Carbon dioxide (13CO2, isotopic purity 99atom% 13C, <1.5atom% 18O) was purchased from Cambridge Isotopes (Andover, MA, USA) in a 10 L lecture bottle. CO2 free air (< 1 ppm CO2, 25% oxygen with a balance of nitrogen) was purchased from Minneapolis Oxygen Company (Minneapolis, MN, USA). The amino acid standard mixture was purchased from Thermo Scientific-Pierce (Rockford, IL, USA) and the [13C]-labeled algal amino acid mixture was purchased from Cambridge Isotope Laboratories. All other chemicals were obtained from Sigma-Aldrich (St Louis, MO, USA) at the highest available purity.
Plant Growth and Labeling Conditions
Arabidopsis thaliana Col-0 seeds were surfaced sterilized using 70% ethanol for 1 min followed by 10% bleach containing 0.1% TritonX-100 for 20 min. After being washed with de-ionized water three times, seeds were sown on pre-rinsed rockwool plugs (3.75 cm × 3,75 cm, Grodan, Milton, ON, Canada), and covered with silicone rubber mats (4.7 cm × 4.7 cm, 8608K151, Extreme-Temp Textured Silicone Foam Rubber, 3.175 mm thick, ordered from McMaster-Carr, Robbinsville, NJ, USA) to restrict algal growth. Half strength Gib hydroponic medium  was used for growing the Arabidopsis in the enclosure. To minimize dissolved 12CO2, the medium was degassed using a helium spurge at 60°C for 30 min, capped, and then cooled to room temperature before being injected into the automated enclosure system. The growth parameters for Arabidopsis were 16 h light/8 h dark with 23°C during the day/18°C night air temperature, a constant 65% relative humidity, and a maximum light intensity of 100 μmol m-2 s-1. The lighting was provided by a combination of fluorescence lamps (F96T12 CW VHO, Phillips, Andover, MA, USA) and incandescence light bulbs (40 W frost, Phillips). The internal atmospheric pressure was maintained at 2 kPa above ambient. In order to remove ambient CO2 from the enclosure after it was sealed, the enclosure was purged with CO2 free air until CO2 levels went down to <5 ppm, then 13CO2 was injected under software control into the enclosure until the levels reached 400 ppm. Air was allowed to circulate throughout the entire system, including the attached tubing, for 1 h then the system was purged again two additional times. After the system was thoroughly purged, the 13CO2 level was kept at 600 ppm for the duration of the experiment. Dilution with 12CO2 after 13CO2 labeling was achieved by exposure of three-week-old labeled plants to the ambient atmosphere in the same walk-in growth chamber that housed the 13C-enclosure, with height adjustments to maintain the same light intensity.
This growth enclosure was built from Plexiglass® acrylic sheets (1.27 cm thickness) and had a two-level design (2 cube-shapedboxes total of 30.4 cm × 30.4 cm × 30.4 cm, volume = ~54 L). This design permitted a single enclosure to be used for routine work and the height doubled for experiments where flowering and seed production were to be studied. Viton® gaskets (2 mm thick) were used as the sealing material between the two enclosures and to seal the lid because of its excellent chemical resistance and low gas permeability. A small Plexiglass® acrylic enclosure with a thermoelectric cold plate (CP-031, TE Technology, Traverse City, MI, USA) installed on the back side of its walls was connected to the main enclosure by one inlet acrylic tube (5.08 cm, OD) and one outlet aluminum tube (5.08 cm, OD), as a heat exchanger for the return air. The cold plate was controlled by a temperature controller (TC-48-20, TE Technology) and was maintained at 10°C. This cold box is used to control the humidity of the main enclosure by circulating air from the main enclosure using a software controlled small fan (3.81 cm × 3.81 cm, 12 V) onto the cold plate such that the excess moisture in the air condensed on the cold surface. The detailed design of this dehumidifier can be found in the user's manual (see Additional file 4). Two software controlled solenoid gas valves were installed on the lid of the enclosure for inputs of CO2 gas and CO2-free air. One was a proportional two-way solenoid (EV-P-10-0925, Clippard Minimatic, Cincinnati, OH, USA) for controlling CO2 gas via a pulse-width modulator (Si5HyUdMC2-30 V-2 × 20A, Signal Consulting, Edgewater, MD, USA). The other was a three-way solenoid valve (EVO-3-24, Clippard Minimatic) that served as a switch between 13CO2 gas and CO2 free air. When energized, it switches to the CO2 free-air port for enclosure purging or maintaining enclosure pressure. In addition, a two-way solenoid valve that functions as a pressure relief valve (MME-2PDS, Clippard Minimatic), a pressure sensor (DPG1000DAR35KPAG-1N-I-CC, Cecomp Electronics, Libertyville, IL, USA) and a humidity/temperature sensor (HX94ACW, Omega Engineering, Stamford, CT, USA), were all installed on the lid. For injection of growth medium and withdrawing liquid from the growth tray, another needle valve (B-4JN2, Nupro, Willoughby, OH, USA) was installed on the lid. A standard luer-lock female fitting (McMaster Carr) was soldered to one end of the valve that faces outside the chamber to accept a 60 mL luer lock syringe (BD, Franklin Lakes, NJ, USA). Also, a standard luer-lock male fitting (McMaster Carr) was soldered to the other side of the valve to which a hypodermal needle (gauge 10; McMaster Carr) with a standard luer-lock female fitting was connected. The needle was long enough to slightly touch the bottom of the tray. Two small 12 V DC fans (5 cm × 5 cm) were also installed on the inside of the lid to provide air circulation within the enclosure. A CO2 analyzer (WMA-4, PP Systems, Amesbury, MA, USA) was located outside the enclosure and was connected via Bev-A-Line IV tubing (0.64 cm OD/0.32 cm ID, US Plastic Corporation, Lima, OH, USA). The CO2 analyzer housed a pneumatic diaphragm pump to continuously circulate air at a flow rate of 1 L/min from the growth enclosure through a small air chamber inside the analyzer where the CO2 sensor was located. After measurement, the air passed through a custom ethylene scrubber column (14 cm ×2.3 cm ID glass tube with Teflon compression fitting caps) containing five sachets of potassium permanganate-coated × pellets (2 × 9 g sachets, Ethylene Control) before flowing back to the main enclosure. A manual stack 4-way valve (224-X-PP-ALL-MS2, EVSCO, Libertyville, IL, USA) was used to direct the air coming out from the CO2 analyzer to either the ethylene scrubber or a bypass route, which allowed easy replacement of ethylene scrubber bags in the column while an experiment was in progress. After replacement, the scrubber column was purged with CO2 free air that was directed by a manually operating 3-way valve (H6800 SSL1/16PST, HAM-LET, Solon, OH, USA) located at the top of the CO2 analyzer. CO2 free air containing less than 1 ppm CO2 was run through a soda lime based CO2 scrubber (PP Systems, Amesbury, MA, USA) before being injected into the enclosure for purging purposes and/or for maintaining the slightly positive pressure of the enclosure. To prevent excessive CO2 gas addition beyond desired concentrations, the flow rate of CO2 gas was controlled by a needle valve (SS-SS2-VH, Swagelok, Chaska, MN, USA) set at the lowest rate. The outlet pressure of 13CO2 was controlled by the regulator (Y11-L244ALB, Airgas, Savage, MN, USA) that was attached directly to the lecture bottle. A pressure of ~10 psi was found to be ideal for optimal control of the 13CO2 flow rate. The CO2 free air flow rate during plant growth was controlled only by the restriction of ~100 cm of the attached 1.59 mm OD/0.51 mm ID stainless steel tubing when outlet pressure of the gas regulator (Y12-244 D, Airgas) was set at 100 psi. During the purging of the enclosure, the pressure was increased to 120 psi.
Power supplied to the system
A 12 V DC power supply (Model 1316, Global Specialities, Wallingford, CT, USA) was used to operate the thermoelectric cooler (cold plate), 3 small fans, and the PWM controller for the proportional solenoid. A 24 V DC power supply (Model 6216B, Agilent/HP) as used to control all other solenoid valves, temperature/relative humidity sensor and enclosure pressure sensor. The power supplies, the laptop computer (Toshiba Satellite M 115-S3094) that served as the controller, the data logger, as well as the CO2 analyzer are all directly plugged into a backup power system (Back-UPS RS 1500VA LCD 120 V; APC, Kingston, RI, USA).
System Control Setup
The current signal values from all sensors are acquired by a data acquisition device, NI Compact DAQ (cDAQ-9172, National Instruments) using an analog current input module (NI 9203, National Instruments) and compared to set points by a virtual control program (see Additional files 1 and 5) written in LabView (version 8.2 or above, National Instruments) on the notebook computer/controller. Output signals were triggered by the controller to turn on a relay module (NI 9481, National Instruments) on the Compact DAQ for the control of the CO2-free air solenoid, the fan in the dehumidifier, and pressure relief solenoid, until the values reached their set points. A proportional-integral-derivative (PID) control loop was used to control the solenoid valve for CO2 gas in a proportional manner. The voltage signal generated by a voltage signal module (NI 9263, National Instruments) on the Compact DAQ was transmitted to a pulse-width modulator where the signal was translated into power load to the solenoid valve. In addition, tuning parameters were also programmed into the control loop to deal with the delay in the report of CO2 levels in the enclosure due to the distance between the injection site and the CO2 sensor. To allow correct data acquisition and PID control using the Labview program provided, both the Compact DAQ data logger driver (NI DAQmx 8.2 or above, National Instruments) and the PID control tool kit (National Instruments) were also installed to the laptop computer/controller.
The enclosure pressure was maintained by a control loop in the program that compared actual pressure with the set minimum pressure and energized the air solenoid via a relay switch to inject enough CO2-free air into the system to maintain the set pressure. To return the enclosure pressure from an overpressure value, the pressure relief valve was energized by another relay switch. The enclosure pressure was kept at 2 kPa to prevent ambient CO2 from entering the enclosure. The control loop also allowed for enclosure purging by continuously opening the pressure relief valve until the purge was complete. While purging, the enclosure pressure was set at 1 kPa so that the pressure relief valve would remain continuously open during the purging process.
The fan inside the dehumidifier was activated by a corresponding relay on the Compact DAQ when the enclosure humidity became higher than the set point. The air in the main enclosure was drawn into the dehumidifier and excess moisture condensed on the surface of the cold plate and was returned by gravity to the hydroponics reservoir.
No lighting or temperature controls were integrated into the system as it was designed to operate within a walk-in growth chamber where lighting and temperature were independently regulated. We found that the acrylic used for the enclosure construction filtered out wavelengths shorter than 389 nm but did not absorb visible light wavelengths necessary for plant growth (Additional file 6). Total light intensity, however, was reduced approximately 16%. Thus, it is highly recommended that light intensity be carefully measured prior to each experiment and balanced against controls growing outside the enclosure by adjustments in the elevation of the control plants.
Amino Acid Purification and Derivatization
Amino acid analysis was carried out according to Chen et al. . Tissue samples of approximately 50 mg fresh weight were excised from Arabidopsis seedlings, transferred to microcentrifuge tubes, weighed and frozen in liquid nitrogen before storage at -80°C. Frozen tissues were ground in 1 mL of 10 mM HCl with 10 μL of methionine sulfone or stable isotope labeled amino acids (20-100 μg/mL) as internal standard using a bead grinding mill (5 min at frequency 25 s-1, MixerMill, Qiagen/Retsch Model MM330, Valencia, CA, USA). Samples were then centrifuged at 14,000 × g for 5 min. SCX SPE columns (100 mg resin, Grace, Deerfield, IL) were first wetted with 2 mL of distilled water three times using a vacuum manifold. The supernatants of the samples were transferred to the column and slowly drawn through. After sample loading, the columns were washed two times with 2 mL of a methanol/water mixture (8:1) and then the amino acids were eluted with 0.25 mL of 1:1 (v/v) 8 M NH4OH:methanol. A 50 μL aliquot of the analyte was transferred to a GC-MS vial insert and derivatized directly by mixing with 5 μL of pyridine and 5 μL of methyl chloroformate (MCF). To separate the MCF derivatives from the reactive mixture, 90 μL of chloroform and 90 μL of sodium bicarbonate solution (50 mM) were added and vortexed well. The bottom (chloroform) layer was transferred to a new GC insert containing few crystals of anhydrous sodium sulfate to dry the samples before they were used for GC-MS analysis.
GC-MS Analysis of Amino Acids
All GC-MS analyses were performed using a Hewlett-Packard 5890 (GC)/5970 (MS) (Agilent) using electron impact (EI) ionization at 70 eV. The GC was equipped with a fused silica capillary column (HP-5MS, 30 m × 25 mm ID, 0.25 μm film thickness; Agilent J&W Scientific, Folsom, CA, USA). A 2 μL sample was injected in the splitless mode. The oven temperature was initially held at 70 °C for 3 min. Thereafter the temperature was increased using a gradient of 25°C/min until 280°C, followed by a temperature hold for 5 min. Helium was used as carrier gas and delivered at a constant flow rate at 1 mL/min during the run. The injector temperature was set at 240°C and the interface temperature was at 280°C. The mass spectra of the MCF derivatized amino acids and internal standards were obtained in either the full-scan or, alternatively SIM acquisition mode using a series of predetermined masses that changed based on the known elution time of specific sets of amino acids .
Protein Extraction, Isolation and Trypsin Digestion
13C-labeled Arabidopsis leaves were harvested 0, 24, 48 and 96 h after the growth chamber was opened to ambient air. Plant material was ground in liquid N2 with a mortar and pestle and then total proteins were extracted and washed twice with ice-cold methanol containing a protease inhibitor cocktail (Roche, Indianapolis, IN, USA), then twice with ice-cold acetone. The protein pellets after centrifugation at 14,000 × g for 10 min were air-dried and resuspended in TE buffer containing 1% SDS. The protein concentration was estimated by the Bradford method  using a commercial kit from Bio-Rad (Hercules, CA, USA). Labeled protein samples were spiked with a known amount of the unlabeled protein (at 4: 1 ratio labeled to unlabeled) and separated by SDS-PAGE. Protein bands around 52 kDa corresponding to the molecular weight of the Rubisco large subunit, were excised manually after visualizing with colloidal Coomassie G-250 stain . Excised bands were subjected to trypsin enzymatic digestion  on a ProPrep™ System (Genomic Solutions, Ann Arbor, MI, USA). Briefly, protein bands were subjected to two series of dehydration and hydration steps by the addition, incubation and removal of acetonitrile followed by the addition, incubation and removal of 25 mM NH4HCO3. Gel plugs were then reduced with 10 mM DTT/25 mM NH4HCO3 at 56°C for 30 minutes. The DTT solution was aspirated and a 55 mM iodacetamide/25 mM NH4HCO3 solution was added and the sample incubated for 30 minutes at room temperature. The iodacetamide solution was aspirated, followed by two series of dehydration and hydration steps as above. Protein bands were then subjected to tryptic digestion using 12 ng/μL trypsin (Sigma-Aldrich) in 25 mM NH4HCO3, 5 mM CaCl2 at 37°C for 10 h. The reaction was stopped with the addition of formic acid to a final concentration of 0.1% (v/v). Sample digests were manually aspirated and dispensed into 1.5 mL tubes with subsequent extraction by addition, incubation and removal to the respective tubes of 70% acetonitrile, 0.1% formic acid. All digested extracts were evaporated in vacuo (SC210A SpeedVac® Plus, ThermoSavant, Asheville, NC USA), resuspended in LC-MS/MS loading buffer (98% H2O, 2% acetonitrile and 0.1% formic acid), and run on a QSTAR Pulsar i quadrupole-TOF MS system (Applied Biosystems, Foster City, CA, USA).
Trypsin-digested peptides were separated and analyzed by a LC-MS/MS method described by Griffin et al. . The LC system (LC Packings/Dionex, Sunnyvale, CA, USA) was interfaced with the QSTAR instrument (Applied Biosystems), which was equipped with a Protana (Odense, Denmark) nanoelectrospray source. Peptides (0.5 μg) were eluted with a linear gradient from 0-35% B (0.1% formic acid in a solution of 95:5 acetonitrile:water) over 45 min, followed by 35-80% B over 2 min, and held isocratic at 80% B for 10 min. Solvent A was 0.1% formic acid in 95:5 acetonitrile:water. Product ion spectra were collected in an information-dependent acquisition (IDA) mode, using continuous cycles of one full scan TOF MS from 400-1200 m/z (1 s) plus four product ion scans from 50-2000 m/z (2 s each). Precursor m/z values were selected starting with the most intense ion, using a selected quadrupole resolution of 3 Da. The rolling collision energy feature was used, which determines collision energy based on precursor m/z and charge state. Dynamic exclusion time for precursor ion m/z values was 60 s. MS/MS data were assigned using ProteinPilot (Applied Biosystems) and using the TAIR9 non-redundant Arabidopsis thaliana protein sequence database from TAIR. The list of identified peptides from confidently identified proteins was then saved in text format. Next, the original MS data in WIFF format was converted to mzXML format using the converter, mzWiff, from The Trans-Proteomic Pipeline developed at the Institute for Systems Biology. After the list of identified peptides and mzXML files were compileded they were then input into a program for the modeling algorithm written in R (described below) for the protein turnover calculation.
Determination of 13C Enrichment
We adopted the method described by MacCoss et al.  for the estimation of 13C enrichment of compounds extracted from the 13CO2 labeled plants. The predicted isotope distribution was based on all natural isotope abundances with the exception of selected elements defined by the user as "enriched". For the enriched element(s) the isotope enrichment is varied from 0 to 100%. Each predicted isotope distribution was then correlated against the measured isotope distribution to find a best-predicted isotope distribution that is most representative of the experimentally measured isotope distribution using the Pearson correlation coefficient (r). The relative intensity for each peak in the predicted isotope distribution was calculated as described by Kubinyi .
A program implemented in R to accomplish these calculations is available from the authors upon request (further information is available at website: http://www.proteinturnover.umn.edu/).
Determination of Amino Acid Turnover
The theoretical mass isotopomer distributions of 100% [13
C]-labeled amino acid fragment ions were calculated according to the binomial distribution model as described previously [38
]. The experimental mass isotopomer distributions of the 13
C-labeled and unlabeled amino acid fragment ions were obtained by GC-MS and were used to calculate the relative isotope abundance (Rt
) at each time, t
, as a ratio of total net experimental fractional abundance of the mass isotopomers of the labeled ions and total net theoretical fractional abundance of mass isotopomers of 100% [13
C]- labeled ions as shown in equation (1
In equation (1), i is the number of carbon atoms in any amino acid derived ion used for turnover calculation. In the mass spectrum, i provides the number of possible isotopic peaks appearing in roughly integer increments above the monoisotopic peak. This value was used as a normalizing parameter as its magnitude reflects the fractional contribution of 13C to 12C in each isotopic peak. To calculate R
, we first calculated the total net isotopic abundance of observed labeled amino acids by summing the normalized differences in fractional abundance for each peak of the distribution between the experimental samples (EMi) and the unlabeled (natural abundance) standards (SMi). Similarly, the total net isotopic abundance of 100% [13C]-labeled amino acids can be calculated by summing the normalized differences in fractional abundance for each peak of the distribution between the 100% [13C]-labeled amino acid (TMi) and the unlabeled standards (SMi). The isotopic peak distribution for unlabeled standards can be obtained from the experimental data or it can be generated from the combination of theoretical binomial distributions of each naturally occurring stable isotope in any given elemental composition; TMi is the theoretical fractional abundance of isotopic peaks occurring in the 100% labeled amino acid ions.
The value of R
changes over time as the amino acids are first prelabeled with 13
C and are then repopulated by 12
C. This shift in distribution occurs through normal intracellular amino acid metabolism following transfer from 13
. Turnover rates can be estimated by nonlinear curve fitting of the plot of R
measured over multiple time points and fitted to an equation for exponential decay either without (Eq. 2) or with a plateau (Eq. 3) parameter.
is relative isotopic abundance from Eq. 1
(time) is zero, meaning at the onset of dilution experiment. Plateau
is relative isotopic abundance R
at infinite times, k
is the rate constant. Once k
is computed, then the half-life of an amino acid can be computed as Eq. 4
The half-lives of amino acids shown in Figure 2 were calculated using Eq. 2 assuming no plateau. The algorithm has been implemented in a Windows Excel 2003 format for ease of use of this MIDA calculation.
Determination of Protein Turnover
An algorithm implemented in R was developed by us to extract isotopic distribution information from raw MS data for multiple peptides identified by tandem MS. Then, the isotopic distributions were modeled by maximum likelyhood estimation using β-binomial distributions for: 1) spiked natural abundance, 2) newly synthesized peptide and 3) old peptide distributions. The workflow of the algorithm is described briefly as below. First, the algorythm was provided with a list of identified peptides (peptide amino acid sequence, and detected m/z and retention information) and the raw MS data in mzXML format. Then the number of carbon atoms were calculated for each peptide using the amino acid composition. Each carbon isotopic channel was assigned an m/z value calculated from the observed monoisotopic m/z value plus the 13C mass defect. Six additional channels were included to take account of the natural abundance of other isotopes (15N, 2H, 18O etc.). Extracted ion chromatograms were generated for each peptide at each isotope channel within a 5 min window centered on the retention time when the identifying MS/MS spectrum was triggered. This set of isotope abundances in each retention window for each peptide served as the data set for all analyses. Next, a linear regression analysis was performed for each isotope channel against the monoisotope channel for each peptide to reduce chemical noise and overlapping uncorelated peptide signals in the extracted spectra. Maximum likelihood estimation was performed to calculate the fractional isotopic abundance of the newly synthesied peptide distribution and the distribution abundance ratio of old to newly synthesized peptide distributions. Finally, half-life of each peptide and protein was calculated from changes in the distribution abundance ratios using non-linear-regression. Development and evaluation of algorithm is ongoing and a β-version of the web-based calculator and standalone software in R is available (further information at website: http://www.proteinturnover.umn.edu/).