Effects of Al3+ and La3+ Trivalent Metal Ions on Tomato Fruit Proteomes

The tomato (Solanum lycopersicum) ripening process from mature green (MG) to turning and then to red stages is accompanied by the occurrences of physiological and biochemical reactions, which ultimately result in the formation of the flavor, color and texture of ripe fruits. The two trivalent metal ions Al3+ and La3+ are known to induce different levels of phytotoxicity in suppressing root growth. This paper aims to understand the impacts of these two metal ions on tomato fruit proteomes. Tomato ‘Micro-Tom’ plants were grown in a hydroponic culture system supplemented with 50 μM aluminum sulfate (Al2 (SO4)3.18H2O) for Al3+ or La2(SO4)3 for La3+. Quantitative proteomics analysis, using isobaric tags for relative and absolute quantitation, were performed for fruits at MG, turning and red stages. Results show that in MG tomatoes, proteins involved in protein biosynthesis, photosynthesis and primary carbohydrate metabolisms were at a significantly lower level in Al-treated compared to La-treated plants. For the turning and red tomatoes, only a few proteins of significant differences between the two metal treatments were identified. Results from this study indicate that compared to La3+, Al3+ had a greater influence on the basic biological activities in green tomatoes, but such an impact became indistinguishable as tomatoes matured into the late ripening stages.


Introduction
Tomato (Solanum lycopersicum) is one of the most important fleshy vegetable crops in the world. According to the United States Department of Agriculture (USDA), based on the external color changes, the ripeness stages for red-fleshed tomatoes are divided into (1) mature green (MG), (2) breaker, (3) turning, (4) pink, (5) light red and (6) red [1]. Tomato fruits are predominantly composed of parenchyma cells enclosed by an unlignified layer of cellulose microfibrils suspended in a matrix of glycoproteins, water, pectic and hemicellulose polysaccharides [2]. Tomato ripening involves a variety of biochemical reactions in the flesh tissues, for example: the accumulation of lycopene leads to tomato turning red; an elevated content of citrate and malate makes the pulp acidic; the biosynthesis of a plentiful assortment of secondary metabolites gives the pleasant flavors and aroma; and the alteration of cell wall structures results in changes in fruit texture [3,4]. Underlying the fruit ripening process is a coordinated expression of sets of genes regulated by ripening-related transcription factors (TFs) and/or the two-leaf stage, plants were placed in net pots with the support of hydroton clay balls and transplanted into hydroponic tanks filled with 30 L of a modified Magnavaca's solution (pH, 4.5) [27]. There were eight tanks each growing 18 plants.
Previously, aluminum sulfate octadecahydrate (Al2(SO4)3·18H2O) and lanthanum ‫)|||(‬ sulfate hydrate (La2(SO4)3) were used in plant treatment experiments [20,28]. In this study of tomatoes, Al was added to a final concentration of 50 μM Al2(SO4)3·18H2O (Fisher Scientific, Pittsburgh, PA, USA) in the Al-treated group and La of 50 μM La2(SO4)3 (Strem chemicals, MA, USA) in the La-treated group. These treatments were initiated when two or more tomatoes on the first fruit cluster turned red. The pH of the solution was monitored daily using pH indicator strips (Fisher Scientific). The solution was refreshed every three days. After nine days of treatments, tomato fruits in the MG, turning and red stages ( Figure 1) were harvested. Two fruits were taken from each plant; samples from one tank were pooled to form one biological replicate. Four biological replicates each for Al-treated and La-treated plants were harvested. Samples were frozen in liquid nitrogen immediately after collection and stored at −80 °C.

Protein Extraction
To extract protein, frozen tomatoes were ground into a fine powder under N2. Total protein was extracted using the Plant Total Protein Extraction Kit (Sigma, St. Louis, MO, USA), following the manufacturer's instruction with modifications. Briefly, 5 g aliquots of the tissue powder were transferred into 50-mL centrifuge tubes to which 50 mL of the extraction solution (10% trichloroacetic acid (TCA)) in acetone was added. Samples were vortexed, incubated overnight at −20 °C and centrifuged at 10,000 rpm for ten minutes at 4 °C. Protein pellets were washed 3 times with pre-chilled acetone to remove residual TCA and dried briefly. Proteins were solubilized in a solution of 6 M urea/500 mM triethylammonium bicarbonate (TEAB) and then diluted (1:3, v/v) with TEAB buffer to reduce urea to a final concentration of 2 M. The protein concentration was determined by Bradford's method using bovine serum albumin (BSA) as the standard (Bio-Rad, Hercules, CA, USA).

iTRAQ Labeling Procedure
An aliquot of 100 μg protein per sample was used for iTRAQ labeling following the instruction in the 8-plex iTRAQ ® labeling kit (AB SCIEX, Foster City, CA, USA). Sodium dodecyl sulfate (SDS) was added to the samples to a final concentration of 0.1%. To reduce the disulfide bonds of proteins, tris (2-carboxyethyl) phosphine (TCEP) was added to a final concentration of 5 mM, and the samples were incubated at 32 °C for 1 h. Next, the cysteines were blocked by adding methyl methane thiosulfonate (MMTS) to a final concentration of 10 mM, and the samples were incubated at room temperature for 30 min. For protein digestion, 2 μg of sequence-grade modified trypsin enzyme ) sulfate hydrate (La 2 (SO 4 ) 3 ) were used in plant treatment experiments [20,28]. In this study of tomatoes, Al was added to a final concentration of 50 µM Al 2 (SO 4 ) 3 ·18H 2 O (Fisher Scientific, Pittsburgh, PA, USA) in the Al-treated group and La of 50 µM La 2 (SO 4 ) 3 (Strem chemicals, MA, USA) in the La-treated group. These treatments were initiated when two or more tomatoes on the first fruit cluster turned red. The pH of the solution was monitored daily using pH indicator strips (Fisher Scientific). The solution was refreshed every three days. After nine days of treatments, tomato fruits in the MG, turning and red stages ( Figure 1) were harvested. Two fruits were taken from each plant; samples from one tank were pooled to form one biological replicate. Four biological replicates each for Al-treated and La-treated plants were harvested. Samples were frozen in liquid nitrogen immediately after collection and stored at −80 • C.  [20,28]. In this study of tomatoes, Al was added to a final concentration of 50 μM Al2(SO4)3·18H2O (Fisher Scientific, Pittsburgh, PA, USA) in the Al-treated group and La of 50 μM La2(SO4)3 (Strem chemicals, MA, USA) in the La-treated group. These treatments were initiated when two or more tomatoes on the first fruit cluster turned red. The pH of the solution was monitored daily using pH indicator strips (Fisher Scientific). The solution was refreshed every three days. After nine days of treatments, tomato fruits in the MG, turning and red stages ( Figure 1) were harvested. Two fruits were taken from each plant; samples from one tank were pooled to form one biological replicate. Four biological replicates each for Al-treated and La-treated plants were harvested. Samples were frozen in liquid nitrogen immediately after collection and stored at −80 °C.

Protein Extraction
To extract protein, frozen tomatoes were ground into a fine powder under N2. Total protein was extracted using the Plant Total Protein Extraction Kit (Sigma, St. Louis, MO, USA), following the manufacturer's instruction with modifications. Briefly, 5 g aliquots of the tissue powder were transferred into 50-mL centrifuge tubes to which 50 mL of the extraction solution (10% trichloroacetic acid (TCA)) in acetone was added. Samples were vortexed, incubated overnight at −20 °C and centrifuged at 10,000 rpm for ten minutes at 4 °C. Protein pellets were washed 3 times with pre-chilled acetone to remove residual TCA and dried briefly. Proteins were solubilized in a solution of 6 M urea/500 mM triethylammonium bicarbonate (TEAB) and then diluted (1:3, v/v) with TEAB buffer to reduce urea to a final concentration of 2 M. The protein concentration was determined by Bradford's method using bovine serum albumin (BSA) as the standard (Bio-Rad, Hercules, CA, USA).

iTRAQ Labeling Procedure
An aliquot of 100 µ g protein per sample was used for iTRAQ labeling following the instruction in the 8-plex iTRAQ ® labeling kit (AB SCIEX, Foster City, CA, USA). Sodium dodecyl sulfate (SDS) was added to the samples to a final concentration of 0.1%. To reduce the disulfide bonds of proteins, tris (2-carboxyethyl) phosphine (TCEP) was added to a final concentration of 5 mM, and the samples were incubated at 32 °C for 1 h. Next, the cysteines were blocked by adding methyl methane thiosulfonate (MMTS) to a final concentration of 10 mM, and the samples were incubated at room temperature for 30 min. For protein digestion, 2 µ g of sequence-grade modified trypsin enzyme

Protein Extraction
To extract protein, frozen tomatoes were ground into a fine powder under N 2 . Total protein was extracted using the Plant Total Protein Extraction Kit (Sigma, St. Louis, MO, USA), following the manufacturer's instruction with modifications. Briefly, 5 g aliquots of the tissue powder were transferred into 50-mL centrifuge tubes to which 50 mL of the extraction solution (10% trichloroacetic acid (TCA)) in acetone was added. Samples were vortexed, incubated overnight at −20 • C and centrifuged at 10,000 rpm for ten minutes at 4 • C. Protein pellets were washed 3 times with pre-chilled acetone to remove residual TCA and dried briefly. Proteins were solubilized in a solution of 6 M urea/500 mM triethylammonium bicarbonate (TEAB) and then diluted (1:3, v/v) with TEAB buffer to reduce urea to a final concentration of 2 M. The protein concentration was determined by Bradford's method using bovine serum albumin (BSA) as the standard (Bio-Rad, Hercules, CA, USA).

iTRAQ Labeling Procedure
An aliquot of 100 µg protein per sample was used for iTRAQ labeling following the instruction in the 8-plex iTRAQ ® labeling kit (AB SCIEX, Foster City, CA, USA). Sodium dodecyl sulfate (SDS) was added to the samples to a final concentration of 0.1%. To reduce the disulfide bonds of proteins, tris (2-carboxyethyl) phosphine (TCEP) was added to a final concentration of 5 mM, and the samples were incubated at 32 • C for 1 h. Next, the cysteines were blocked by adding methyl methane thiosulfonate (MMTS) to a final concentration of 10 mM, and the samples were incubated at room temperature for 30 min. For protein digestion, 2 µg of sequence-grade modified trypsin enzyme (Promega, Madison, WI, USA) were added to each sample, and samples were incubated at 30 • C for 18 h. After digestion, the volume of each sample was reduced to 20 µL, by vacuum centrifugation. The four Al-treated samples were labeled with iTRAQ Reagents 113, 114, 117 and 118, while the four La-treated control samples were labeled with iTRAQ Reagents 115, 116, 119 and 121. The labeling reaction was allowed to proceed for 2 h at room temperature, protected from light.

Cleaning of the Peptides
After the labeling reaction, the 8 samples were pooled into one tube, mixed and evaporated by vacuum centrifugation. Excess TCEP, SDS and iTRAQ reagents were removed from the sample using an iCAT cation exchange cartridge (AB Sciex, Framingham, MA, USA). The dried sample was resuspended in a cation exchange load buffer (10 mM potassium phosphate in 25% acetonitrile, pH 3.0). The pH of the sample was adjusted to pH 2.5 using trifluoroacetic acid (TFA). The iCAT cartridge was cleaned with 1 mL of the cation exchange clean buffer (10 mM potassium phosphate in 25% acetonitrile/1M potassium chloride, pH 3.0). The cartridge was conditioned with 2 mL of the cation exchange load buffer. The sample mixture was slowly injected onto the iCAT cartridge. To wash the peptides, 2 mL of the cation exchange load buffer were injected through the cartridge. The labeled peptides were eluted from the cartridge and collected by injecting 600 uL of the cation exchange elution buffer (10 mM potassium phosphate in 25% acetonitrile/350 mM potassium chloride, pH 3.0) through the cartridge. The elution was dried to completion by vacuum centrifugation.
The sample was resuspended in 0.1% TFA for a solid phase extraction (SPE) step to remove salts. The pH of the sample mixture solution was brought to 3 with TFA. A SepPak C18, 1 mL, vacuum cartridge (Waters, Milford, MA, USA) was cleaned using acetonitrile. Then, the cartridge was conditioned with 5 mL of 0.1% TFA. The sample mixture was slowly loaded through the cartridge using a needle valve to control the vacuum. The cartridge was washed with 5 mL of 2% acetonitrile/0.1% TFA. The peptides were then eluted with 600 uL of 50% acetonitrile/0.1% TFA. The eluate was dried to completion and submitted for analysis of the iTRAQ-labeled peptides.

High pH First Dimension UPLC Separation
A Waters Acquity UPLC (Milford, MA, USA) coupled with a Dionex Probot (Sunnyvale, CA, USA) robotic fraction collector was used to separate the peptides. A Waters Acquity UPLC BEH C18 1.7 µm, 2.1 mm × 100 mm column was used. Mobile Phase A consisted of 20 mM ammonium formate dissolved in water (pH 10), and Mobile Phase B consisted of 90:10 acetonitrile: water (v/v) containing 20 mM ammonium formate (pH 10). For the separation, a segmented gradient was used employing a flow rate of 300 µL/min. The initial conditions were 0% B; Segment 1 was 0% Proteomes 2017, 5,7 (Promega, Madison, WI, USA) were added to each sample 18 h. After digestion, the volume of each sample was red The four Al-treated samples were labeled with iTRAQ Rea La-treated control samples were labeled with iTRAQ Rea reaction was allowed to proceed for 2 h at room temperatu 2.2.3. Cleaning of the Peptides After the labeling reaction, the 8 samples were pooled vacuum centrifugation. Excess TCEP, SDS and iTRAQ r using an iCAT cation exchange cartridge (AB Sciex, Frami resuspended in a cation exchange load buffer (10 mM p pH 3.0). The pH of the sample was adjusted to pH 2.5 u cartridge was cleaned with 1 mL of the cation exchange cle 25% acetonitrile/1M potassium chloride, pH 3.0). The car cation exchange load buffer. The sample mixture was slo wash the peptides, 2 mL of the cation exchange load buffe labeled peptides were eluted from the cartridge and co exchange elution buffer (10 mM potassium phosphate chloride, pH 3.0) through the cartridge. The el vacuum centrifugation.
The sample was resuspended in 0.1% TFA for a sol salts. The pH of the sample mixture solution was broug vacuum cartridge (Waters, Milford, MA, USA) was clean was conditioned with 5 mL of 0.1% TFA. The sample cartridge using a needle valve to control the vacuum. Th acetonitrile/0.1% TFA. The peptides were then eluted with eluate was dried to completion and submitted for analysis

High pH First Dimension UPLC Separation
A Waters Acquity UPLC (Milford, MA, USA) couple USA) robotic fraction collector was used to separate the pe 1.7 µm, 2.1 mm × 100 mm column was used. Mobile P formate dissolved in water (pH 10), and Mobile Phase B containing 20 mM ammonium formate (pH 10). For the se employing a flow rate of 300 µL/min. The initial condition over 0.5 min; Segment 2 was 5%  35% B (linear) over 8 m 1 min. The column was then returned to initial conditio hundred micrograms of total labeled protein digest were in 96-well plate. The fraction collector was triggered via co delayed for 30 s before commencing collection. A total fraction being collected every 14 s. Fractions were then Fractions, 1, 2, 3 and 4 were mixed to produce second dim mixed to make Sample 2; Fractions 6 and 26 were mixed to an additional 20 fractionally-concatenated second dim produced by pooling Fractions 45-48.

Low pH Second Dimension RP Separation
Dried samples were reconstituted with 24 µL of 3 nanoACQUITY system (Waters), equipped with a Symme column and a UPLC BEH C18 1.7 µm, 15 cm × 75 µm perform the nano-LC separation of tryptic peptides. The s 5% B over 0.5 min; Segment 2 was 5%

Cleaning of the Peptides
After the labeling reaction, the 8 samples were pooled into one tube, mixed and evaporated by vacuum centrifugation. Excess TCEP, SDS and iTRAQ reagents were removed from the sample using an iCAT cation exchange cartridge (AB Sciex, Framingham, MA, USA). The dried sample was resuspended in a cation exchange load buffer (10 mM potassium phosphate in 25% acetonitrile, pH 3.0). The pH of the sample was adjusted to pH 2.5 using trifluoroacetic acid (TFA). The iCAT cartridge was cleaned with 1 mL of the cation exchange clean buffer (10 mM potassium phosphate in 25% acetonitrile/1M potassium chloride, pH 3.0). The cartridge was conditioned with 2 mL of the cation exchange load buffer. The sample mixture was slowly injected onto the iCAT cartridge. To wash the peptides, 2 mL of the cation exchange load buffer were injected through the cartridge. The labeled peptides were eluted from the cartridge and collected by injecting 600 uL of the cation exchange elution buffer (10 mM potassium phosphate in 25% acetonitrile/350 mM potassium chloride, pH 3.0) through the cartridge. The elution was dried to completion by vacuum centrifugation.
The sample was resuspended in 0.1% TFA for a solid phase extraction (SPE) step to remove salts. The pH of the sample mixture solution was brought to 3 with TFA. A SepPak C18, 1 mL, vacuum cartridge (Waters, Milford, MA, USA) was cleaned using acetonitrile. Then, the cartridge was conditioned with 5 mL of 0.1% TFA. The sample mixture was slowly loaded through the cartridge using a needle valve to control the vacuum. The cartridge was washed with 5 mL of 2% acetonitrile/0.1% TFA. The peptides were then eluted with 600 uL of 50% acetonitrile/0.1% TFA. The eluate was dried to completion and submitted for analysis of the iTRAQ-labeled peptides.

High pH First Dimension UPLC Separation
A Waters Acquity UPLC (Milford, MA, USA) coupled with a Dionex Probot (Sunnyvale, CA, USA) robotic fraction collector was used to separate the peptides. A Waters Acquity UPLC BEH C18 1.7 µm, 2.1 mm × 100 mm column was used. Mobile Phase A consisted of 20 mM ammonium formate dissolved in water (pH 10), and Mobile Phase B consisted of 90:10 acetonitrile: water (v/v) containing 20 mM ammonium formate (pH 10). For the separation, a segmented gradient was used employing a flow rate of 300 µL/min. The initial conditions were 0% B; Segment 1 was 0%  5% B over 0.5 min; Segment 2 was 5%  35% B (linear) over 8 min; and Segment 3 was 35%  95% B over 1 min. The column was then returned to initial conditions and re-equilibrated over 4 min. One hundred micrograms of total labeled protein digest were injected and separated into 48 fractions in a 96-well plate. The fraction collector was triggered via contact closure after the injection and was delayed for 30 s before commencing collection. A total of 48 fractions were collected, with one fraction being collected every 14 s. Fractions were then concatenated as follows: 1st Dimension Fractions, 1, 2, 3 and 4 were mixed to produce second dimension Sample 1; Fractions 5 and 25 were mixed to make Sample 2; Fractions 6 and 26 were mixed to make Sample 3; and so forth, to produce an additional 20 fractionally-concatenated second dimension samples. The 22nd sample was produced by pooling Fractions 45-48.

Low pH Second Dimension RP Separation
Dried samples were reconstituted with 24 µL of 3% acetonitrile mixed with 0.1% TFA. A nanoACQUITY system (Waters), equipped with a Symmetry C18 5 µm, 20 mm × 180 µm trapping column and a UPLC BEH C18 1.7 µm, 15 cm × 75 µm analytical column (Waters), was used to perform the nano-LC separation of tryptic peptides. The samples, 3-µL partial loop injections, were 35% B (linear) over 8 min; and Segment 3 was 35% Proteomes 2017, 5,7 (Promega, Madison, WI, USA) were added to each sample, and 18 h. After digestion, the volume of each sample was reduced t The four Al-treated samples were labeled with iTRAQ Reagents La-treated control samples were labeled with iTRAQ Reagents reaction was allowed to proceed for 2 h at room temperature, pr

Cleaning of the Peptides
After the labeling reaction, the 8 samples were pooled into vacuum centrifugation. Excess TCEP, SDS and iTRAQ reagen using an iCAT cation exchange cartridge (AB Sciex, Framingham resuspended in a cation exchange load buffer (10 mM potass pH 3.0). The pH of the sample was adjusted to pH 2.5 using tr cartridge was cleaned with 1 mL of the cation exchange clean bu 25% acetonitrile/1M potassium chloride, pH 3.0). The cartridge cation exchange load buffer. The sample mixture was slowly in wash the peptides, 2 mL of the cation exchange load buffer were labeled peptides were eluted from the cartridge and collected exchange elution buffer (10 mM potassium phosphate in 25 chloride, pH 3.0) through the cartridge. The elution vacuum centrifugation.
The sample was resuspended in 0.1% TFA for a solid pha salts. The pH of the sample mixture solution was brought to vacuum cartridge (Waters, Milford, MA, USA) was cleaned us was conditioned with 5 mL of 0.1% TFA. The sample mixtu cartridge using a needle valve to control the vacuum. The cartr acetonitrile/0.1% TFA. The peptides were then eluted with 600 u eluate was dried to completion and submitted for analysis of the

High pH First Dimension UPLC Separation
A Waters Acquity UPLC (Milford, MA, USA) coupled wit USA) robotic fraction collector was used to separate the peptides 1.7 µm, 2.1 mm × 100 mm column was used. Mobile Phase formate dissolved in water (pH 10), and Mobile Phase B consis containing 20 mM ammonium formate (pH 10). For the separati employing a flow rate of 300 µL/min. The initial conditions wer over 0.5 min; Segment 2 was 5%  35% B (linear) over 8 min; an 1 min. The column was then returned to initial conditions an hundred micrograms of total labeled protein digest were injected 96-well plate. The fraction collector was triggered via contact delayed for 30 s before commencing collection. A total of 48 fraction being collected every 14 s. Fractions were then conca Fractions, 1, 2, 3 and 4 were mixed to produce second dimensio mixed to make Sample 2; Fractions 6 and 26 were mixed to mak an additional 20 fractionally-concatenated second dimension produced by pooling Fractions 45-48.

Low pH Second Dimension RP Separation
Dried samples were reconstituted with 24 µL of 3% ace nanoACQUITY system (Waters), equipped with a Symmetry C column and a UPLC BEH C18 1.7 µm, 15 cm × 75 µm analyt perform the nano-LC separation of tryptic peptides. The sample 95% B over 1 min. The column was then returned to initial conditions and re-equilibrated over 4 min. One hundred micrograms of total labeled protein digest were injected and separated into 48 fractions in a 96-well plate. The fraction collector was triggered via contact closure after the injection and was delayed for 30 s before commencing collection. A total of 48 fractions were collected, with one fraction being collected every 14 s. Fractions were then concatenated as follows: 1st Dimension Fractions, 1, 2, 3 and 4 were mixed to produce second dimension Sample 1; Fractions 5 and 25 were mixed to make Sample 2; Fractions 6 and 26 were mixed to make Sample 3; and so forth, to produce an additional 20 fractionally-concatenated second dimension samples. The 22nd sample was produced by pooling Fractions 45-48.

Low pH Second Dimension RP Separation
Dried samples were reconstituted with 24 µL of 3% acetonitrile mixed with 0.1% TFA. A nanoACQUITY system (Waters), equipped with a Symmetry C18 5 µm, 20 mm × 180 µm trapping column and a UPLC BEH C18 1.7 µm, 15 cm × 75 µm analytical column (Waters), was used to perform the nano-LC separation of tryptic peptides. The samples, 3-µL partial loop injections, were transferred to the trapping column via a 0.1% solution of formic acid in water at a flow rate of 7 µL/min for three minutes. Mobile Phase A contained water with 0.1% formic acid, and B contained acetonitrile with 0.1% formic acid. The trapping column was subjected to a reverse flush to the analytical column after the desalting and concentration and was separated at a flow rate of 300 nL/min. The initial conditions were run using 2% B. A 3-segment gradient was used to elute the peptides: Segment 1, linear 2% was conditioned with 5 mL of 0.1% TFA. The sam cartridge using a needle valve to control the vacuum acetonitrile/0.1% TFA. The peptides were then eluted eluate was dried to completion and submitted for an

High pH First Dimension UPLC Separation
A Waters Acquity UPLC (Milford, MA, USA) c USA) robotic fraction collector was used to separate 1.7 µm, 2.1 mm × 100 mm column was used. Mo formate dissolved in water (pH 10), and Mobile Pha containing 20 mM ammonium formate (pH 10). For employing a flow rate of 300 µL/min. The initial con over 0.5 min; Segment 2 was 5%  35% B (linear) ove 1 min. The column was then returned to initial co hundred micrograms of total labeled protein digest w 96-well plate. The fraction collector was triggered delayed for 30 s before commencing collection. A fraction being collected every 14 s. Fractions were Fractions, 1, 2, 3 and 4 were mixed to produce secon mixed to make Sample 2; Fractions 6 and 26 were mi an additional 20 fractionally-concatenated second produced by pooling Fractions 45-48.

Low pH Second Dimension RP Separation
Dried samples were reconstituted with 24 µL nanoACQUITY system (Waters), equipped with a S column and a UPLC BEH C18 1.7 µm, 15 cm × 75 perform the nano-LC separation of tryptic peptides. 5% B over 2 min; Segment 2, linear 5% salts. The pH of the sample mixture solution was brought to 3 with TFA. A SepPak C18, 1 mL, vacuum cartridge (Waters, Milford, MA, USA) was cleaned using acetonitrile. Then, the cartridge was conditioned with 5 mL of 0.1% TFA. The sample mixture was slowly loaded through the cartridge using a needle valve to control the vacuum. The cartridge was washed with 5 mL of 2% acetonitrile/0.1% TFA. The peptides were then eluted with 600 uL of 50% acetonitrile/0.1% TFA. The eluate was dried to completion and submitted for analysis of the iTRAQ-labeled peptides.

High pH First Dimension UPLC Separation
A Waters Acquity UPLC (Milford, MA, USA) coupled with a Dionex Probot (Sunnyvale, CA, USA) robotic fraction collector was used to separate the peptides. A Waters Acquity UPLC BEH C18 1.7 µm, 2.1 mm × 100 mm column was used. Mobile Phase A consisted of 20 mM ammonium formate dissolved in water (pH 10), and Mobile Phase B consisted of 90:10 acetonitrile: water (v/v) containing 20 mM ammonium formate (pH 10). For the separation, a segmented gradient was used employing a flow rate of 300 µL/min. The initial conditions were 0% B; Segment 1 was 0%  5% B over 0.5 min; Segment 2 was 5%  35% B (linear) over 8 min; and Segment 3 was 35%  95% B over 1 min. The column was then returned to initial conditions and re-equilibrated over 4 min. One hundred micrograms of total labeled protein digest were injected and separated into 48 fractions in a 96-well plate. The fraction collector was triggered via contact closure after the injection and was delayed for 30 s before commencing collection. A total of 48 fractions were collected, with one fraction being collected every 14 s. Fractions were then concatenated as follows: 1st Dimension Fractions, 1, 2, 3 and 4 were mixed to produce second dimension Sample 1; Fractions 5 and 25 were mixed to make Sample 2; Fractions 6 and 26 were mixed to make Sample 3; and so forth, to produce an additional 20 fractionally-concatenated second dimension samples. The 22nd sample was produced by pooling Fractions 45-48.

Low pH Second Dimension RP Separation
Dried samples were reconstituted with 24 µL of 3% acetonitrile mixed with 0.1% TFA. A nanoACQUITY system (Waters), equipped with a Symmetry C18 5 µm, 20 mm × 180 µm trapping column and a UPLC BEH C18 1.7 µm, 15 cm × 75 µm analytical column (Waters), was used to perform the nano-LC separation of tryptic peptides. The samples, 3-µL partial loop injections, were 10% B over 1 min; Segment 3, linear 10% salts. The pH of the sample mixture solution was brought to vacuum cartridge (Waters, Milford, MA, USA) was cleaned usi was conditioned with 5 mL of 0.1% TFA. The sample mixtur cartridge using a needle valve to control the vacuum. The cartr acetonitrile/0.1% TFA. The peptides were then eluted with 600 uL eluate was dried to completion and submitted for analysis of the

High pH First Dimension UPLC Separation
A Waters Acquity UPLC (Milford, MA, USA) coupled with USA) robotic fraction collector was used to separate the peptides 1.7 µm, 2.1 mm × 100 mm column was used. Mobile Phase A formate dissolved in water (pH 10), and Mobile Phase B consist containing 20 mM ammonium formate (pH 10). For the separati employing a flow rate of 300 µL/min. The initial conditions wer over 0.5 min; Segment 2 was 5%  35% B (linear) over 8 min; and 1 min. The column was then returned to initial conditions an hundred micrograms of total labeled protein digest were injected 96-well plate. The fraction collector was triggered via contact delayed for 30 s before commencing collection. A total of 48 fraction being collected every 14 s. Fractions were then conca Fractions, 1, 2, 3 and 4 were mixed to produce second dimension mixed to make Sample 2; Fractions 6 and 26 were mixed to make an additional 20 fractionally-concatenated second dimension produced by pooling Fractions 45-48.

Low pH Second Dimension RP Separation
Dried samples were reconstituted with 24 µL of 3% ace nanoACQUITY system (Waters), equipped with a Symmetry C1 column and a UPLC BEH C18 1.7 µm, 15 cm × 75 µm analyt perform the nano-LC separation of tryptic peptides. The sample 35% B over 2 h. The column was then washed for 5 min with 95% of Mobile Phase B before returning to initial conditions and a 20-min equilibration period. The column temperature was maintained at 35 • C and re-equilibrated at initial conditions for 20 min prior to the next injection. One hundred femtomoles per microliter [Glu1]-fibrinopeptide B in 25% acetonitrile with 0.1% formic acid were used as the lock mass compound and were delivered via the auxiliary pump of the LC system at a flow rate of 300 nL/min to the reference sprayer of the Nano Lock Spray source of the mass spectrometer. From the analytical column, the eluent was delivered to the analytical sprayer of the same source through a PicoTip emitter (New Objective, Woburn, MA, USA) with a 10-µm tip diameter.

Mass Spectrometric Analysis
Tryptic peptides were subjected to mass spectrometric analysis, using a Synapt HDMS mass spectrometer (Waters). The Synapt was operated in Q-TOF V mode with a typical resolution of at least 10,000 full width at half maximum (FWHM). The TOF analyzer of the mass spectrometer was externally calibrated by fragmenting the doubly-protonated mono isotopic ion of [Glu] 1 -fibrinopeptide B and delivering it via the lock mass reference sprayer. Calibration was performed over the m/z range from 50-2000, and the collected data were then post-acquisition lock mass corrected. The reference sprayer was sampled for one second after every 100 s. Accurate mass LC-MS/MS data-dependent acquisition (DDA) data were obtained as follows: MS survey scans of one-second duration with an inter-scan delay of 0.02 s were acquired for the m/z range from 300-1500. The intensity of a single ion rising above a set 60 counts per second threshold triggered MS/MS fragmentation for the ion provided the ion met the charge state criteria. MS/MS data were obtained for up to four ions of charge 2 + , 3 + , or 4 + detected in the survey scans using charge state selection. MS/MS spectra were acquired for the m/z ranges from 50-1400 at a scan rate of 1 s with an inter-scan delay of 0.02 s. To improve the quality of MS/MS spectra, charge state-dependent collision energy ramps were optimized and employed. A real-time dynamic exclusion window of 35 s was applied to each precursor selected for fragmentation. The instrument returned to MS mode when the total ion current for an MS/MS acquisition exceeded 30,000 counts per second (cps) or after 2.5 s had elapsed.

Data Processing
All raw data files were processed using ProteinLynx Global Server 2.4 into a .pkl peak list format compatible with Mascot (Matrix Science, Boston, MA,). The software was adjusted for mass accuracy by performing lock mass correction using the doubly-protonated, monoisotopic peak for [Glu] 1 -fibrinopeptide B at m/z 785.8426, utilizing 2 lock spray scans with a tolerance of 0.25 Dalton (Da). For noise reduction, both precursor and fragment functions were set to perform "adaptive" background subtraction. For deisotoping and centroiding, both functions instructed the software to perform deisotoping at the "medium" setting with automatic thresholds.

Database Searching/iTRAQ Quantitation
The 22 packing list file (pkl) files from the 22 injections of each sample were combined using Mascot Daemon v. 2.3.2 to make a single query against an iTAG 2.3 tomato protein database (downloaded 16 September 2011; solgenomics.net/tomato/). The enzymatic cleavage specificity was set to trypsin with 1 missed cleavage allowed. Precursor tolerance was set at 25 ppm, while Proteomes 2017, 5, 7 6 of 14 fragment tolerance was set to 0.1 Da. The instrument selected was electrospray ionization quadrupole time-of-flight mass spectrometry (ESI-QUAD-TOF). Individual iTRAQ quantitation methods were set up and used depending on the control/treatment/pool reporter labels used for each sample. Each quantitation method used "average" for the protein ratio type and required a minimum of 2 peptides for protein quantitation. For normalization, "average" was used. N-terminal and lysine modification with iTRAQ were set as fixed modifications, and tyrosine labeling was set as variable modification. Upon completion of the searches, results were exported after setting the ion score filter to 0.1, thereby exporting only results with an expectation value below 0.1 and requiring bold red in order to limit results to the highest scoring match to a particular query listed under the highest scoring protein containing that match. Functional pathways of the listed proteins were analyzed using the MapMan tools (Version 3.6.0RC1) [29,30]. Additionally, a literature search was also used to identify the functions of particular proteins relevant to tomato fruit ripening and Al and La stresses.

Statistical Analysis
Statistical analysis was done by using SAS (Version 9.3; SAS Institute, Cary, NC, USA). Raw expression ratios were log 2 transformed and then fit to a normal distribution [25,31]. The log 2 fold values were subjected to t-test using the general linear model (GLM) procedure followed by the false discovery rate corrections using SAS. Significant proteins were listed for those passing the two steps of statistical tests (p ≤ 0.05) and with the fold value either lower than 0.65 (<0.65) or higher than 1.51 (>1.51-fold), from Al-treated to La-treated groups.

The Al-Treatment-Induced Proteome Composition in Tomatoes
Using the criteria of two unique peptides per protein, 809 proteins were quantified for MG tomatoes and 600 proteins for both turning and red tomatoes each. When these proteins were analyzed for functional pathways in MapMan, it was found that protein expression in MG tomatoes differs from those from turning and red tomatoes, which display more similar protein expression patterns, as can be seen from Figure 2. time-of-flight mass spectrometry (ESI-QUAD-TOF). Individual iTRAQ quantitation methods were set up and used depending on the control/treatment/pool reporter labels used for each sample. Each quantitation method used "average" for the protein ratio type and required a minimum of 2 peptides for protein quantitation. For normalization, "average" was used. N-terminal and lysine modification with iTRAQ were set as fixed modifications, and tyrosine labeling was set as variable modification. Upon completion of the searches, results were exported after setting the ion score filter to 0.1, thereby exporting only results with an expectation value below 0.1 and requiring bold red in order to limit results to the highest scoring match to a particular query listed under the highest scoring protein containing that match. Functional pathways of the listed proteins were analyzed using the MapMan tools (Version 3.6.0RC1) [29,30]. Additionally, a literature search was also used to identify the functions of particular proteins relevant to tomato fruit ripening and Al and La stresses.

Statistical Analysis
Statistical analysis was done by using SAS (Version 9.3; SAS Institute, Cary, NC, USA). Raw expression ratios were log2 transformed and then fit to a normal distribution [25,31]. The log2 fold values were subjected to t-test using the general linear model (GLM) procedure followed by the false discovery rate corrections using SAS. Significant proteins were listed for those passing the two steps of statistical tests (p ≤ 0.05) and with the fold value either lower than 0.65 (<0.65) or higher than 1.51 (>1.51-fold), from Al-treated to La-treated groups.

The Al-Treatment-Induced Proteome Composition in Tomatoes
Using the criteria of two unique peptides per protein, 809 proteins were quantified for MG tomatoes and 600 proteins for both turning and red tomatoes each. When these proteins were analyzed for functional pathways in MapMan, it was found that protein expression in MG tomatoes differs from those from turning and red tomatoes, which display more similar protein expression patterns, as can be seen from Figure 2.  Distribution of functional pathways for proteins from mature green, turning and red stage tomatoes (Tomatoes at turning and red stages contain nearly the same protein distribution, and the mature green (MG) tomatoes show distinct distribution in enzymes, hormone metabolism, protein degradation and stress responses. The cellular pathways were constructed using the tools in MapMan [29,30].).

Differentially-Expressed Proteins in Tomatoes under Al and La Treatments
In MG tomatoes, compared to La 3+ , the treatments with Al 3+ induced a more pronounced and consistent repression of proteins in the functional pathways for protein biosynthesis, photosynthesis, primary carbohydrate metabolic pathways and redox status regulation. When tomatoes reached the turning or red stages, very few proteins showed differential responses to these two metal ions ( Figure S1). In Al-treated MG tomatoes, 17 significantly changed proteins were repressed (Table 1). These Al-repressed proteins include two cell wall proteins, sucrose synthase in sucrose metabolism/catabolism, phosphoglycerate kinase, transketolase, NADP-malic enzyme and aldehyde dehydrogenase in glycolysis and the TCA cycle and the cinnamyl-alcohol dehydrogenase in the secondary metabolic pathway. Six repressed proteins were classified in the RNA-protein synthesis and post-translational modification pathways. Four ribosomal proteins and the 26S proteasome regulatory subunit S2 1A were also repressed. The identification of the chloroplast photosystem II PsbB and RUBISCO in the MG tomatoes concurs with the physiological property that green tomatoes have some photosynthetic activity [32].
In the turning and red tomatoes, only a few proteins expressed significant differences between the Al-treated and La-treated groups. The lectin protein was consistently repressed, while the Kunitz trypsin inhibitor at 1.86-fold (Log2, 0.90-fold) was induced in Al-treated compared to La-treated groups. In turning tomatoes, acid beta-fructofuranosidase (Log2, −0.60-fold) catalyzing the hydrolysis of sucrose was repressed at −1.5-fold, whereas the ABSCISIC ACID STRESS RIPENING 1 protein (Log2, 0.60-fold) was induced at 1.51-fold, in the Al-treated as opposed to the La-treated group. Methionine sulfoxide reductase catalyzes the reduction of methionine sulfoxide in proteins back to methionine, and thus, it has been shown to play an important role in protecting cells from damage from reactive oxygen and nitrogen intermediates [33]. The protein was repressed in Al-treated red tomatoes at −1.53-fold (log2, −0.61-fold).  a Proteins that have passed the t-test followed by FDR correction (p < 0.05) when comparing protein abundance levels between Al-treated and La-treated groups. b Log2 transformed ratio from the Al-treated to the La-treated group; these datasets were used in pathway analysis using MapMan tools (Version 3.6.0RC1) [29,30]. c Accession number in the International Tomato Annotation Group (ITAG) 2.4 protein database.

Patterns of Proteome Changes during the Tomato Ripening Process Associated with Al and La Treatments
Proteins identified in MG, turning and red tomatoes were placed into 10 clusters (Figure 3), and those showing significant changes in any one of the three tomato ripening stages between Al-treated and La-treated groups (log2, ±0.6-fold, Al-treated/untreated; p < 0.05) are listed in Table 2. In Clusters 1 and 2, proteins were either induced by Al treatment or showed no difference with La-treated tomatoes. In this group, only one protein (Cc-nbs-lrr, resistance protein) passed the threshold for significant proteins in MG fruits. In Cluster 3, 15 proteins were significantly repressed in MG tomatoes, but not in turning and red tomatoes. Cluster 5 contains proteins that were repressed by the Al-treated compared to the La-treated group across the three ripening stages. Cluster 6 contains proteins showing no differences between Al-treated and La-treated groups. Cluster 7 contains proteins that were repressed by Al treatment in MG tomatoes, but induced in turning and red tomatoes. Cluster 9 contains proteins showing Al inducement in MG, but repressed in turning and red tomatoes compared to La treatments. Cluster 10 contains protein induced by Al-treatment in MG tomatoes, but repressed in turning and red tomatoes.
in MG tomatoes, but not in turning and red tomatoes. Cluster 5 contains proteins that were repressed by the Al-treated compared to the La-treated group across the three ripening stages. Cluster 6 contains proteins showing no differences between Al-treated and La-treated groups. Cluster 7 contains proteins that were repressed by Al treatment in MG tomatoes, but induced in turning and red tomatoes. Cluster 9 contains proteins showing Al inducement in MG, but repressed in turning and red tomatoes compared to La treatments. Cluster 10 contains protein induced by Al-treatment in MG tomatoes, but repressed in turning and red tomatoes.   Figure 3. Cluster analysis of protein fold changes in Al-treated compared to La-treated tomatoes (the three points showing MG, turning, red stages of tomatoes; the analysis was conducted in MapMan [29,30]).

Discussion
This study identified changes in tomato proteomes as they ripened from MG to turning and then to the red fruits under Al or La treatment conditions. Green tomatoes contain intact chloroplasts and plastids [34]. Photosynthesis of these green tomatoes contributes a portion of carbohydrates to provide for fruit development. Studies have shown that disrupting the photosynthetic activities of these green tomatoes can lead to poor fruit quality [35][36][37], as well as early seed development [38]. Compared to the La treatment, MG tomatoes in the Al-treated group expressed a decline in proteins that are structural components of the photosystem II (PSII) and enzymes in nearly every step of the Calvin cycle, including RUBISCOs ( Figure S2). These protein changes clearly show that Al 3+ treatments induced a more harmful effect than La 3+ in green tomatoes. Similar Al-induced leaf protein changes in the same biological pathway were reported previously [25]. These protein changes would explain the Al-induced decline in photosynthesis activity of tomato plants [10,11], and it is obvious that Al 3+ is more damaging than La 3+ to the respective biological process.
As tomato ripens from MG to turning and red stages, chloroplasts are converted into chromoplasts, and this photosynthetic machinery is dismantled. According to the proteomics analysis, those photosynthetic proteins also diminished at these two latter ripening stages. Thus, the fruit proteome changes appear to concur with the physiological status of these tomatoes. Glycolysis and the citric acid cycle (TCA) provide substrates for the biosynthesis of organic acids, mainly malate and citrate in tomatoes [39,40]. The turning and red tomatoes are characterized by carotenoids' accumulation in chromoplasts, as well as hexoses, glucose and fructose, organic acids, aromatic amino acids and secondary metabolites [41][42][43]. At the turning and red stages, the Al-treated tomatoes were more enriched with the N-carbamoylputrescine amidase, which catalyzes the final step for the conversion of putrescine. Accumulation of this polyamine can serve two roles: one is to provide some stress tolerance mechanisms to the cells, and the second mechanism is related to climacteric ethylene production to promote fruit ripening. Both endogenous and exogenous putrescine were shown to have a positive role in promoting fruit ripening in bananas (Musa AAA group) [44]. However, no such direct correlation has been established in tomato. Instead, several studies have shown that putrescine content in tomato pericarp is not related to normal ripening [45], and an increasing free polyamine level in ripe pericarp in the tomato cultivar Liberty may account for the reduction of climacteric ethylene production [46]. The ripening-specific accumulation of polyamines, spermidine and spermine allows these to act as antiapoptotic regulatory molecules that are able to revive metabolic memory in the tomato fruit [47]. Due to such significant roles of polyamines described above, the enrichment of proteins toward the production of putrescine in Al-treated tomatoes suggests that Al treatment may have some influences on the length of the tomato ripening process.
Tomato fruit ripening is accompanied by seed development and maturation. In the MG stage tomatoes, the embryo has reached full size, but the seed coat is still light-colored with a tender structure. Beyond this point, the seed maturation stage starts with the characteristic dehydration process. In the turning stage tomatoes, Al treatment induced a higher abundance level of ABSCISIC ACID STRESS RIPENING1 protein 1 (ASR1), which is a low molecular weight plant-specific protein encoded by a salt-, drought-and ABA-regulated gene [48][49][50]. Additionally, the Al-treated red tomatoes were also more enriched with hydrophilic proteins, as well as stress proteins. Among these proteins, vicilin-like 7S globulins are a class of main seed storage proteins; they perform a plethora of functions in maintaining protein structural stability against oxidative cellular stresses during embryo development [51]. Previous studies also found that stress proteins were more enriched in tomato seeds derived from Al-treated tomato plants [25], as well as in the germinating radicles upon Al treatments [12]. Thus, we can conclude that tomatoes and seeds grown under Al-treated conditions contain an abundance of stress proteins, which may have a role in preparing the progeny of these plants for the same stress factors.

Conclusions
This study has established the influences of different trivalent cations (Al versus La) on tomato proteomes during fruit development and ripening. Results from the proteomics analysis clearly indicate that an excessive Al level has a much greater impact than La on protein abundance for enzymes in nearly all of the primary and secondary metabolic pathways in MG tomatoes. However, these influences became attenuated or diminished in more mature tomatoes as they ripened to the turning and red tomato stages. These protein changes in tomato fruits in response to Al treatments firmly confirmed the influences of Al toxicity on tomato fruit development and ripening. The study has identified a list of proteins and the candidate cellular functional pathways that responded to Al or/and La toxicity. These results clearly indicate that for fruit metabolome studies, it is imperative to take into consideration the impacts of environmental conditions on fruit physiological properties, including fruit photosynthesis, primary and secondary metabolites, flavor development and many others.