopic compositions of sample peaks were calibrated against external standards (by principle of identical treatment at least in the instrumentation process) and iii) any peak-topeak contamination (intra-run memory effect) and injectionto-injection (inter-run memory effect) could be monitored and minimized. We acknowledge that due to the lack of an accepted international standard for sugar O isotope analysis, the principle of identical treatment of sample and standard was not strictly followed in the sample treatment (up to instrumentation stage). Also the reference material V, sensu stricto, is structurally not equivalent to the derivatized sugars, which may lead to slight difference in pyrolytic behaviour. These factors need to be taken into account when interpreting the isotopic data.

ACS Paragon Plus Environment

Analytical Chemistry Choosing suitable reference material. We tested the suitability of vanillin provided by Thermofisher Scientific. We failed to obtain reproducibility even after many attempts. Instead veratraldehyde (V) was found to give much better reproducibility on both EA/Py/IRMS and GC/Py/IRMS (Table 1). Veratraldehyde was consequently used as an inhouse reference material throughout the instrumentation analysis of our derivatized sugars. Amount effect. When the amount of samples (to be exact, the amount that ends up in the ionization chamber of the mass spectrometer) varies, isotopic composition also varies, a phenomenon observed by both oxidation/reduction-based C/N isotopes analysis in both EA or GC-combustion/reductionIRMS set up and EA or GC/Py/IRMS for O and H isotope analysis. The most relevant work to our current effort was Zech and Glaser28 where methylated sugar mixtures were analyzed by GC/Py/IRMS on a Delta V IRMS. The δ18O of sugars were generally found to be positively correlated with sample size (actual peak area). In our work here, the same positive correlations (Figure 2) were found for the acetonated and formylated sugars. The only exception was veratraldehyde (the in-house reference we employed) which was found to exhibit a negative correlation (see also Table 1 for the correlation coefficients and regression equations). 27 δ 18O (mUr, V-SMOW)

To minimize the measurement error, efforts were made to ensure the amount of injected sample was similar to when stable isotopic values were obtained. This was achieved by concentrating the sample and/or increasing injecting volume when a peak was too small or diluting the sample and/or decreasing injection volume when a peak was too big. This necessitated multiple sample injections to optimize the amounts of particular peaks in a multi-component sample. We followed the procedure as outlined by Zech and Glaser.28, 30 to drift-correct our O isotope measurements for sugar derivatives in this study. All isotopic measurements are reported in the standard delta notation relative to V-SMOW: δ18O = (Rsample − RV-SMOW)/RV18 O/16O SMOW*1000 mUr, where Rsample and RV-SMOW are the ratios of sample and IAEA V-SMOW standard, respectively. Statistical analysis. Statistical analyses were conducted with Excel® 2010. RESULTS and DISCUSSION The desired products, residual reactants, and byproducts were either isolated and analyzed individually on an EA/Py/IRMS (for better accuracy) or as a mixture on a GC/Py/IRMS taking advantage of the resolving power of the GC interfaced to the IRMS. In this work, both instrumental approaches were employed. The isotopic difference for an external reference veratraldehyde (V) obtained from the two approaches was used to calibrate the sample isotopic values after drift correction. By using this dual instrumentation approach, we have been able to harness the better accuracy of the EA, and at the same time demonstrate the feasibility of avoiding the tedious isolation and purification process, therefore minimizing sample loss, and most important of all, any chance of fractionation due to incomplete recovery of target molecules. Stability of reference gas CO. Since our samples were run as mixtures, the retention time and resolution of neighbouring peaks needed to be optimized to minimize interference via peak cross-talk. Over the entire course of a blank/sample run (see Figure 1), the average variation in δ18O was < 0.2 mUr with the worst case being 0.45 mUr. This small variation was used to drift-correct the isotopic compositions of all sample and (external) reference peaks following the procedure described in Zech and Glaser.28-29 Carbon surplus effect. If residual O is released from the ceramic tube during pyrolysis, even if no O is present in the sample, a CO signal (as background) is still possible when carbon is supplied. This phenomenon observed by Zech et al. (2009), also known as carbon surplus effect (CSE), was tested with a mixture of n-alkanes (C16-C30) standards. Only minimal CSE was observed over the period of analysis (see SI). Pyrolytic conversion yield. It is known from earlier work that even at elevated temperatures, the yield of CO is temperaturedependent as there is a dynamic equilibrium between CO and CO2, and if there is N in the sample, generation of N2 is possible, which may be a source of isobaric interference for CO measurement.31-32 We therefore monitored mass 44 (for CO2) and 28 (for N2 when conducting EA/Py/IRMS-based measurements) in the same way as Zech and Glaser28 even though theoretically there is no N in the analytes. The results (see SI) showed that over 98% of the O in the analyte was converted to CO, therefore any isotopic effect associated with incomplete conversion (the generation of CO2) was negligible.

(a) 25

26

23

24 23

δ18OB=2.387Ln(area)+19.64

22

R2=0.8697, n=35

22

δ18OC=2.2989Ln(area) + 20.101

21

R2=0.8363, n=25

20 5

32

(b)

24

25

21

δ 18O (mUr, V-SMOW)

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 4 of 9

10

15

20

2

25

(c)

4

6

8

(d) -3

δ18OV=-10.7Ln(area)+12.574

31

R2=0.7638, n=20 -4

30

δ18OD=0.7697Ln(area)+28.43

R2=0.7919, n=25

29 8

10

12

Area

14

-5 16

18

4

5 Area

6

7

Figure 2. The correlations between drift-corrected δ18O and peak area for compounds B (a), C (b), D (c) and V (d).

Correction of δ18O obtained with GC/Py/IRMS. The Zech and Glaser28 method was followed to correct the measured δ18O using the following the equation (1): δ 18 O = δ 18 Odrf&amt −cor (GC) + +

δ 18 O(EA) − δ 18 Odrf&amt-cor (GC) 1 + δ 18 O(EA)/1000

δ 18 Odrf&amt-cor (GC)* δ 18 O(EA)-δ 18 Odrf&amt-cor (GC) 

(1)

1000 + δ 18 O(EA)

where the subscript dft & amt refers to first drift-corrected and then amount-corrected isotopic values obtained with GC/Py/IRMS and calibrated against the external reference veratraldehye (V) values obtained on both EA-and GC/Py/IRMS. Conversion and preservation of isotopic fidelity. Theoretically acetonation of glucose (A) introduces no oxygen to DAGF (B) (Figure 3), as in an acidic solution, the reaction proceeds such that no nucleophilic displacement (SN) of any of the –OH groups of glucose occurs (Scheme 2). Nevertheless, to confirm the isotopic integrity of glucose in B, we derivatized the same glucose in two acetones with different

ACS Paragon Plus Environment

Page 5 of 9

18 18 δ O values. The δ O values of the two DAGFs produced showed no statistically significant difference (data not shown here), further confirming the preservation of isotopic fidelity of A in B.

B

100 80 60 40 20

Mass 28 Mass 29 Mass 30

electrical signal

3000

Intensity (mV)

Relative Abundance

 (1 − f )ln(1 − f )  7 * δ 18 O predicted C, Rayleigh = 6*  δ 18 O B -ε p s  f   + 1* δ 18 OPhOC(=S)Cl

B

0 0

substitution and elimination mechanism. Using this KIE , i.e., an εp/s = -6 mUr (note that the KIE for the thionoformylation of a stereohindered –OH group at C-3 of the protected sugar should be greater than this as the thionoformylation should proceed at a lower speed than methyl benzoate hydrolysis), the predicted isotopic composition of C following this approach can be formulated as:

1000

Time (s)

7 * δ 18 O predicted C, Mass Balance = ( 6 * δ 18 O B + ε p

0 4

6

8 Time (min)

10

12

14

Figure 3. The GC/MS (lower) and GC/Py/IRMS (upper) chromatograms showing B as the only product of acetonation.

Scheme 2. The mechanism of diacetonation of glucose, highlighting the isotopic intactness of glucose O. Ketalization involves the nucleophilic addition of an alcohol to a ketone under acid catalysis with the elimination of water.

18

18

(1 − f ) ln (1 − f ) f

(2)

where the δ Ο6 Os in C, δ ΟB and εp/s are the isotopic compositions of the 6 Os in C and B at the beginning of the reaction and the intrinsic isotopic enrichment (product versus substrate) during formylation, respectively. There is a lack of a directly applicable isotope fractionation value associated with esterification on the two Os in the ester. The comparable kinetic isotope fractionation (KIE=1.006) is that reported for the hydrolysis of methyl benzoate in alkaline solution34 as esterification and hydrolysis both proceed via a nucleophilic

) + 1* δ

18

O PhOC( =S)Cl (4)

40

eletrical singal

60

Mass 28 Mass 29 Mass 30

Intensity (mV)

80

δ 18 O6 Os in C = δ 18 OB − ε p s

s

The predicted isotopic compositions following the two approaches are listed in Table 2. There is no significant difference within measurement error between the two instrumental methods. Since both predictions are close to the measured δ18O for C, the 5 inherited Os (O-1~O-2 and O4~O-6) of C (from B) were passed onto D without any isotopic change. This justifies the calculation of the isotopic composition of O-3 by applying mass balance between B and D. 100 3000

When B was converted to C, the conversion rate was greater than 92%. The final reaction mixture contains 8% unreacted (residual) B, 80% of C and 12% byproduct Cˊ (Figure 4). The Cˊ is different from C in that it is a phenoxyformate, rather than a phenoxythionoformate of B. Although in C the extra O (the blue C-bound O in Scheme 1) clearly originates from the phenyl chlorothionoformate (the O isotopic composition of which is listed in Table 1), Cˊ contains one additional O (the blue carbonyl O) of unknown origin (although it was suspected to be derived from a rearrangement of the phenyl chlorothionoformate). This makes it difficult to calculate the average isotopic composition of the 6 Os originally from B. Nevertheless, as the 6 Os originally from B were kept isotopically intact during the formylation, the average isotopic compositions of these 6 Os in C and Cˊ should be equivalent to each other and can be calculated from the isotopic composition of B, the fractionation associated with the esterification (thionoformylation) and the percentage conversion (f) to combined C and Cˊ using a Rayleigh fractionation model (note that the conversion is irreversible and occurred in a closed system).

(3)

Alternatively, it can be formulated following a mass balance approach:

Relative Abundance

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Analytical Chemistry

C

C

0 0

300

600 900 Time (s)

1200

20

C’ B

0

4

6

8 Time (min)

10

12

14

Figure 4. The GC/MS (lower) and GC/Py/IRMS (upper) chromatograms showing C as the desired, Cˊ as the byproduct, and B as the residual reactant in the thionoformylation procedure.

When C was deoxygenated to D, the conversion was nearly quantitative (98% conversion rate ). However a byproduct Dˊ (structurally identical to the original B was also produced (Figure 5). Dˊ was isotopically different from reactant B. Even though Dˊ accounts for 13% of the products, the extra O at C-3 of the Dˊ was from the HCOONa as the reaction proceeds via a free radical mechanism (Scheme 1, see Park et al for further details of the free radical mechanism involved in the Barton-McCombie deoxygenation25). The coupling of the . R + formed as a result of heterolytic fissure of the thiocarbonyl .

C-O bond of C and the R - formed via heterolytic fissure of the C-bond of HCOONa leads to the formation of by product Dˊ. This was supported by the isotope mass balance showing (Table 3) that the predicted isotopic composition of Dˊ was equivalent to that of the regenerated Dˊwithin statistical error using the following equation: 6 * δ 18 O = δ 18 O + 5 * δ 18 O (5)

ACS Paragon Plus Environment

D ′, predicted

HCOONa

D

Analytical Chemistry where δ18OB,predicted, δ18OHCOONa and δ18OC are the isotopic compositions of predicted Dˊ, HCOONa and the (purified) C.

60

Mass 28 Mass 29 Mass 30

eletrical singal

80

σ 18 O3 = 62 * (σ 18 O B ) + 52 * (σ 18 O D ) 2

d δ 18 O

D

d δ Odrf&amt-cor (GC)  18

D’

300

Time (s)

600

4

6 8 Time (min)

10

12

3-deoxyDAGF (D) (mUr)

D' (regenerated DAGF) (mUr)

D' (predicted Difference DAGF) (mUr) (mUr)

(% )

avg

stdev

avg

stdev

68.8

32.99

0.42

29.72

0.44

29.29

0.43

85.7

32.58

0.20

29.36

0.36

28.97

0.39

92.2

32.87

0.24

29.58

0.40

29.21

0.37

98.0

32.84

0.39

29.05

0.35

29.19

-0.14

σ 18 O ≈

Calculation of δ18O3. Theoretically the δ18O of glucose O-3 can be calculated by applying mass balance between either A, or B or C and D. Of the three possibilities, the isotopic measurement of B is considered to be most reliable as i) glucose (A) is hydroscopic so its isotopic measurement has to be corrected for the adsorbed moisture; ii) although C is less hygroscopic (and therefore a more suitable candidate than B as the latter has an unprotected -OH group at C-3) and thus adsorption of moisture is less of a problem, the uncertainty associated with the byproduct Cˊ means that it is less representative of the original glucose isotopic composition; iii) both C and Cˊ contain O from the formylation reagent and it is hard to predict with accuracy the isotopic compositions for C and Cˊ without reliable knowledge of the isotopic effect associated with the formylation process. Applying isotope mass balance, the δ18O3 can be calculated as:

δ O 3 = 6 * δ O B − 5 * δ O D (6) We use both EA/Py/IRMS and GC/Py/IRMS analytical results to calculate δ18O3 (shown in Table 2). The EA/Pyrolysis/IRMS-based measurement gave a ∆18O3 vs 1~6 =12.90 mUr, higher by 3.3 mUr than that of the GC/Py/IRMSbased measurement. The 3.3 mUr discrepancy between the two instrumentation approaches is still close to the estimated propagated error associated with the GC/Py/IRMS-based measurements (see the section immediately below). 18

18

Error associated with δ18O3 measurement. Since δ18O3 is calculated using equation (1) and as the errors associated with B and D measurements are independent of each other, according to the theory of Gaussian error propagation, the error for O-3 (σ18O3) can be calculated as follows:

δ 18 O(EA)-2δ 18 Odrf&amt-cor (GC) 18 σ Odrt&amt-cor (GC) 1000 + δ 18 O(EA)

(10)

+ 1*σ 18 O(EA) ≈ 0.03σ 18 Odrf&amt-cor + σ 18 O(EA)

Using this equation, the estimated maximum σ18O3 is 1.73 mUr. Enrichment of O-3 over the average of 5 non-exchangeable Os (O-2~O-6) in glucose. Due to the hydroscopic nature of glucose, isotopic compositions measured with EA/Py/IRMS should be corrected for the adsorbed moisture (equilibrium between vapour and condensate achieved). We followed the procedure of Zhou et al.16 to calculate the true isotopic composition of glucose used for diacetonation (A in Table 1 and Figure 1):

a All average and standard deviations are calculated based on 8 replicated analysis.

18

(9)

18

14

Figure 5. The GC/MS (lower) and GC/Py/IRMS (upper) chromatograms showing D as the desired, and Dˊas the byproduct in the thionoformylation procedure. Table 3. The percentage conversion of C to D and Dˊ versus and the isotopic difference between the measured and predicted Dˊs. a Conversion of C

2

Since the δ O(EA) and δ O(GC) values are around 30 mUr, the σ18O3 can then be approximated as: 18

D’

2

(7)

δ 18 O(EA)-2δ 18 O(GC) (8) 1000 + δ 18 O(EA)

1000 + δ 18 O(GC)  = ≈1 2 d δ 18 O(EA)  1000 + δ 18 O(EA)   

900

20

0

≈

d δ 18 O 

40 0 0

2

To determine the σ18O for B and D, we take the derivatives of δ18O3 against EA & GC/Py/IRMS-based δ18O values (see equation 1),

D

3000

Intensity (mV)

100

Relative Abundance

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 6 of 9

δ Otrue 18

 18 16  96  δ Omeasured 180 * (1 − f air _ moisture ) + 18 * f air _ moisture       16 − δ 18O  α 1000ln * * f + ) air _ moisture 18 air _ moisture  (  ≈ 96 * (1 − f air _ moisture ) 180

(11)

where δ18Oair-moisture, α and fair-moisture are the isotopic composition of air moisture, the fractionation factor associated with condensation of air moisture onto the sugar surface and the percentage of moisture (by weight) in the glucose, respectively. With this true isotopic value (Table 1), the isotopic composition of the 5 non-exchangeable Os in glucose (O-2~O-6) can then be calculated by assuming a full equilibrium between the carbonyl O (O-1) and the adsorbed moisture (condensate) using

(

)

5* δ 18 O2-6 = 6*δ 18 O true − 1* δ 18 Oair_moisture + ε O1 condensate (12) where εO

condensateis the isotopic enrichment of O-1 (the carbonyl O) of glucose over condensate which can be calculated using the empirical equation of Sternberg et al.35 Note that the calculated δ18Otrue of A (30.01 mUr) is different from that of DAGF (B, 32.70 mUr). This is due to the exchange of the carbonyl O (O-1) with the water generated in the acetonation process and/or the trace quantity of water in the concentrated H2SO4 used as a catalyst. The δ18O2-6 was calculated to be 32.21 mUr (Table 1). The enrichment of O-3 over the five non-exchangeable Os (∆18O3 vs 2~6 = δ18O3 - δ18O2-6) in glucose is calculated to be 12.18 mUr (Table 2). 1

Reasonability of the inferred δ18O3. In higher plants, regardless of their photosynthetic modes, both non-leaf starch and cellulose are synthesized heterotrophically primarily from sucrose exported from the leaf. Prior to being exported to non-

ACS Paragon Plus Environment

Page 7 of 9 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Analytical Chemistry photosynthetic tissues, sucrose is already imprinted with the evaporative isotopic enrichment signal of leaf water8-9, 36-38 at the molecular level and the metabolic isotopic signal at the position-level. The pathway from CO2 fixation by RubisCO in the chloroplast to the biosynthesis of sucrose involves many steps (formation/fission of C-O and P-O bonds, and exchange of O with metabolic water) where an isotope effect can occur. Further complications at both the molecular and positionlevels of isotope distribution also comes from the breakdown of sucrose to hexoses, as catalyzed by invertase and sucrose synthase, the exchange opportunities with local (unenriched or less enriched) water and the fission/formation of C-O and P-O bonds provided by the cycling of sucrose through the futile triose phosphate cycle39 before being incorporated into starch and cellulose.2-4,40 It is therefore highly likely that an oxygen isotope inhomogeneity exists intramolecularly. However, due to the lack of knowledge of these IEs, it is difficult at this stage to predict the oxygen isotope distribution within the glucose/fructose unit of sucrose, cellulose and starch. CONCLUSIONS Accessing O isotope composition (18O/16O or 17O/16O) at a specific position in a carbohydrate poses great technical challenges as the advantage of position-specific isotopic information obtainable with quantitative-NMR isotope measurement at natural abundance level cannot be harnessed due to the lack of exploitable nuclear resonance in either 18O or 16O. The novel wet-chemistry based method reported here offered an opportunity to unlock the valuable metabolic, biochemical and physiological information recorded in the 18 O/16O ratio of the glucose O-3. The strategy of selective protection and removal to access O-3 in glucose can be easily applied to access other O atoms in glucose and other sugars (and even non-carbohydrates). Although our analysis started with 5g glucose, it can be easily scaled down to 50mg glucose or equivalent. Glucose and other monomeric sugars locked in cellulose and starch can be easily analyzed following chemical or enzyme-assisted hydrolysis. The finding that O-3 of glucose in a starch from a C4 plant is about 12 mUr more enriched in 18O than the average of the 5 non-exchangeable Os in the glucose unit highlights the intramolecular heterogeneity of O at natural abundance levels in glucose. Work is ongoing to apply this method to soluble sugars, starch and cellulose synthesized in plants of different photosynthetic carbon assimilation modes and growth conditions.

ASSOCIATED CONTENT Supporting Information Detailed chemistry involved in the removal of O-3 from glucose and carbon surplus effect and pyrolytic efficiency (PDF)

AUTHOR INFORMATION Corresponding Author *Email: [email protected]; Isotopomics in Chemical Biology, School of Chemistry & Chemical Engineering, Shaanxi University of Science & Technology; Xi’an, China. Ph: +86 29 8616 8973.

Author Contributions YZ and RM designed the experiment and prepared the manuscript. RM, YZ, FL, XY, YZ, YW and JS conducted experiments. ZZ, BW and CHH contributed to discussion.

Notes The authors declare no competing financial interest.

ACKNOWLEDGMENT YP acknowledges financial support from the Chinese NSF [41773032], a Shaanxi Provincial Talent 1000 Fellowship and a Shaanxi Department of Education grant [7JS013]. Undergraduate intern Mr Haimiao Zhao was involved in part of the derivatization work. This is contribution #7 from the Isotopomics in Chemical Biology (ICB) group. We thank the two anonymous reviewers whose suggestions have greatly improved the manuscript.

REFERENCES (1) Treydte, K. S.; Schlesser, G. H.; Helle, G.; Frank, D. C.; Winiger, M.; Haug, G. H.; Esper, J. Nature, 2006, 440, 1179-1182. (2) Sternberg, L.; Anderson, W; Morrison, K. Geochim. Cosmochim. Acta, 2003, 67, 2561-2566. (3) Sternberg, L.; Pinzon, M. C.; Anderson, W. T.; Jahren, A. H. Plant Cell Environ., 2006, 29, 1881-1889. (4) Waterhouse, J. S.; Cheng, S. Y.; Juchelka, D.; Loader, N. J.; McCarroll, D.; Switsur V. R.; Gautam, L. Geochim. Cosmochim. Acta, 2013, 112, 178-191. (5) Ehler, I.; August, A.; Betson, T. R.; Nilson, M. B.; Marshall, J. D.; Schleucher, J. Proc. Natl. Acad. Sci. U. S. A., 2015, 112, 15585-15590. (6) Tcherkez, G; Farquhar, G. D.; Beadeck, F.; Ghashgaie, G. Funct. Plant Biol., 2004, 3, 857-877. (7) Gilbert, A.; Robins, R. J.; Remaud, G.; Tcherkez, G. G. B. Proc. Natl. Acad. Sci. U. S. A., 2012, 109, 18204-18209. (8) Farquhar, G. D.; Gan, K. S. Plant Cell Environ., 2003, 26, 15791597. (9) Gan, K. S.; Wong, S. C.; Yong, J. W. H.; Farquhar, G. D. Plant Physiol., 2003, 130, 1009-1021. (10) Kahmen, A.; Sachse, D.; Arndt S. K.; Tu, K. P.; Farrington, H.; Vitousek, P. M.; Dawson, T. E. Proc. Natl. Acad. Sci., U. S. A., 2011, 108, 1981-1986. (11) Song, X.; Farquhar G. D.; Gessler, A.; Barbour, M. M. Plant Cell Environ., 2014, 37, 2500-2507. (12) Lehmann, M. M.; Gamarra, B.; Kahmen, A.; Siegwolf, R. T. W.; Saurer, M. Plant Cell Environ., 2017, 40, 1658-1670. (13) Helliker, B. R.; Ehleringer, J. R. Proc. Natl. Acad. Sci. U. S. A., 2000, 97, 7894-7898. (14) Roden, J. S.; Farquhar, G. D. Tree Physiol., 2012, 32, 490-503. (15) Zhou, Y. P.; Stuart-Williams, H.; Farquhar, G. D.; Hocart, C. H. Phytochemistry, 2010, 71, 982-993. (16) Zhou, Y. P.; Yin, X. J.; Yang, H. B.; Su, J.; Yu, H. M.; Wang, Y.; Zhou, S. X.; Zavadlav, S. ACS Sust. Chem. Eng., 2017, 5, 3250− 3260. (17) Schleucher, J. Stable Isotopes: Integration of Biological, Ecological and Geochemical Processes; Griffiths, H., Ed.; Bios Scientific Publishers Ltd.: Oxford, 1998; pp. 63-73. (18) Zhang, B. L.; Billault, I.; Li, X. B.; Mabon, F.; Remaud, G.; Martin, M. L. J. Agric. Food Chem., 2002, 50, 1574-580. (19) Zhou, Y. P.; Zhang, B. L.; Stuart-Williams, S.; Grice, K.; Hocart, C. H.; Gessler, A.; Kayler, Z. E.; Farquhar, G. D. Phytochemistry, 2018, 145, 197-206. (20) Gerothanassis, I. P. Prog. Nuc. Mag. Res. Spec., 2010, 57, 1-110. (21) Robins, R. J.; Pétavy, F.; Nemmaoui, Y.; Ayadi, F.; Silvestre, V.; Zhang, B. L. J. Biol. Chem., 2008, 283, 9704-9712. (22) Jung, J. C.; Choi, H C.; Kim, Y. H. Tetrahedron Lett., 1993, 34, 3581-3684. (23) Kakitsubata, Y.; Aramaki, R.; Nishioka K; Wakao, M.; Suda, Y. Tetrahedron Lett., 2016, 57(10), 1154-1157. (24) Roy, B. G.; Maity, J. K.; Drew, M. G. B.; Achari, B.; Mandal, S. B. Tetrahedron Lett., 2006, 47, 8821-8825. (25) Park, H. S.; Lee, H. Y.; Kim, Y. H. Org. Lett., 2005, 7, 31873190. (26) Brandt, W. A.; Coplen, T. B. Isot. Environ. Health Stud., 2012, 46, 393-409.

ACS Paragon Plus Environment

Analytical Chemistry 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

(27) Jung, J.; Sewenig, S.; Hener, U.; Mosandle, A. Eur. Food Res. Technol., 2005, 220, 232-237. (28) Zech, M.; Glaser, B. Rapid Commun. Mass Spectrom., 2009, 23 3522-3532. (29) Zech, M.; Glaser, B. Rapid Commun. Mass Spectrom., 2008, 22, 135-142. (30) Hitzfeld, K. L.; Gehre, M.; Richnow, H.-H. Isot. Environ. Health Stud., 2017, 53, 116-133. (31) Farquhar, G. D.; Henry, B. K.; Styles, J. M. Rapid Commun. Mass Spectrom., 1997, 11, 1554-1560. (32) Loader, N. J.; Buhay, W. M. Rapid Comm. Mass Spectrom., 1999, 13, 1828-1832. (33) Majoube, M. J. Chimie Physique, 1971, 10, 1423. (34) O’Leary, M. H.; Marlier, H. F. J. Am. Chem. Soc., 1979, 101, 3300-3306.

Page 8 of 9

(35) Sternberg, L.; Ellsworth, P. F. V. PLoS ONE, 2011, 6: e28040. (36) Farquhar, G. D.; Barbour, M. M.; Henry, B. K. Stable Isotopes: Integration of Biological, Ecological and Geochemical Processes; Griffiths, H., Ed.; Bios Scientific Publishers Ltd.: Oxford, 1998; pp. 27-62. (37) Barbour, M. M; Walcroft A. S.; Farquhar, G. D. Plant Cell Environ., 2002, 25, 1483-1499. (38) Cernusak, L. A.; Farquhar, G. D.; Pate, J. S. Tree Physiol., 2005, 25, 129-146. (39) Hill, A.; Waterhouse, J. S.; Field, E. M.; Switsur, V. R., ap Ress T. Plant Cell Environ., 1995, 18, 931-936. (40) Gessler, A.; Tcherkez, G.; Karyanto, O.; Keitel, C.; Ferrio, J. P.; Ghashghaie, J. New Phytol., 2007, 181, 374-386.

Table 1. (i) EA/Py/IRMS δ18O values and standard deviations for standards, reference material and water (used for calculating trueglucose δ18O), (ii) functions describing the area dependence of the drift-corrected GC/Py/RMS δ18O values of B, C, D and V based on 35 or 25 data points obtained for a concentration series, and (iii) final (drift- & amount-corrected) δ18O values and errors for B, C, D and V. EA/Py/IRMS δ18O (mUr) avg stdev (n)

Compound

Amount dependence of the driftcorrected GC/Py/IMRS δ18O (mUr)

Coefficiency (R2)

Area-independent GC/Py/IRMS δ18O (mUr) mean stdev (n)

σ18Odrf-amt (mUr)

EA/Py/IRMS benzoic acid (IAEA601 standard)

24.80

H2O (Xi'an laboratrory moisture)

-25.10

0.11 (8)

veratraldehyde (V)

-1.80

0.22 (8)

glucose (condensate included, A)

27.08

0.18 (8)

glucose (condensate excluded, A, true)R

30.01

O2-6 of glucoseR

32.21

DAGF (B)

32.70

0.17 (8)

DAGF phenoxythionoformate (C)

29.14

0.35 (8)

3-deoxyDAGF (D)

30.12

0.13 (8)

HCOONa

10.79

0.35 (8)

[(n-Bu)4N]2S2O8

10.93

0.38 (8)

PhOC(=S)Cl

16.60

0.41 (8)

GC/Py/IRMS veratraldehyde (V)

δ18O = -10.7Ln(area) + 12.574

0.7638**

-3.10

0.17 (8)

1.7

DAGF (B)

δ O = 2.3827Ln(area) + 19.640

0.8697**

34.76

0.22 (8)

2.1

DAGF phenoxythionoformate (C)

δ18O = 2.2989Ln(area) + 20.101

0.8363**

31.52

0.21 (8)

2.7

35.00

0.32 (8)

2.4

25.64

0.31 (8)

2.7

32.84

0.39 (8)

2.5

29.05

0.35 (8)

2.9

18

DAGF (residual B) DAGF phenoxyformate (C')

δ18O = 0.7697Ln(area) + 28.43

3-deoxyDAGF (D)

0.7919**

DAGF (D', regenerated B)

n: number of replicates; **: correlation is significant at the 0.01 level (2-tailed). R: calculated using δ Oair-rmoisture = -25.0 mUr, a 4% mois18

o

ture in glucose, a temperature of 19 C at which equilibrations between condenstate and air moisture (αEIE=1.0099 according to Majoube)33 and between carbonyl O (O-1) and condensate (αEIE = 1.0265 according to Sternberg et al)35 were reached.

Table 2. Calculated δ18O of O-3, O-3 enrichment over molecular average ∆18O3 and the consistence in the predicted δ18O values of C following the Rayleigh fractionation and the mass balance approaches. Instrumentation

GC/Py/IRMS EA/Py/IRMS

Μ easured δ18O (mUr) of

Predicted δ18O (mUr) of C by

Calculated (mUr)

B

C

C'

Residual B

D

δ18O3

σ18O3

∆18O3 vs 1~6

∆18O3 vs 2~6

Rayleigh fract*

Mass balance**

Diff

34.76 32.70

31.52 29.39

25.64

35.00

32.84 30.12

44.37 45.60

1.73

9.60 12.90

12.18

31.04 29.28

31.31 30.01

0.27 0.11

∗: A fractionation factor αformylation=1.006 and a conversion of 92% are used for the calculation; **: δ18OHCOONa =10.79 mUr and δ18OPhOC(=S)Cl = 16.60 mUr are used for the calculation.

ACS Paragon Plus Environment

Page 9 of 9 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Analytical Chemistry

For Table of Contents Only

ACS Paragon Plus Environment

9