4 Advice to Choose a Sodium Deoxycholate

Protein solubility is a critical prerequisite to any proteomics analysis. Combination of urea/thiourea and 3-[(3-cholamidopropyl)dimethylammonio]-1-propanesulfonate (CHAPS) have been routinely used to enhance protein solubilization for oil palm proteomics studies in recent years. The goals of these proteomics analysis are essentially to complement the knowledge regarding the regulation networks and mechanisms of the oil palm fatty acid biosynthesis. Through omics integration, the information is able to build a regulatory model to support efforts in improving the economic value and sustainability of palm oil in the global oil and vegetable market. Our study evaluated the utilization of sodium deoxycholate as an alternative solubilization buffer/additive to urea/thiourea and CHAPS. Efficiency of urea/thiourea/CHAPS, urea/CHAPS, urea/sodium deoxycholate and sodium deoxycholate buffers in solubilizing the oil palm (Elaeis guineensis var. Tenera) mesocarp proteins were compared. Based on the protein yields and electrophoretic profile, combination of urea/thiourea/CHAPS were shown to remain a better solubilization buffer and additive, but the differences with sodium deoxycholate buffer was insignificant. A deeper mass spectrometric and statistical analyses on the identified proteins and peptides from all the evaluated solubilization buffers revealed that sodium deoxycholate had increased the number of identified proteins from oil palm mesocarps, enriched their gene ontologies and reduced the number of carbamylated lysine residues by more than 67.0%, compared to urea/thiourea/CHAPS buffer. Although only 62.0% of the total identified proteins were shared between the urea/thiourea/CHAPS and sodium deoxycholate buffers, the importance of the remaining 38.0% proteins depends on the applications. The only observed limitations to the application of sodium deoxycholate in protein solubilization were the interference with protein quantitation and but it could be easily rectified through a 4-fold dilution. All the proteomics data are available via ProteomeXchange with identifier PXD. In conclusion, sodium deoxycholate is applicable in the solubilization of proteins extracted from oil palm mesocarps with higher efficiency compared to urea/thiourea/CHAPS buffer. The sodium deoxycholate buffer is more favorable for proteomics analysis due to its proven advantages over urea/thiourea/CHAPS buffer.

View Details

Proteomics techniques have been routinely utilized to study protein compositions and cellular functions of plants, animals and microorganisms [ 16 ]. One of the prerequisites to an effective proteomics analysis is a good protein solubility [ 17 ]. However, it is well known that extracts from plant materials, such as oil palm origin, consists of contaminants like phenolics, polyphenols and lipids. These contaminants strongly interfere with subsequent protein extraction process [ 12 , 18 &#; 20 ]. One of the typical interferences is the inability for the proteins to dissolve completely after protein enrichment with trichloroacetic acid/acetone or ammonium acetate/methanol [ 21 &#; 26 ]. A complete dissolution of proteins in any given sample is highly crucial to enable further downstream mass spectrometric analyses. The use of different buffers, detergents and surfactants to dissolve proteins depend strictly on their compatibility with downstream proteomics analyses. Many studies have employed denaturing buffers containing guanidine hydrochloride [ 27 , 28 ], urea and/or thiourea to solubilize proteins from recalcitrant tissues [ 29 &#; 31 ]. Although sodium dodecyl sulfate has the strongest solubilization power, this detergent is incompatible with protease activity and mass spectrometry [ 32 , 33 ]. Meanwhile, some of the major drawbacks of urea/thiourea are the resulting additional carbamylation modification of N-termini and lysine residues, resulted in raising the false discovery rate of identified proteins [ 34 &#; 37 ] and its incompatibility with tryptic digestion at high concentration [ 38 ]. Guanidine hydrochloride could be used with endoprotease Lys-C but not with a more routine trypsin for protein digestion due to inhibition of the digestion enzyme [ 36 , 39 ]. Another less common detergents like sodium deoxycholate is widely used to solubilize membrane proteins [ 40 &#; 43 ], in addition to improving protein digestions in some studies due to its compatibility with mass spectrometry [ 44 &#; 48 ]. However, there has been no documented work until now that described the use of sodium deoxycholate in solubilizing proteins extracted from recalcitrant and oily plant tissues such as oil palm fruit mesocarps.

Palm oil remains the most efficient oil crop in the world based on its land use (0.36% of the world agricultural land) and productivity (34% of world oils and fats production) [ 1 ]. There has been an increasing interest in studying the oil palm proteome to answer many physiological questions, for instance, the machinery of fatty acid production [ 2 &#; 6 ], fungal disease affecting the oil palm plantations [ 7 , 8 ] and the flowering process [ 9 ] to enhance the sustainability of oil palm. Our proteomics studies revolved around establishing a quantitative model for oil palm lipid metabolism that would coincide with the biochemical, genomics and transcriptomics analyses. Previously, the oil palm transcriptomic studies have revealed elevated transcripts of several fatty acid biosynthetic enzymes in the fruit mesocarp, that lead to an increase in lipid production [ 10 &#; 15 ]. Expression of the proteins related to fatty acid production were also reported to be distinctive throughout oil palm development stages [ 2 , 4 ]. Integration of these omics datasets could be exploited as a platform to further scrutinize the oil palm fruit mesocarp in order to comprehend the exact regulation control of high-value fatty acid production in the effort to optimize the economic value and sustainability of palm oil.

Supervised PLS-DA using MetaboAnalyst 4.0 ( http://www.metaboanalyst.ca/ ) [ 51 ] was employed to determine the correlation of the identified proteins (based on their peak intensities) and different solubilization buffers. Data inputs containing measured m/z value for each peptide and their corresponding retention time and intensities were extracted from the Thermo RAW files. Four replicates representing each of the evaluated solubilization buffers were used (total of 93,777 peaks, with an average of .1 peaks per sample). Peaks of the same group were summed, if they are from one sample, resulting in 5,483 peak groups. For peak matching, these variables were grouped based on their retention time. Mass and retention time were set at 0.025 m/z and 30 secs, respectively. Interquartile range (IQR) filtered out the unusable variables [ 52 ] to improve the regression model. These variables are normally the uninformative regions or noise of mass spectra. Normalization and data scaling based on data dispersion were performed using the sum of intensities and Pareto scaling [ 53 ]. Normalization of the datasets improves the interpretability of the model. Pareto scaling (square root of the standard deviation as the scaling factor) was applied because of the dynamism of the proteomics datasets [ 54 , 55 ]. Statistical model was validated using permutation test as PLS-DA tends to over fit data [ 56 , 57 ]. This test determined if the differences between the evaluated buffers were significant. In the permutation test, the Y-block (class assignment) was permutated times. For every PLS-DA model built, a sum of squares between/within (B/W) ratio was calculated for the class assignment predictions. These ratios were plotted in a histogram. The further to the right the B/W ratio of the original class assignment to the distribution based on the permuted class assignment, the more significant the contrast between the two class assignments from a statistical point of view.

Data acquisitions in positive mode were executed with Thermo Scientific Xcalibur (Version 4.1.31.9) (Thermo Scientific, MA, USA). Generated raw data (.RAW) was processed with Thermo Scientific Proteome Discover, version 2.1 (Thermo Scientific, MA, USA) to generate peak lists in .DTA format for database searching. Tandem (MS 2 ) mass spectra were searched with SEQUEST HT engine against Elaeis guineensis (TaxID = ) and Phoenix dactylifera (TaxID = ) taxonomies (containing 35,972 and 33,101 protein sequences, respectively, as of 30 th October ) in NCBI protein database. Mass tolerances for peptide and product ions were set to 20 ppm and 0.5 Da. Trypsin was designated as the protease with two missing cleavages allowed. Carbamidomethylation on cysteine and lysine was set as the fixed modification while oxidation of methionine and deamidation of asparagine and glutamine were searched as variable modifications. Proteins were accepted if they had at least one Rank 1 peptide. A decoy database contained randomized sequences of searched taxonomies. All database searches were also performed against the decoy database to determine the false discovery rate. All peptide spectral matches were validated using the Percolator version 2.04 (component of Proteome Discover) based on q-value at a 1% false discovery rate. Venn diagram of the identified proteins from the evaluated solubilization buffers was created using a free web-based program ( http://bioinformatics.psb.ugent.be/webtools/Venn/ ). Biological process, cellular component and molecular function of the identified proteins were annotated using the Retrieve/ID mapping tool in Uniprot ( https://www.uniprot.org/uploadlists/ ). Gene ontology (GO) terms associated with the identified proteins from all the evaluated solubilization buffers were collected from the Uniprot-GOA database ( http://www.ebi.ac.uk/GOA ).

Separation and spectra acquisition of the protein digests was conducted with an EASY-nano liquid chromatography (EASY-nLC) System (Thermo Scientific, MA, USA), coupled to a Q Exactive Plus Hybrid Quadrupole-Orbitrap mass spectrometer (Thermo Scientific, MA, USA). Tryptic digests were reconstituted in 20 μL of 0.1% FA and 5% ACN. A sample volume of 2 μL was injected into an Acclaim PepMap 100 C18 reversed phase column (3 μm, 0.075 x 150 mm) (Thermo Scientific, MA, USA) for peptide separation. The column was equilibrated with 95% mobile phase A (0.1% FA) and 5% mobile phase B (0.1% FA in ACN). A gradient of 5&#;35% mobile phase B in 70 min was employed to elute the bound peptides at a flow rate of 300 nL min -1 . Gas-phase peptide ions were generated by electrospray ionization using a spray voltage of V. Peptide precursors survey scan was acquired in the Orbitrap mass analyzer with a mass range of m/z 310&#; and resolving power of 70,000. Maximum injection time applied was 100 ms. Peptide precursors with charge state of 2&#;8 were chosen for tandem MS (MS 2 ). Tandem MS conditions consisted of rapid scan rate with the linear ion trap mass analyzer using a resolving power of 17,500, 0.7 m/z isolation window and an maximum injection time of 60 ms. Precursors were fragmented using collision-induced and high-energy collision-induced (CID and HCD) at a normalized collision energy of 28%, respectively. Mass range scanned was from m/z 110&#;. The mass spectrometry proteomics data have been deposited to the ProteomeXchange Consortium via the PRIDE [ 50 ] partner repository with the dataset identifier PXD and 10./PXD.

To obtain 100 μg of proteins for electrophoretic separation, the solubilized proteins were re-precipitated with cold ammonium acetate-saturated methanol. The precipitated proteins were then dissolved in Laemlli buffer (62.5 mM Tris-HCl, pH 6.8, 2% SDS, 25% glycerol, 0.01% bromophenol blue, 0.005% β-mercaptoethanol) and denatured by boiling at 95°C for 4 min [ 3 ]. 100 μg protein was loaded into each lane on a 1.0 mm in-house casted 12% polyacrylamide gel. Electrophoresis was conducted in a Bio-Rad mini-PROTEAN Tetra Cell apparatus (Bio-Rad Laboratories Inc., Hercules, CA) at 200 V for 1 h. Following electrophoresis, the separated proteins were fixed for 30 min in a fixing solution (50% ethanol, 10% acetic acid) and stained with an in-house prepared Colloidal Coomassie G-250. The gel was destained with Milli-Q water until the gel background was clear. The gel was scanned as digital image using Bio- Plus scanner (Microtek, Hsinchu, Taiwan) according to the manufacturer&#;s instructions.

Protein digestion was performed according to Lau and co-workers [ 3 , 4 ]. To obtain 50 μg of proteins for digestion, solubilized proteins in Buffer A, B and C were re-precipitated with cold ammonium acetate-saturated methanol. The precipitated proteins were re-suspended in 0.1 M ammonium bicarbonate and 1 M urea before reduction and alkylation using 50 mM tris(2-carboxyethyl)phosphine and 150 mM iodoacetamide, respectively. Sodium deoxycholate (1% w/v) was added to the protein solution prior to digestion with 4 μg of modified sequencing grade trypsin (Promega, Madison, WI, USA) in 50 mM NH 4 HCO 3 for 16 h at 37°C. Sodium deoxycholate was removed after tryptic digestion by acidification using 0.5% formic acid and centrifugation at 14 000 g (RA-300, Kubota ) for 15 min at ambient temperature. The peptide solution was then dried in a centrifugal evaporator (CentriVap Concentrator, Labconco, MO, USA). Peptide clean-up&#;The dried peptide pellet was resuspended in 200 μL of 0.1% formic acid. Acetonitrile, methanol and 0.1% formic acid-conditioned Empore solid phase extraction disks (3M Purification, Inc., MN, USA) were added to the peptide solution and incubated at ambient temperature with slight agitation for 4 h. The bound peptides on the C18 membrane disks were sequentially eluted with 50% ACN in 0.1% FA for 2.5 h.

In this study, four different solubilization buffers were used to solubilize the ammonium acetate/methanol precipitated proteins. A volume of 600 μL of each evaluated buffers was added to the protein pellet. Buffer A: Urea/thiourea/CHAPS&#; 7 M urea, 2 M thiourea, 4% CHAPS, 0.4% DTT, 10 mM Tris base; Buffer B: Urea/CHAPS&#;7 M urea, 4% CHAPS, 0.4% DTT, 10 mM Tris base; Buffer C: Urea/sodium deoxycholate&#; 7 M urea, 4% sodium deoxycholate, 0.4% DTT, 10 mM Tris base; Buffer D: Sodium deoxycholate&#; 4% sodium deoxycholate, 0.4% DTT, 10 mM Tris base. Commercially available 2D Quant Kit (GE Healthcare Life Sciences, Uppsala, Sweden) was then utilized to determine protein content in the samples. Bovine serum albumin provided with the kit was used as the protein calibration standard and each quantitation was performed in duplicate. A 4-fold dilution was performed on the proteins solubilized with sodium deoxycholate-containing buffers when Pierce 660 nm Protein Assay Reagent (Thermo Scientific, IL, USA) or Coomassie-based Bradford was used. Without the dilution, precipitated sodium deoxycholate would interfere with the absorbance readings.

Proteins were extracted according to Lau and co-workers with some modifications [ 49 ]. 10 g of sliced mesocarps were ground and mixed well with 25 mL of cold acetone containing 10% trichloroacetic acid and 1 mM dithiothreitol on ice. The slurry was then centrifuged at 13,000 g for 10 min at 4°C (RA-300 rotor, Kubota , Kubota Corporation, Tokyo, Japan). The washing step was repeated once before adding 25 mL of cold 80% methanol containing 0.1 M ammonium acetate to the precipitate; mixed well and centrifuged as before, on ice. The precipitated mesocarp pellet was washed with 25 mL of cold 80% acetone. The mixture was mixed well and centrifuged again at 13,000 g for 10 min at 4°C. Pellet was gently re-suspended in 15 mL of extraction buffer containing 0.7 M sucrose, 1 M Tris-HCl, pH 8.3, 5 M NaCl, 50 mM DTT, 1 mM EDTA and a tablet of Roche protease inhibitors. The resuspension was sonicated using ultrasonic bath for 30 mins (Townson & Mercer Ltd., England, UK). The mixture was then sieved through two layers of Miracloth (Calbiochem, EMB Millipore Corporation, Billerica, MA) to separate non-macerated plant materials. An equal volume of fresh 50 mM, pH 8.0 Tris-saturated phenol (15 mL) was added to the mixture, mixed well and centrifuged at 15,000 g for 15 min at 4°C (RA-300 rotor, Kubota ) for phase separation. Proteins in the upper phase were precipitated by adding five volumes of cold ammonium acetate-saturated methanol (25 mL) to one volume of phenol phase, mixed well and incubated at -20°C overnight before being centrifuged at 15,000 g for 15 min at 4°C (RA-300 rotor, Kubota ). The protein pellets were then rinsed with 5 mL of cold ammonium acetate-saturated methanol and washed three times with 5 mL of cold 80% acetone. The protein pellet was air-dried for 5 min.

Results and discussion

Proteomics studies are critically dependent on soluble and good quality proteins. The study was conducted to determine the efficiency of sodium deoxycholate (SDC) in solubilizing oil palm mesocarp proteins compared to other urea/CHAPS-containing buffers. The efficiency was compared in terms of their total protein yields, electrophoretic patterns, chromatographic and mass spectra patterns, as well as number of identified proteins and their resulting gene ontologies. A statistical analysis using partial least squares-discriminant analysis (PLS-DA) was also incorporated into the evaluation criteria to determine the variability of the solubilization buffers. In this study, comparisons were made between SDC and a routinely used urea/thiourea/3-[(3-cholamidopropyl)dimethylammonio]-1-propanesulfonate (CHAPS) buffers for solubilization of proteins derived from oil-rich plant tissues such as oil palm [29&#;31]. Effect of SDC in replacing CHAPS (in urea/SDC and urea/CHAPS buffers) were also compared to evaluate the different detergents and the complimentary of urea/SDC in protein solubilization efficiency.

The first criteria to determine the efficiency of SDC as solubilization buffer was to investigate the total protein yields after solubilization in different buffers. The protein quantitation for all the proteins solubilized in urea/thiourea/CHAPS, urea/CHAPS, urea/SDC and SDC buffers was repeated three times. As shown in S1 Fig, the solubilization power of all buffers tested was quite satisfactory and no extensive loss in protein yield was recorded. Protein yield from urea/thiourea/CHAPS buffer was 1.13 ± 0.07 μg/μL. Protein yield from urea/CHAPS buffer was 1.17 ± 0.11 μg/μL. Meanwhile, protein yields from urea/SDC and SDC buffers decreased to 0.90 ± 0.07 μg/μL and 0.86 ± 0.05 μg/μL, respectively (compared to urea/thiourea/CHAPS). The total protein yield for SDC buffer was 0.86 ± 0.05 μg/μL. This was a 0.27 μg/μL reduction compared to the urea/thiourea/CHAPS buffer. Clearly, the presence of strong chaotropic agents and detergent is unparalleled in their solubilization efficiency. Meanwhile, the combination of urea/SDC did not improve the solubilization efficiency as compared to urea/thiourea/CHAPS, urea/CHAPS or even SDC buffers. We would expect a contrasting effect of the combination as both urea and SDC are also chaotropic agent and surfactant. Thus, the results suggested that CHAPS was not substitutable by SDC as detergent. Our observations also established that prior to protein quantitation, the assays containing SDC surfactant needed to be diluted about 4-fold (< 1% SDC) to avoid interference to the absorbance reading (data not shown). Unless the proteins were precipitated before the quantitation assay, or GE 2-D Quant kit was used to determine the protein content, this step is critical to achieve accurate and reproducible protein yield.

Qualitative comparison of the solubilization efficiencies using polyacrylamide gel showed that proteins solubilized by all the buffers were separated into well resolved and good intensity bands without any apparent sign of degradation or interference due to impurities (S2 Fig). Relative number of protein bands was identical except for proteins solubilized in urea/SDC buffer. The important pattern shown by these data was that, although solubilized proteins in SDC buffer resulted in lower yield compared to urea/thiourea/CHAPS buffer, majority of the solubilized proteins from both buffers were still detected in gel. However, a number of bands were missing for SDC buffer, as indicated in S2 Fig. That might explain the lower total protein yield for SDC buffer compared to urea/thiourea/CHAPS and urea/CHAPS buffers. Meanwhile, electrophoretic profile of urea/CHAPS and urea/SDC buffers showed a reduction in the number of protein band for the latter. The combination effect of urea and SDC seemed to lower the number of separated proteins on polyacrylamide gel or reduce the band intensities. This electrophoretic pattern profile was in agreement with the protein yield measurements obtained earlier for both solubilization buffers. We deduced that the possible reason might be due to interference from an incomplete removal of high concentration of SDC (4%) prior to gel electrophoresis but more works to elucidate this observation was necessary.

The solubilized proteins were subsequently tryptic digested and analyzed mass spectrometrically. An EASY-nano liquid chromatography (EASY-nLC) System (Thermo Scientific, MA, USA), coupled to a Q Exactive Plus Hybrid Quadrupole-Orbitrap mass spectrometer (Thermo Scientific, MA, USA) was used to detect the separated peptides. Base peak chromatograms for the separated peptides from the four different solubilization buffers were presented in . Comparison of the chromatograms for all the tested solubilization buffers revealed similar profiles. However, unlike the urea/SDC and SDC buffers, proteins solubilized in urea/thiourea/CHAPS and urea/CHAPS buffers gave a signature peak at approximately 72 minutes into the chromatographic separation (indicated with a red box in ). Complete removal of excess CHAPS was challenging although ammonium acetate-saturated methanol had removed most of the CHAPS (performed prior to protein digestion). As a result, an intense peak ion would still be noticeable at 615 m/z (MH+) in the mass spectra (fragmentation of peak at 72 minutes). However, this particular contaminant peak was not detected with urea/SDC and SDC buffers given that CHAPS was not added. Although CHAPS had been reported to prevent protein loss through precipitation or aggregation [58], this was not observed in this study as major peaks were present in all of the evaluated buffer chromatograms. The relative abundance of the ions was also comparable among all the solubilization buffers.

A more in-depth comparison between urea/thiourea/CHAPS and SDC buffers was made using three representative peptides corresponded to some of the targeted fatty acid biosynthetic enzymes in the oil palm proteomics works. For this study, the peak intensities, ion scores, detected unique peptides and the coverage of b and y ion series of these representative peptides were compared ( ). Peptide AALESDTMVLAFEAGR, with a SEQUEST ion score of .45 was identified to enoyl-ACP reductase in urea/thiourea/CHAPS buffer ( ). In total, 11 unique peptides were identified to enoyl-ACP reductase. With SDC buffer, slightly lower SEQUEST ion score (.80) and total number of unique peptides (9) were acquired. More importantly, the coverage of b and y ions for AALESDTMVLAFEAGR from both buffers was similar. Nevertheless, the b and y ion intensities for AALESDTMVLAFEAGR in SDC buffer were relatively higher. An ideal MS/MS spectrum would have high signal to noise ratio and contain all the N-terminal b ions and C-terminal y ion fragments, as observed with the described peptides detected in urea/thiourea/CHAPS and SDC buffers. also showed the comparison of another two peptides detected from urea/thiourea/CHAPS and SDC buffers in term of their SEQUEST ion scores, unique peptides and coverage of b and y ions. Both peptide KGGEYEPEEQPEADTDYSR and EEQDSYAIQSNER corresponded to phospholipase D alpha 1 and acetyl-CoA acetyltransferase, respectively. Ion score for peptide KGGEYEPEEQPEADTDYSR was increased in SDC buffer (715.90) relatively to urea/thiourea/CHAPS buffer (629.01) ( ). Total detected unique peptides for phospholipase D alpha 1 in both buffers remained at 14 peptides. In another comparison, ion score for peptide EEQDSYAIQSNER for both buffers were almost similar at 216.04 (urea/thiourea/CHAPS) and 240.10 (SDC buffer). The total number of detected unique peptides was the same at 6 for both buffers ( ). In both cases, their b and y ion coverages remained similar but with relatively higher intensities for SDC buffer. Acetyl-CoA acetyltransferase, enoyl-ACP reductase and phospholipase D are important enzymes for initiation, synthesis of fatty acids and their metabolism, respectively [59]. They are of interest in the proteomics studies of oil palm mesocarps and therefore, it is crucial to be able to identify them using high quality mass spectra regardless of the solubilization buffers used for oil palm mesocarp proteins.

Proteomics analysis was subsequently performed on the oil palm mesocarp proteins solubilized in four different solubilization buffers. A total of proteins ( peptides) was identified from urea/thiourea/CHAPS buffer compared to proteins ( peptides) from SDC buffer. Urea-induced carbamylation on lysine residues was found on 225 peptides from urea/thiourea/CHAPS buffer. In contrast, only 44 peptides were modified on the same amino acid in SDC buffer, a reduction of 67.3%. Removal of thiourea did not affect the protein solubilization significantly as proteins ( peptides) were identified from urea/CHAPS buffer. The combination of urea/SDC appeared to reduce the number of identified proteins by 1.29% only ( proteins, peptides), compared to urea/CHAPS buffer. Of the total identified peptides, carbamylation on lysine residue occurred on 356 and 336 peptides for urea/CHAPS and urea/SDC buffers, respectively. The outcome of the proteomics analysis clearly strengthened the results acquired from their protein quantitation assays (S1 Fig) and one-dimensional gel electrophoresis profile (S2 Fig). Furthermore, the modification search revealed that peptide carbamylation had occurred in all buffers involving urea, with varying degrees. Results from further examination of the protein ( ) and peptide ( ) identifications were shown in four-way Venn diagrams. Urea/thiourea/CHAPS and SDC buffers both shared 763 proteins in common (62.0% of total identified proteins). About 34.0&#;41.2% of the total identified proteins from urea/thiourea/CHAPS (399) and SDC (534) buffers were unique to each buffer. Urea/thiourea/CHAPS and SDC buffers both shared 820 peptides while 501 and 641 peptides were unique to each respective buffer. 954 or 75.0% of the total identified proteins from urea/CHAPS and urea/SDC buffers were shared. Urea/CHAPS and urea/SDC buffers had 335 (415 unique peptides) and 302 unique proteins (371 unique peptides), respectively.

With competitive price and timely delivery, chenlv sincerely hope to be your supplier and partner.

Urea/thiourea/CHAPS had better solubilization efficiency than SDC buffer. However, the proteomics results indicated that SDC buffer was able to elevate the total identified proteins to a greater extent, although only 62.0% of the total identified proteins were shared between the buffers. Depending on the biological questions to be elucidated, the remaining 38.0% identified proteins might not be crucial, at least not in the oil palm proteomics studies. Further works are in progress to look into these unique proteins [60]. The differences could be due to the characteristic of urea/thiourea/CHAPS in disrupting hydrogen bonds and hydrophobic interactions of the proteins for solubilization. As mentioned by Broeckx and co-workers [61], protein crosslinking reversion was improved in an alkaline environment. Therefore, a slightly basic environment provided by a fresh urea/thiourea/CHAPS buffer might facilitate the protein solubilization. Unlike urea/thiourea/CHAPS, SDC is a deoxycholic acid derivative. However, the basic environment in the SDC buffer was conferred by the addition of Tris. The presence of dithiothreitol could also assist in the reduction of internal disulfide bonds. In the assessment of SDC as a detergent substitute, SDC was evidently not able to perform as effectively as CHAPS (in urea buffer) based on the number of identified proteins. Unlike SDC, CHAPS could protect the protein activity due to its zwitterionic characteristic while SDC might induce denaturation of the proteins to some extent [62&#;64]. This was a very likely reason as to the slight differences observed in this study relating to the number of identified proteins and peptides.

To further evaluate the efficiency of the solubilization buffers, identified proteins from all the buffers were categorized according to their biological processes, subcellular localizations and molecular functions ( ). All proteins identified were annotated with same gene ontology terms regardless of the solubilization buffers used. Majority of the gene ontology terms for both SDC and urea/CHAPS buffers were higher relatively, compared to urea/thiourea/CHAPS and urea/SDC buffers ( ). In most biological processes, number of annotated proteins from urea/CHAPS buffer was slightly higher compared to SDC buffer (except in response to stimulus, cellular component organization, multiorganism processes and reproductive process). Note that number of identified proteins for both buffers were comparable ( proteins for SDC and for urea/CHAPS, respectively) and higher relatively to the rest of the evaluated buffers. Majority of the proteins were involved in metabolic and cellular processes. illustrates the cellular components of the identified proteins. In overall, proteins from all the buffers were annotated with the same cellular localization. Most proteins were located in cell, followed by membrane, protein-containing complex and organelles. Least proteins were localized in the extracellular, plasmodesma, mitochondrial matrix and microtubule. Molecular activity of the proteins identified was also classified using gene ontology analysis ( ). There were 11 activities associated with all the proteins from the different solubilization. Most proteins were implicated in binding and catalytic activities. Less than 10 proteins were associated with transcription regulator, nutrient reservoir, phosphorelay sensor kinase and photoreceptor activities.

A more detailed comparison of the gene ontology for biological process, subcellular location and molecular activity was made on the identified proteins from urea/thiourea/CHAPS and SDC buffers ( ). The comparisons revealed that SDC buffer had profound effects on the resulting proteome. In particular, SDC buffer had enriched proteins in every functional categories. The enrichment in the biological regulation and metabolic processes of protein identified using the SDC buffer, could contribute significantly to our efforts in understanding the regulation of oil palm fatty acid biosynthesis mechanism. In terms of cellular components, there was no obvious significant difference or additional gene ontology terms observed in the comparative analysis. Proteins localized in the membrane (5.6%) and cell (6%) were slightly enriched, which coincide with the use of SDC buffer. Molecular activities of the identified proteins from SDC buffer were also enriched, particularly catalytic activity (5.8%) and binding (6.6%).

For additional examination of the efficiency variation between all the solubilization buffers, the proteomics data analysis of four replicates corresponded to the four different solubilization buffers were statistically compared using Partial Least Squares Discriminant Analysis (PLS-DA). PLS-DA comprehensively determined the linear relationship between different buffers (Y response matrix) and the corresponding peptide spectra (X predictor matrix). PLS-DA was applied in the context of our study due to its ability to analyze data with complicated, noisy, collinear and incomplete variables in both X and Y. The PLS-DA model qualities were cross-validated with a 10-fold cross-validation method based on R2 and Q2 parameters [65]. R2 = 1 is an indication of a perfect data description by the model. In this study, the corresponding R2 and Q2 values for each component were listed in Tables and . The value of R2 &#;Q2 is less than 0.3 for up two components, which indicated that the model has good predictability. The cross-validation correlation coefficient R2, Q2 revealed a value of 0., which was an indicator of a model with high predictive model. A three-component model was the best classifier (S3A Fig). The performed permutation tests, another PLS-DA model cross-validation method, showed that the group separation was statistically insignificant at p = 0.451 (S3B Fig). The original model (indicated with red arrow) was part of the permutated models. The result showed a good elucidation and buffer type classification information [56]. In the supervised PLS-DA of peptide intensities, a clear grouping based on the buffers evaluated were achieved ( ). The model was built between dependent variables (Principal Component 2), represented the urea/thiourea/CHAPS (A), urea/CHAPS (B), urea/SDC (C) and SDC (D) buffers; and independent variables (Principal Component 1) (peptide spectra). The explained variance for the first and second principal components were 25.0% (PC1) and 16.7% (PC2), respectively. Groups of urea/thiourea/CHAPS, urea/CHAPS and urea/SDC were clustered negatively. However, SDC buffer group was positively clustered compared to the rest of the buffer groups. Clearer correlation between the buffer groups was projected in a three-dimensional scores plot, based on three principal components (PC1, 2 and 3) ( ). In this model, it was apparent that urea/thiourea/CHAPS, urea/CHAPS buffers and urea/SDC (A, B and C) were closely related to each other, suggesting that the buffers shared similar characteristics. In this study, the characteristic was possibly the urea additive. Conversely, SDC buffer was located away because there was no similarity in the buffer components. Loadings of the buffer groups (A-D) in this study were explained by the first two principal components (PC1 and PC2) ( ). The principal component loadings used to detect variability, showed that the solubilization buffers were indistinguishable by the profile of peptide spectra. Most loadings were clustered together except for several &#;outliers&#;. The directions of the loading variables indicated that they were positively (to the right of the x-axis) and negatively (to the left of the x-axis) correlated.

Table 1

Measure1 components2 components3 components4 componentsAccuracy0.250.250..R20....Q20....Open in a separate window

Table 2

Measure1 components2 components3 components4 componentsAccuracy0.250.250..R20....Q20....Open in a separate window

All the four buffers were able to solubilize the extracted proteins from oil palm fruit mesocarps to a variety of extents. A minimum concentration of 4% (w/v) SDC was used in this study as any concentration less than 4% would not able to solubilize the proteins completely (based on qualitative observations, data not shown). The fluctuation in the total protein yields and electrophoretic profile indicated that urea/thiourea/CHAPS buffer remained the most effective solubilization buffer. Nonetheless, the total protein yields determined from all the solubilization buffers were still satisfactory and the difference was only 0.2 μg/μL between urea/sodium deoxycholate and sodium deoxycholate buffers and urea/thiourea/CHAPS and urea /CHAPS buffers. For urea/CHAPS and urea/SDC comparisons to determine additive effects on protein solubilization, CHAPS&#;containing buffer was more proficient to solubilize the proteins. However, when chromatogram, spectra, identified protein and peptide numbers, gene ontologies of all the solubilization buffers were compared, at 4% (w/v), SDC alone was broadly applicable to the oil palm mesocarp proteins, despite the lower protein yield and additive efficiency. Detailed statistical approach to analyze oil palm proteomics datasets was also presented in this study. The results were in agreement with the mass spectrometric analysis that there were only minor variations (based on the group clustering) between the different solubilization buffers. These results were significant as four replicates (n = 4) were used for each buffer in the PLS-DA. Inability to find similar experimental set-up in the literatures prevented the comparison of the results acquired from this study in accessing the solubilization efficiency. Currently, SDC has only been used to solubilize membrane proteins [40&#;43] and to enhance tryptic protein digestion [44&#;48]. SDC is an acid-removable detergent that able to disrupt cell membranes and protein to protein interactions, similar to sodium dodecyl sulphate. The major advantage of employing SDC is that the detergent is removable through acid precipitation either before or after enzymatic digestion without causing any loss or variability to protein identification rate [46, 66]. Removal of sodium dodecyl sulphate, urea, CHAPS using filter-aided sample preparation (FASP) and zip tips [66&#;68] still resulted in interferences to liquid chromatography runs and mass spectrometric analysis [68, 69]. The limitation of the utilization of SDC in protein solubilization, which was observed from the study, was the interference to the protein quantitation. However, this limitation could be circumvented by incorporating a 4-fold dilution prior to protein content determination using a colorimetric approach. Alternatively, the proteins could be precipitated before quantitation. Finally, further studies are necessary to determine if SDC could also be applied to animal and human-based proteins for solubilization.