The authors have declared that no competing interests exist.
Conceived and designed the experiments: JOW MJT CJF. Performed the experiments: JOW. Analyzed the data: JOW CJF. Contributed reagents/materials/analysis tools: MJT CJF. Wrote the paper: JOW CJF.
In the search for a replacement for fossil fuels bioethanol comes out as one of the most promising alternatives. For sustainable production without interference with food production it is necessary to use lignocellulosic sources such as agricultural or forestry residues as raw materials
The most widely used microorganism for production of fuel ethanol, be it 1st or 2nd generation, is
Cell immobilization can be done in a number of ways, but the one giving the highest local cell density is undoubtedly encapsulation in a semi permeable membrane. Local cell densities of several hundred grams dry weight per litre of capsule volume have been achieved
Quantitative proteomics is of utmost importance for the understanding of changes in cellular physiology arising from different treatments of the cells, as different proteins, including post translationally modified variants, are directly linked to metabolic fluxes and cellular structure, and therefore ultimately determine the physiology. There are a number of different quantitative proteomic methods available, with two major types of protein separation, namely two dimensional gel electrophoresis (2-DE) and multidimensional liquid chromatography (multidimensional protein identification, MudPIT), often called nLC-MS/MS. For identification of proteins both methods depend on mass spectrometry in combination with database searches. For 2-DE one of the currently most popular methods is 2-D difference gel electrophoresis (2-D DIGE), where proteins from different samples are labelled with different fluorescent probes, enabling quantification of proteins from different samples in the same gel
In this study, we compare the protein expression levels in yeast cells growing anaerobically either in liquid core capsules enclosed by alginate-chitosan membranes or in suspension, using both the 2-D DIGE approach and a MudPIT approach with TMT®. In addition to the overall elucidation of physiological changes in the cells due to encapsulation, the aim was to find possible reasons for the increased tolerance towards lignocellulosic derived inhibitors as well as for the enhanced yield of ethanol and lower glycerol yield of encapsulated yeast
The
Aerobic cultures for cell propagation were grown in 250 ml cotton-plugged conical flasks. Anaerobic batch cultivations were performed in 250 ml conical flasks equipped with a rubber stopper fitted with a loop trap filled with sterile water to permit produced CO2 to leave the flasks, and stainless steel capillaries for sample removal. The growth medium used for the batch cultivations was a defined glucose medium (DGM), as previously reported
The capsules were prepared by the liquid-droplet-forming method
Capsules were formed by dripping the CMC-yeast-CaCl2 solution into the stirred sodium alginate solution through syringe needles. The capsules were gelled for 10 minutes, washed with ultra-pure water for 10 minutes and hardened in 1.3% (w/v) CaCl2 solution for 20 minutes. The calcium-alginate capsules were thereafter treated in a 0.2% (w/v) low molecular weight chitosan (Product number 448869, Aldrich) solution with 300 mM CaCl2 in 0.040 M acetate buffer, pH 4.5, at a ratio of 1∶10 of capsules to solution for 24 hours. The treatment was performed in 0.5 l Erlenmeyer flasks in a water bath at 30°C at 130 rpm. Chitosan molecules are incorporated in the alginate matrix, thereby improving the capsules' strength by creating alginate-chitosan membranes
Approximately 15 ml of cell seeded capsules were cultivated aerobically in 100 ml DGM containing 40 g l−1 glucose for 24 hours in a shaker bath (130 rpm) at 30°C. The capsules were thereafter rinsed with sterile 0.9% NaCl (Scharlau) and transferred to fresh medium for another 7 hours. Fifty capsules were subsequently transferred to 100 ml of DGM, 40 g l−1 glucose, for anaerobic batch cultivations, giving a starting cell concentration of 0.77±0.04 g DW l−1 liquid volume. Samples for protein expression analysis were taken after 25.4±0.3 hours, when the glucose concentration had reached 12.6±0.7 g l−1. Cells from 8 capsules were washed out with ice-cold sterile ultra-pure water, and harvested at 10,000×g for 1 minute at 4°C. The pellet was immediately frozen in N2(l) and subsequently kept at −80°C until analysis.
Propagation of suspended yeast was started from aerobic 24 hours cultivations in 100 ml DGM (50 g l−1 glucose), by a 1% dilution into fresh DGM. After 23.5 hours, cells were harvested (3500 g, 4 minutes) and resuspended in 100 ml fresh DGM (40 g l−1 glucose) for anaerobic cultivations, at an initial cell concentration of 0.74±0.01 g DW l−1. The cells were grown anaerobically and samples for protein expression analysis were taken at 8.3±0.1 hours, when 15.8±0.1 g l−1 glucose remained in the cultivations. Cells were harvested by centrifugation of 20 ml cell suspension for 1 min at 3,500×g, 4°C, followed by washing of the cell pellet with ice-cold ultrapure water and centrifugation for 1 min at 10,000×g, 4°C. The pellet was immediately frozen in N2(l) and subsequently kept at −80°C until preparation of protein extracts and analysis.
After a series of trial cultivations, the sampling times were chosen to obtain samples when the residual extracellular glucose concentration and total amount of cells in the cultivations were approximately equal, and glucose was being consumed at constant rates. Free cells were sampled at a slightly higher residual glucose concentration, due to a higher biomass yield on glucose.
Cells from five biological replicates each of free and encapsulated yeast were lysed and the proteins extracted and cleaned up prior to dividing the samples for 2-D DIGE and quantitative nLC-MS/MS.
For the nLC-MS/MS, precipitated protein pellets of 100 µg of each samples (three biological replicates each of free and encapsulated yeast) were digested with trypsin and labelled with a six-plex set of Tandem Mass Tag reagents following the manufacturer's instructions (Thermo Fisher Scientific). Nano LC-MS/MS analysis was performed on a LTQ Orbitrap Velos instrument (Thermo Fisher Scientific, Inc., Waltham, MA, USA) interfaced with an in-house constructed nano-LC column. MS data analysis was performed using Proteome Discoverer version 1.2 (Thermo Fisher Scientific).
2-D DIGE analyses
Further details of the proteomic analyses are described in
The ratios between the abundances of proteins in suspended and encapsulated yeast are presented by one to three numbers in parentheses after the name of the protein. The first, and in most cases the only, number represents the ratio (fold change, FC) obtained by n-LC-MS/MS. When applicable, this is followed by the average ratio obtained from 2-D DIGE spots with unique significant protein hits, and lastly the average ratio obtained from spots with significant hits for more than one co-migrating proteins in 2-D DIGE. Thus, Tps1p (1.77, 1.59, 1.63) indicates a FC of 1.77 estimated by the nLC-MS/MS approach, an average FC of 1.59 estimated from spots with a unique significant hit for Tps1p, and an average FC of 1.63 estimated from spots with significant hits from both Tps1 and co-migrating proteins.
The amounts of metabolites were quantified by HPLC using an Aminex HPX-87H column (Bio-Rad) at 60°C with 5 mM H2SO4 as eluent at a flow rate of 0.6 ml min−1. A refractive index detector was used for the detection and quantification of glucose, acetic acid, lactic acid, glycerol and ethanol.
The cell dry weight was measured in predried and preweighed glass tubes. Cells were separated by centrifugation and washed once with ultra-pure water before drying for approximately 24 h at 105°C. Cells from capsules were released by crushing the capsule followed by extensive washing of the capsule debris with ultra-pure water.
The biomass and metabolite yields as well as the carbon balance were calculated from the determined concentrations at the end of the fermentations, i.e. the time of sampling for proteome analysis. Error intervals are given as 95% confidence intervals of the mean, unless otherwise stated.
Free and encapsulated cells were grown anaerobically in shake flasks starting at the same initial cell concentration, 0.75±0.02 g DW l−1 medium volume. The cells were sampled at residual glucose concentrations of 14.2±1.2 g l−1 while consuming glucose and producing ethanol at constant rates (
Glucose and ethanol concentration profiles of encapsulated (◊, □) and free (▵, ○) cells during anaerobic batch cultivations.
The yields of major metabolites and biomass were significantly different between cells grown in the two ways. The encapsulated cells had higher ethanol yield than the free cells, whereas free cells had higher glycerol, acetate and biomass yields (
YSE | YSGly | YSAce | YSLac | YSBiomass | Carbon recovery (%) | |
Free | 427±4 | 56±1 | 5±0 | n.d. | 74±2 | 98.8±1.0 |
Encapsulated | 439±3 | 48±2 | 2±0 | 3±0 | 41±3 | 96.3±0.8 |
Yields are shown as mg product per g consumed glucose from the start until the sampling of cells in the anaerobic batch cultivations. The molar CO2 production was assumed to be the same as the sum of ethanol and acetate. Error intervals shown are 95% confidence intervals, with n = 5. YSE – Ethanol yield, YSAce – Acetate yield, YSGly – Glycerol yield, YSLac – Lactate yield, YSBiomass – Biomass yield, n.d. – not detected.
The slower growth rate and lower biomass yields of encapsulated cells are likely an effect of mass transfer limitations to the cells in the middle of the capsules. The amount of biomass per capsule was 1.5±0.1 mg at the start of the cultivation, which increased to 3.9±0.1 mg per capsule at the time of sampling. This corresponds to approximately 300 g DW (l capsule volume)−1. It has previously been shown, that for a flocculating yeast strain mass transfer limitations occur in flocs larger than 100 µm
Capsules full of cells at the time of sampling for proteome analysis. Major unit of the ruler is in centimetres.
Of the 842 proteins identified with nLC-MS/MS (
Volcano plot illustrating the distribution of all proteins identified with the nLC-MS/MS approach. Significantly up- and down-regulated proteins (|fold change| ≥1.3, x-axis; FDR adjusted p value≤0.05, y-axis) are highlighted in green and red respectively. Statistically up- and down-regulated proteins with non-significant biological changes (|fold change| <1.3) are shown in light green and orange, respectively, and proteins with non-significant differences between the free and encapsulated yeast are shown in grey.
Distribution of functional categories (A) and cellular localizations (B) of identified proteins in encapsulated and free
Category | Sub-category | p value | Proteins |
Metabolism (30) | Metabolism of the cysteine – aromatic group (5) | 7.7E-03 | Aro2p Aro7p Gly1p Cys4p Gcv1p |
Metabolism of glycine (2) | 6.3E-03 | Gcv1p Gly1p | |
Protein Synthesis (26) | Ribosome biogenesis (15) | 7.2E-05 | Tif5p Ygr054wp Drs1p Rna1p Ria1p Nsr1p Rpl16bp Rpl17ap Rps24ap Rps21ap Rpl31bp Nog1p Rpl19ap Prp20p Ubi3p |
Ribosomal proteins (10) | 4.5E-03 | Ubi3p Rps21ap Rpl17ap Rpl19ap Rpl31bp Drs1p Rps24ap Ygr054wp Nsr1p Rpl16bp | |
Translation (11) | 8.2E-08 | Egd1p Eft1p Cdc33p Ria1p Efb1p Hyp2p Pab1p Ygr054Wp Tif5p Tif3p Caf20p | |
Translation initiation (5) | 3.4E-04 | Hyp2p Ygr054Wp Tif3p Cdc33p Tif5p | |
Translation elongation (3) | 3.9E-03 | Ria1p Efb1p Eft1p | |
Protein fate (21) | Protein folding & stabilization (6) | 3.0E-03 | Cct5p Cct2p Ydj1p Zuo1p Sti1p Caj1p |
Protein w. binding function (26) | Protein binding proteins (13) | 9.1E-03 | Srv2p Scs2p Cct2p Hyp2p Zuo1p Sti1p Egd1p Pea2p Rvs167p Bbc1p Ssz1p Cct5p Abp1p |
RNA binding proteins (9) | 2.6E-03 | Pab1p Gbp2p Rpl16Bp Nop13p Scp160p Nsr1p Bfr1p Tma22p Arc1p | |
Cellular transport (19) | RNA transport (6) | 2.1E-03 | Rna1p Scp160p Prp20p Gbp2p Pab1p Arc1p |
Cellular communication (7) | Small GTPase mediated signal transduction (5) | 1.9E-03 | Srv2p Ras2p Zeo1p Cla4p Pea2p |
Cell rescue, defence and virulence (16) | Stress response (16) | 1.3E-03 | Cct5p Nsr1p Zuo1p Zeo1p Gbp2p Rhr2p Ssz1p Cct2p Ras2p Sod1p Rvs167p Yhb1p Stm1p Sti1p Egd1p Ydj1p |
Unfolded protein response (6) | 6.5E-04 | Sti1p Ssz1p Cct5p Egd1p Zuo1p Cct2p |
Enriched (p<0.01) functional categories among down-regulated proteins in encapsulated yeast, as analysed using the MIPS functional category enrichment tool (FUNCAT,
Category | Sub-category | p value | Proteins |
Metabolism (32) | Metabolism of glutamine (2) | 4.9E-03 | Gln1p Fas1p |
Phosphate metabolism (21) | 2.2E-05 | Rix7p Tpk1p Ugp1p Hsp78p Cka1p Vph1p Ypk1p Pex6p Ssa1p Pro1p Tpk2p Rli1p Hxk1p Ssb2p Tps2p Glc7p Glk1p Stt4p Rpt5p His2p Hsp104p | |
C-compound and carbohydrate metabolism (35) | 2.1E-12 | Glk1p Emi2p Adh5p Pmt7p Hsp12p Ach1p Ayr1p Gsy2p Ybr056Wp Glc3p Tps1p Adh1p Gph1p Gre3p Gln1p Tdh1p Mal62p Tsl1p Dld2p Gnd1p Tps2p Hxk1p Kgd1p Ald4p Ugp1p Dpm1p Nth1p Uga1p Ynr071Cp Gdb1p Dsf1p Gsy1p Pgm2p Glc7p Tal1p | |
Sugar, glucoside, polyol and carboxylate metabolism (10) | 2.1E-06 | Tps1p Tdh1p Pgm2p Kgd1p Nth1p Tsl1p Mal62p Gre3p Tal1p Ugp1p | |
Sugar, glucoside, polyol and carboxylate anabolism (7) | 2.8E-06 | Nth1p Ugp1p Tsl1p Tps1p Mal62p Tal1p Pgm2p | |
Sugar, glucoside, polyol and carboxylate catabolism (9) | 1.3E-05 | Nth1p Mal62p Kgd1p Tal1p Tps1p Tdh1p Pgm2p Gre3p Ugp1p | |
Polysaccharide metabolism (7) | 1.8E-04 | Glc3p Gln1p Gdb1p Gsy1p Gph1p Dpm1p Gsy2p | |
Glycogen metabolism (2) | 3.3E-03 | Gsy1p Gsy2p | |
Glycogen anabolism (2) | 3.3E-03 | Gsy1p Gsy2p | |
Lipid, fatty acid and isoprenoid metabolism (22) | 8.7E-09 | Stt4p Erg13p Dpm1p Erg11p Erg3p Cat2p Pdx3p Scs3p Fas1p Slc1p Fas2p Mcr1p Erg25p Ayr1p Ypk1p Ach1p Hsp12p Yml131Wp Ole1p Ura8p Fas3p Mrs6p | |
Fatty acid metabolism (4) | 9.0E-04 | Fas1p Fas2p Fas3p Ole1p | |
Tetracyclic and pentacyclic triterpenes metabolism (5) | 4.8E-04 | Erg25p Erg3p Erg11p Erg13p Mcr1p | |
Energy (27) | Pentose phosphate pathway (4) | 9.0E-04 | Gnd1p Tal1p Ynr034Wp Pgm2p |
Alcohol fermentation (3) | 1.6E-03 | Adh1p Adh5p Ald4p | |
Metabolism of energy reserves (13) | 1.5E-11 | Nth1p Tps2p Gsy1p Gph1p Glc7p Gdb1p Gsy2p Mal62p Ugp1p Tsl1p Glc3p Tps1p Pgm2p | |
Protein w. binding function (31) | Nucleotide/nucleoside/nucleobase binding (12) | 9.0E-04 | Ssa1p Rix7p Tpk2p Hnt1p Tpk1p Hsp104p Pex6p Hsp78p Rpt5p Rli1p Ssb2p Lap3p |
Cyclic nucleotide binding (2) | 2.0E-03 | Tpk1p Tpk2p | |
FAD/FMN binding | 8.9E-03 | Mcr1p Oye3p | |
Cellular transport (24) | Electron transport (7) | 8.3E-04 | Oye3p Mcr1p Vma2p Vma13p Atp7p Vma6p Vph1p |
Transport ATPases (5) | 3.1E-03 | Vma6p Vma2p Vph1p Atp7p Vma13p | |
Cell rescue, defence and virulence (24) | Stress response (20) | 2.0E-04 | Sip18p Tps1p Ssb2p Tps2p Mcr1p Hsp26p Mdj1p Hsp78p Pst2p Tsl1p Nth1p Hsp12p Hsp104p Pep4p Def1p Glc7p Ssa1p Aip1p Cka1p Gre3p |
Heat shock response (3) | 5.7E-03 | Glc7p Hsp12p Gre3p | |
Unfolded protein response (5) | 8.9E-03 | Ssa1p Mdj1p Hsp78p Hsp26p Ssb2p | |
Interaction with the environment (11) | Homeostasis of protons (5) | 1.8E-03 | Vma2p Vma13p Vma6p Atp7p Vph1p |
Enriched (p<0.01) functional categories among up-regulated proteins in encapsulated yeast, as analysed using the MIPS functional category enrichment tool (FUNCAT,
Of utmost importance for the production of bioethanol are the metabolism and energy turnover of the yeast. The differences in metabolite yields between the two modes of cultivation indicate that the metabolism of the yeast changed significantly upon encapsulation (
The central carbon metabolism is presented with up-regulated proteins with fold changes (encapsulated cells compared to free cells) in green, down-regulated proteins with fold changes in red and unaffected proteins with the measured fold changes in grey. The first number represents the fold change obtained by n-LC-MS/MS. Where applicable, this is followed by the average fold change obtained from 2-D DIGE spots with unique significant protein hits, and the average fold change obtained from spots with significant hits for co-migrating proteins in 2-D DIGE.
The two most up-regulated proteins in encapsulated cells were the high affinity hexose transporters Hxt6p (10.63) and Hxt7p (13.04). The expression of these transporters is repressed by high glucose levels and they have high expression levels on non-fermentable carbon sources and at low concentrations of glucose
Another enzyme in the glycolytic pathway, Tdh1p catalysing the oxidative phosphorylation of glyceraldehyde-3-phosphate to 1,3-bisphosphoglycerate, is known to be expressed during stationary phase and other conditions of slow growth, while the Tdh2p and Tdh3p are detected in the exponential phase
The cytoplasmic alcohol dehydrogenases Adh1p and Adh5p, and the glucose-repressed mitochondrial Adh3p, that all reduce acetaldehyde to ethanol, were up-regulated in the encapsulated cells (
The proteins involved in synthesis and utilization of the storage carbohydrates trehalose and glycogen were up-regulated (
In addition to carbon starvation, it is possible that the cells in the middle of the capsules also experienced nitrogen and phosphate limitation, since both glutamine and phosphate metabolism were significantly up-regulated in the encapsulated cells (
Another variable likely to change with a change in nutrient availability is the growth, and hence, also the protein synthesis should be affected. Previous reports have stated that both the total protein levels and the total RNA levels (with ribosomal RNA as the main contributor) decrease in prolonged growth of encapsulated yeast
In the PRIMA analysis of the genes coding for regulated proteins, the usage of transcription factor Abf1p was enriched among the down-regulated proteins (
Certain proteins in the unfolded protein response (UPR) related to growth were also down-regulated, e.g. Ssz1p (−1.49) and Zuo1p (−1.34) that act together in a complex involved in ribosome biogenesis
Considering the seemingly starvation-stressed cells, the expression of stress response proteins are of particular interest to study. Of the 91 proteins in the data set identified as stress response proteins according to the FunCat analysis, 21 were up-regulated, while 16 were down-regulated. Notable is that proteins under the control of the transcription factors Msn2 and Msn4 were enriched in the encapsulated yeast (
One of the major benefits of using encapsulated yeast in 2nd generation bioethanol production is its ability to tolerate otherwise too toxic dilute-acid hydrolysates. The tolerance of encapsulated yeast towards high levels of the pentose-derived furan aldehyde inhibitor furfural has been specifically studied
Lin et al.
In addition to the nLC-MS/MS with TMT® labelling, relative quantification of proteins from cells encapsulated in alginate-chitosan gel membranes compared to free cells was carried out also with 2-D DIGE. In this approach 103 spots with differential protein expression (Anova p<0.05) were chosen for subsequent MALDI TOF/TOF analysis for protein identification. Proteins were identified in 100 spots, of which 93 had highly significant hits, with a total of 52 different proteins detected (
Correlation between the average ratios of 31 proteins (with single significant hits in spots on the gel) (A and B) and 33 co-migrating proteins (with two or more significant hits per spot on the gel, giving uncertainties in the quantification of each individual protein in the spot) (C and D) proteins obtained by 2-D DIGE (RDIGE) and nLC-MS/MS (RMS/MS). A and C, the ratios obtained by nLC-MS/MS divided by the mean ratios obtained by 2-D DIGE, for single significant hit spots and spots with co-migrating proteins respectively. Triangles indicate extremely up-regulated proteins (fold change >2.5) as measured by the nLC-MS/MS approach and squares indicate proteins showing different sign of the fold change in the two approaches. The proteins were sorted by increasing fold change values obtained by the 2-D DIGE approach and divided into three groups, depending on the expression according to 2-D DIGE. Proteins marked with * had invariant expression in the nLC-MS/MS approach, and those marked with “were up-regulated. Proteins in parentheses (Sam2p) had extremely large RSD among the replicates in nLC-MS/MS and missing values indicate that the protein was not detected in the nLC-MS/MS approach (Ssa2p, Rpl9bp, Rps0bp). B and D, correlation plots of the ratios obtained by DIGE (x-axis) against the ratios obtained by nLC-MS/MS (y-axis), for unique (B) and co-migrating (D) protein spots, respectively. Extremely up- or down-regulated proteins (triangles) as well as the three and eight proteins showing different expression with the two methods (squares) were excluded from the calculation of the correlation.
In 34 of the 93 spots with significant protein hits, more than one protein was identified with significance. Such co-migration
On average, the two different methods gave similar expression values and the combination of the methods is thus a good way of verifying proteomic results. A comparison of DIGE quantification with metabolic labelling quantification of spots picked from a gel showed correlations similar to what was observed in our study
A FunCat localization analysis of the proteins detected as single proteins in the spots on the gels showed that all of them localized to the cytoplasm, nucleus, mitochondria or vacuole. This showed a drawback of the 2-D DIGE approach when it comes to proteome-wide studies. Mainly abundant proteins are detected, such as the glycolytic ones that made up 30% of the proteins detected in the 2-D DIGE approach, and not those that are less abundant, such as those involved in e.g. transcription.
Comparative proteomics of free and encapsulated
(TIF)
(TIF)
(TIF)
(TIF)
Protein sample preparation, nLC-MS/MS and 2-D DIGE methods.
(DOCX)
Proteins identified by the nLC-MS/MS approach.
(XLSX)
Functional classification and cellular localization of proteins identified by the nLC-MS/MS approach.
(XLSX)
Proteins identified by the 2-D DIGE approach.
(XLSX)
We acknowledge Carina Sihlbohm and Jörgen Bergström at The Proteomics Core Facility at Sahlgrenska Academy, University of Gothenburg, for performing the 2-D DIGE and nLC-MS/MS analysis.