Accepted Manuscript

Analysis of proteomic responses of freeze-dried Oenococcus oeni to access the molecular mechanism of acid acclimation on cell freeze-drying resistance

Kun Yang, Yang Zhu, Yiman Qi, Tingjing Zhang, Miaomiao Liu, Jie Zhang, Xinyuan Wei, Mingtao Fan, Guoqiang Zhang

PII: S0308-8146(19)30188-8 DOI: https://doi.org/10.1016/j.foodchem.2019.01.120 Reference: FOCH 24215

To appear in: Food Chemistry

Received Date: 2 October 2018 Revised Date: 24 December 2018 Accepted Date: 17 January 2019

Please cite this article as: Yang, K., Zhu, Y., Qi, Y., Zhang, T., Liu, M., Zhang, J., Wei, X., Fan, M., Zhang, G., Analysis of proteomic responses of freeze-dried Oenococcus oeni to access the molecular mechanism of acid acclimation on cell freeze-drying resistance, Food Chemistry (2019), doi: https://doi.org/10.1016/j.foodchem. 2019.01.120

This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Analysis of proteomic responses of freeze-dried Oenococcus oeni to access the molecular

mechanism of acid acclimation on cell freeze-drying resistance

Kun Yang1,2, Yang Zhu3, Yiman Qi2,Tingjing Zhang4, Miaomiao Liu2, Jie Zhang2, Xinyuan Wei2, Mingtao Fan2,*,

Guoqiang Zhang1,*

1 College of Biological and Chemical Engineering, Anhui Polytechnic University, Wuhu, 241000, China

2 College of Food Science and Engineering, Northwest A ＆ F University, Yangling, 712100, China

3 School of Agriculture and Food Sciences, University of Queensland, QLD, 4046, Australia

4 College of Food Science and Technology, Henan University of Technology, Zhenzhou, 450001, China

* Corresponding author: 1. Tel.: +86-13618940517; E-Mail: [email protected] (Guoqiang Zhang).

* Co-Corresponding author: 2. Tel.: +86-13892877726; E-Mail: [email protected] (Mingtao Fan)

1

Abstract

Malolactic (MLF), usually induced by Oenococcus oeni (O. oeni), is an important process to improve wine quality. Acid acclimation has been proven to be useful for enhancing the viability of lyophilized O. oeni. To explain the involved mechanisms, cell integrity, morphology and protein patterns of lyophilized O. oeni SD-2a were investigated with acid acclimation. After lyophilization, improvement of cell integrity and more extracellular polymeric substances (EPS) were observed in acid acclimated cells. Combined with GO and KEGG analysis, different abundant proteins were noticeably enriched in the process, especially amino sugar and nucleotide sugar metabolism. The most significant result was the over-expression of proteins participating in cell wall biosynthesis, EPS production, ATP binding and the bacterial secretion system. This result indicated the important role of acid acclimation on cell envelope properties. In addition, protein response to stress and arginine deiminase pathway were also proven to be over- expressed.

Keywords: Oenococcus oeni; Acid acclimation; Freeze-drying resistance; Cross-protection;

Comparative proteomics.

1. Introduction

During wine making, the alcoholic fermentation (AF) by yeast and the following process of malolactic fermentation (MLF) are usually promoted with starter cultures to make sure a controlled and efficient process. MLF is crucial for wine making, as it deacidifies wine by the conversion of malate to lactate, prevents the growth of spoilage microbes and contributes wine flavor. Generally,

MLF is mainly driven by O. oeni, though members of other lactic acid bacteria (LAB) such as 2

Lactobacillus and Leuconostoc are also found in wine and may contribute to wine quality.

Nonetheless, the growth of O. oeni is often inhibited in wine, leading to the delay of MLF. This failure was usually attributed to two aspects. Firstly, the growth of LAB for MLF was inhibited in wine conditions owing to its harsh environment, such as high content of ethanol, low pH, cold and

SO2 concentrations (Spano & Massa, 2006). Based on this, strains with better tolerance performance in wine could be regarded as the better candidate of MLF starters. Secondly, strains chosen for industrial application should withstand preservation processes (freeze-drying, freezing, or spray-drying). However, these processes might generate cell membrane damage, protein and

DNA denaturation, thereby significantly reducing cell viability (Santivarangkna, Kulozik, & Foerst,

2008). Taking these into consideration, the improvement of preservation processes plays an important role in MLF performance.

The low pH (< 3.5) in wines is one of the main parameters that decreases LAB viability. To overcome this extreme condition, wine LAB have generated several mechanisms to escape or tolerate acid stress: (a) alteration of membrane fatty acid composition and modification of membrane fluidity (Grandvalet et al., 2008); (b) induction of stress protein synthesis (Liu et al.,

2017); (c) activation of the proton extruding ATPase to keep proton motive force (PMF) maintenance and pH homeostasis (Tourdot-Marechal, Fortier, Guzzo, Lee, & Divies, 1999). More and more studies reported that pre-adaptation by treating strains to a lethal or sub-lethal stress for a limited time could generate cross-protection (Wang, Cui, & Qu, 2018). Correspondingly, acid pretreatment was proven to be useful for improving the viability of freeze-dried O. oeni SD-2a in our previous study (Zhang et al., 2013).

Currently, proteomics is often applied to clarify molecular mechanisms related to stress responses. Briz-Cid et al. (2016) reported the proteome changes of Garnacha Tintorera red grapes 3 during harvest drying and indicated that the changes are important for the quality of sweet wine.

As for O. oeni, most proteome studies have focused on its adaptation to wine stress conditions

(Costantini et al., 2015; Margalef-Catala, Araque, Bordons, Reguant, & Bautista-Gallego, 2016), and there was little proteome research concerning acid stress. Though adaptation of O. oeni to acid stress has been researched via transcriptome analysis and traditional methods (Liu et al., 2017), mechanisms of its effects on freeze-drying resistance are yet to be well understood. Therefore, cell integrity and cell morphology of freeze-dried O. oeni SD-2a were evaluated by flow cytometry

(FCM) and scanning electron microscope (SEM) in the present study. Furthermore, two- dimensional gel electrophoresis (2-DE) was used to compare the protein patterns of freeze-dried O. oeni SD-2a to provide molecular information of acid adaptation treatment on the cell lyophilized resistance.

2. Materials and methods

2.1. Bacterial culture and growth conditions

MLF strain O. oeni SD-2a (Patent number, 02123444.2), originally isolated from wine in

Shandong province (China), was selected as a MLF starter due to its good performance in deacidification during MLF (Liu et al., 2017). The cultivation of O. oeni SD-2a was performed in acidic tomato broth (ATB medium) at pH4.8 for 80 h to attain the early stationary phase and the preparation of ATB medium was according to the method of Zhang et al. (2013).

2.2. Acclimation conditions and lyoprotection

Cells grown to the early stationary phase (80 h, 1×109 CFU/ml) were divided into four equal parts and washed with sterile saline (0.85%, m/v) twice. Three cell pellets were suspended in the same volume of acid acclimation ATB medium (pH3.2, 3.5 and 4.0 respectively) for 2 h, and the 4 pellet suspended in fresh ATB medium (pH4.8) as a control. After incubation at 25 °C for 2 h, O. oeni SD-2a were centrifuged at 3000×g for 10 min and washed with sterile saline (0.85%, m/v) twice. Then cell pellets were resuspended in 2.5% (m/v) monosodium glutamate solution (MSG solution) and homogenized slightly at room temperature (15 min) for freeze-drying.

2.3. Freeze-drying procedure and rehydration

The lyophilization procedure was conducted in accordance with the method of Zhang et al.

(2013). Aliquots of 10 ml cell suspension (MSG solution) in 50-ml sterile vials were frozen at

−20 °C overnight, then vacuum drying was performed in a freeze dryer (Labogene, Denmark) for

28 h. The condenser temperature was −50 °C and the chamber pressure was lower than 0.06 mbar during this process. After freeze-drying, ATB medium with the original volume was applied to rehydrate the lyophilized cells. They were mixed gently and incubated at room temperature for 15 min for protein extraction, FCM and SEM analysis.

2.4. Bacterial damage and cell surface properties

Bacterial damage was evaluated by FCM on the basis of the method of Pan et al. (2014).

Samples prepared previously were diluted according to the instruction of the LIVE/DEAD BacLight bacterial viability kit (Invitrogen, USA). Then cell suspensions were stained with 5 μM SYTO 9 green fluorescent nucleic acid dye and 30μM propidium iodide (PI) and incubated at 25 °C for 15 min in the dark. Cells without stain were used as background control and single stained samples with PI and SYTO9 respectively were used for instrument adjustment. Then bacterial cell damage assay was conducted on a FACS Calibur flow cytometer (BD, USA). Data were analyzed using the

FlowJo software and populations of live, damaged and dead cells were gated for analysis of different acid treatments.

Rehydrated cell morphology was detected using SEM (Hitachi, Ltd., Japan). Samples were 5 prepared according to the study described by Bastard et al. (2016). Briefly, rehydrated bacteria were fixed with 2.5% (v/v) glutaraldehyde for 3 h at room temperature and washed twice with phosphate buffer (pH7.0). Next, samples were dehydrated with a graded series of ethanol solutions (30%,

50%, 70%, 90%, 100% (v/v)) every 10 min. Finally, samples were dried for 4 h in a critical point dryer to obtain absolute dry samples, sputter-coated with gold and viewed by SEM.

2.5. Protein extraction, two-dimensional gel electrophoresis and image analysis

Whole-cell proteins were extracted for 2-DE from three independent experiments. Cell disruption and protein precipitation were performed according to the method of Yang et al. (2018).

Then protein pellets were dissolved in Tris buffer (pH10). To remove residual DNA in the protein pellet, the phenol/chloroform/isoamyl alcohol solution (25/24/1, v/v/v) was transferred into the protein solution and mixed thoroughly (Antonioli, Bachi, Fasoli, & Righetti, 2009). After that, protein aqueous solution was separated from the organic phases by centrifugation and precipitated again with cold acetone. Protein concentration was analyzed as reported by Yang et al. (2018).

Protein extracts (450 μg) were loaded onto 17 cm strips (pH linear range: 5‒8, Bio-Rad, USA), focused for 85 kV•h in an IPGphor (Bio-Rad, USA) at 20 °C, equilibrated with DTT and iodoacetamide, and then separated on a 12% (m/v) SDS acrylamide gel as reported previously

(Yang et al., 2018). Gels from three independent experiments were stained with the method of blue silver as reported by Candiano et al. (2004). Image analysis (normalization of spot intensities and changes in protein expression) was carried out on software PDQuest 2-DE 8.0.1 (Bio-Rad, USA).

The student’s t-test (at least 1.5 fold change of difference in abundance, P < 0.05) was used to detect significant differences of protein abundance between lyophilized O. oeni SD-2a with acid acclimated (AC) and non-acid acclimated (NAC) pretreatments.

2.6 Mass spectrometry analysis and bioinformatics analysis 6

DAPs obtained from gels were washed and digested according to the method described by

Shevchenko et al. (2006). The peptide mix was analyzed on AB SCIEX 4800 plus MALDI

TOF/TOF (Applied Biosystems, USA) as reported by Yang et al. (2018). MS/MS mass spectra of

DAPs were conducted on TOF/TOF Explorer™ Software and searched against the NCBI databases of O. oeni with the search engine MASCOT (Matrix Science, USA). The peptide MS/MS search was conducted with the following settings: taxonomy, bacteria; mass tolerance for peptides, 100 ppm; mass tolerance for TOF/TOF fragments, 0.5 Da; fixed modification, cysteine carbamidomethylation; and variable modification, methionine oxidation. Identification with the

Mascot score > 30 was considered as significant.

The function of DAPs was elucidated using the Uniprot database (http://www.uniprot.org).

Then prediction of locations was carried out using Cell-PLoc as described by Chou and Shen

(2008). In order to determine the protein expression level, hierarchical clustering was presented by

Multiple Experiment Viewer 4.9. DAPs were organized into different GO categories on the basis of their function information in the DAVID toolkit. Moreover, DAPs were searched against the KEGG database to enrich pathway analysis, and protein networks combined with KEGG analysis were performed to better understand protein interaction and metabolic processes.

3. Results

3.1. Comparison of cell damage in freeze-dried AC and NAC O. oeni SD-2a

Cell damage of lyophilized O. oeni SD-2a was detected by FCM as shown in Figure 1, the PI and SYTO9 double-staining provided three distinct gates (live, damaged and dead cells) to distinguish intact bacteria from membrane-compromised ones. FCM results (Figure 1, B‒D) showed that percentages of intact cells in freeze-dried AC cells (pH4.0, pH3.5, pH3.2) were 68.1%, 7

70.7% and 74.8% respectively, which were increased with the enhancement of acid stress.

Relatively, freeze-dried NAC cells showed the lowest proportion (62.2%) of live cells, indicating that acid acclimation was helpful in maintaining the cell membrane integrity of lyophilized O. oeni

SD-2a.

3.2. Cell morphology of freeze-dried AC and NAC O. oeni SD-2a

Cell surface characteristics were assessed using SEM on freeze-dried O. oeni SD- 2a after rehydration (Figure 2). Images of freeze-dried NAC and AC cells showed that there were some differences in surface and morphology properties. Extracellular polymeric substances (EPS), the linking structure between cells, could be clearly observed on the surface of freeze-dried AC cells marked in Figure 2 F‒H, whereas the freeze-dried NAC cells (Figure 2 E) had a relatively smooth phenotype with little linking structures. As is well known, O. oeni is usually organized in chains (2‒

8) or pairs with a relatively smooth surface, but freeze-dried O. oeni SD -2a performed cell aggregation with linking structures. Besides, freeze-dried AC cells presented a much larger mass in comparison with freeze-dried NAC cells (Figure 2 A‒D).

3.3. Differential abundance of proteins between freeze-dried AC and NAC O. oeni SD-2a

In order to study the relationship between freeze-dried O. oeni SD-2a proteome and acid acclimation pretreatments, only remarkably expressed main proteins with acid pretreatments were emphasized on representative 2-DE maps as shown in Figure 3 A. Differential protein expression was assessed according to the following comparisons (a) pH4.0/ pH4.8, (b) pH3.5/ pH4.8 and (c) pH3.2/ pH4.8: freeze-dried cells with acid acclimation at (a) pH4.0, (b) pH3.5 and (c) pH3.2 for 2 h vs. freeze-dried cells without acid acclimation at pH4.8 for 2 h. In order to analyze 2-DE images, fold-change of 1.5 was selected for consideration as the minimum level of differentially expressed protein. A total of 83 protein spots were analyzed as statistically significant (P-value < 0.05) as 8 shown in Supplementary Table S1 and 82 protein spots were successfully identified using MALDI-

TOF/TOF-MS-based analysis.

The distribution of DAPs was performed as shown in Figure 4 (A‒C) and Venn diagram analysis (D and E) was used to understand DAPs profile patterns with different acid treatments.

There were 52 (31 up- and 21 down-regulated proteins), 50 (36 up- and 14 down-regulated proteins) and 52 (36 up- and 16 down-regulated proteins) DAPs in comparisons of pH4.0/pH4.8, pH3.5/pH4.8 and pH3.2/pH4.8 respectively. Meanwhile, there were 13 spots over-expressed at all acid acclimation treatments, and 6 spots were down-regulated with all acid pretreatments (Figure 4

D, E). Hierarchical clustering analysis of DAPs was performed to better visualize the protein abundance profiles among freeze-dried AC and NAC O. oeni SD-2a (Supplementary Figure S1).

The result generated a pattern which consisted of two major clusters and DAPs involved in these three comparisons displayed similar color distributions except some specific proteins such as peptidase C69 in the comparison of pH4.0/pH4.8, aryl-alcohol dehydrogenase in the comparison of pH3.5/pH4.8 and 3-oxoacyl-acyl carrier protein reductase and diacetyl reductase in the comparison of pH3.2/pH4.8.

3.4. Functional annotation and categories of DAPs

All DAPs were analyzed via the NCBI BLAST database and the function of identified protein was explicated by inputting the Uniprot accession number into the Uniprot database as shown in

Supplementary Table S2. The identified proteins were mainly involved in (a) carbon metabolism,

(b) amino-acid metabolism and transportation, (c) nucleotide metabolism and transportation, (d) genetic information processing, (e) lipid biosynthetic process, (f) oxidative phosphorylation, (g) oxidoreductase activity, (h) stress response and (i) quorum sensing and signal peptide processing.

Bioinformatics approaches were used to extract information of DAPs relevant to biological 9 interpretation. GO terms (Supplementary Table S4) were categorized into groups based on biological process (BP), cellular component (CC) and molecular function (MF) and the top significantly enriched terms in different gene ontology hierarchies were shown in Supplementary

Figure S2. Carbohydrate metabolic process and carbohydrate derivative metabolic process were the most significantly enriched terms in BP analysis (level 4 gene ontology hierarchy) in comparisons of pH4.0/pH4.8, pH3.5/pH4.8 and pH3.2/pH4.8. In CC analysis, most annotated proteins in all comparisons (pH4.0/pH4.8, pH3.5/pH4.8 and pH3.2/pH4.8) were predicted to be located in cytoplasm. Besides, coenzyme binding and oxidoreductase activity (level 4 gene ontology hierarchy) were the dominant molecular function in comparisons of pH4.0/pH4.8 and pH3.5/pH4.8.

3.5. KEGG pathway analysis

For DAPs, KEGG analysis presented some different biological functional properties in

Supplementary Figure S3 (Supplementary Table S5) and revealed that the enriched metabolism pathways included , biosynthesis of other secondary metabolites, amino acid metabolism, as well as biosynthesis of antibiotics. Notably, enrichment of amino sugar and nucleotide sugar metabolism in carbohydrate metabolism became increasingly significant in comparisons of pH4.0/pH4.8, pH3.5/pH4.8 and pH3.2/pH4.8. Correspondingly, carbohydrate metabolism processes and carbohydrate derivative metabolism processes were also significantly enriched in BP analysis, indicating that these processes responded actively to acid pretreatments in lyophilized O. oeni SD-2a.

Protein-protein interaction networks (Figure 5) were constructed to better understand the relationship between the proteins and enriched pathways. In comparisons of pH4.0/pH4.8, pH3.5/pH4.8 and pH3.2/pH4.8, 36, 39 and 38 DAPs were included in the networks, while the rest of the proteins were found with no interactions in the STRING database. Proteins of 10 phosphoglucomutase (spot 81, OEOE_0367), -6-phosphate dehydrogenase (spot35, 68, zwf), glucosamine-6-phosphate deaminase (spot 31, nagB) and 30S ribosomal protein S6 (spot 25, rpsF) were the main connectors in the network of the pH4.0/pH4.8 comparison. ATP synthase subunit alpha (spot42, atpA), glutamine--6-phosphate aminotransferase (spot 36, OEOE_0635) and

30S ribosomal protein S2 (spot77, rpsB) were the top three proteins with the most connected proteins in the comparison of pH3.5/pH3.2. As for the comparison of pH3.2/pH4.8, main connectors in this network (Figure 6 C) were glutamine-fructose-6-phosphate aminotransferase

(spot 36, OEOE_0635), 30S ribosomal protein S2 (spot77, rpsB) and glucosamine-6-phosphate deaminase (spot 31, nagB). Moreover, amino sugar and nucleotide sugar metabolism was increasingly significantly enriched and glutathione metabolism was predicted to be less significantly enriched with the decrease of pH value.

4. Discussion

As an important microorganism associated with grapes and wines, O. oeni has adapted itself to the harsh environment with many adaptation mechanisms and exhibited its importance in the production of quality wines. During winemaking, dried starters of O. oeni are usually used to promote the completion of MLF and the viability of starters was considered to be crucial in this process. The O. oeni SD-2a strain was selected as a potential candidate for MLF starter in the study, as it showed effective abilities to survive in wine conditions and good performance in MLF capabilities. According to the previous study, O. oeni SD-2a was proven to be more resistant to freeze-drying stress with acid acclimation treatments (Zhang et al., 2013). In order to understand the mechanisms involved in acid acclimation on freeze-drying resistance, this study detected the damage properties and surface characteristics of O. oeni acclimated in modified ATB broth with pH 11 from 3.2 to 4.8 after freeze-drying and identified the protein patterns of lyophilized O. oeni SD-2a involved in the acid pretreatments.

As for FCM, the combination of fluorescent dyes (SYTO9 and PI) has often been used to monitor the structural integrity of bacteria in food industry. SYTO9 stains nucleic acids in all cells, whereas the other one selectively labels nucleic acids in membrane-compromised cells. As reported previously, Bensch et al. (2014) evaluated the viability of Lactobacillus plantarum produced by fluidized bed-drying with FCM. Similarly, FCM was used in the present study to detect the membrane damage of freeze-dried O. oeni SD-2a. The significant finding (Figure 1) was that acid acclimation before freeze-drying improved cell membrane integrity of lyophilized O. oeni and treatment with pH3.2 showed the highest percent of intact cells. As mentioned in the report of

Zhang et al. (2013), freeze-dried O. oeni with acid stress pretreatment also showed better performance in cell survival with plate counting. However, lyophilized O. oeni with treatment of pH3.5 presented the highest survival rate in the previous study. A possible explanation for this might be that plate counting could not provide the number of viable-but-nonculturable cells (Bensch et al., 2014).

Apart from cell integrity of freeze-dried O. oeni, cell surface morphology (Figure 2) was detected using SEM and more extracellular substances between freeze-dried cells were found with acid acclimation. Comparably, Dimopoulou et al. (2018) found that the production rate of exopolysaccharides was increased in the presence of single stress factors such as pH and ethanol.

Endogenous exopolysaccharides have been considered to be useful for producing freeze dried strains (Dimopoulou et al., 2016) and previous work has shown that proteins related to production and quorum sensing were over-expressed during the freeze-drying process (Yang et al., 2018). Taking these into account, EPS production was considered to be an 12 efficient way to protect cells from environmental stresses.

Obviously, improvement of cell integrity and production of EPS were direct responses of acid acclimation to freeze-drying resistance, while the underlying mechanism was not fully elucidated.

According to the result of GO enrichment (Supplementary Figure S2) and KEGG pathway analysis

(Supplementary Figure S3), proteins related to carbohydrate metabolism were highlighted with treatments of acid acclimation. Notably, amino sugar and nucleotide sugar metabolism was especially enriched as it became more significant with the lower of the pH values, and most involved proteins were related to UDP-sugar metabolism (Figure 5, Supplementary Table S5).

Based on the KEGG database, pathways of DAPs involved in carbohydrate metabolism pathways were phosphate pathway, , amino sugar and nucleotide sugar metabolism (UDP- sugar), dTDP sugar metabolism and butanoate metabolism as shown in Figure 6. Eleven of these proteins were related to UDP-sugar metabolism and the corresponding metabolites were UDP- glucose, UDP-glucuronate and UDP- et al. Combined with SEM results, we considered that these results could support the ideas of previous studies (Dimopoulou et al., 2014; Boels, van

Kranenburg, Hugenholtz, Kleerebezem, & de Vos, 2001), in which EPS could be generated from intracellular sugar nucleotides.

As described in the review of production in LAB (Zeidan et al., 2017), the

Wzy-dependent pathway represented the synthesis of gram-positive bacteria polysaccharides including two distinct steps, one of them was considered to be the production of activated sugar precursors (sugar nucleotides). This step always starts from the glycolytic intermediates including glucose-6-phosphate (glucose-6p) and fructose-6-phosphate (fructose-6p). Consistent with the literature, the present research found the over-expression of glucokinase (spot38, OEOE_0920) and glucose-6p isomerase (spot52, OEOE_0636), which were corresponded to the production of 13 glucose-6p and fructose-6p respectively. Then different sugar nucleotides were produced dictated by particular proteins. As for the branch of glucose-6p (Figure 6), phosphoglucomutase (spot81,

OEOE_0367) catalyzed the conversion of glucose-6p to glucose-1-phosphate (glucose-1p) and changed the carbon flux into the UDP-glucose synthetic pathway. Then glucose-1p uridylyltransferase (spot56, OEOE_0565) transferred glucose-1p into UDP-glucose, which subsequently could be converted into UDP-glucuronate by UDP-glucose dehydrogenase (spot61,

OEOE_1737) or UDP-galactose by UDP-glucose 4-epimerase (spot17, OEOE_1402) as shown in

Figure 6. Levander et al. (2002) found that over-expression of glucose-1-phosphate uridylyltransferase in combination with phosphoglucomutase in Streptococcus thermophilus LY03

(S. thermophiles LY03) resulted in higher production levels of polysaccharide in comparison with the original strain. In the present study, these two proteins were over-expressed with acid acclimation. Besides, UDP-glucose dehydrogenase (spot61, OEOE_1737), responsible for the synthesis of UDP-glucuronic acid, was also found to be over-expressed in this study and UDP- glucuronic acid was considered to be the source of glucuronic acid for biosynthesis of many exopolysaccharides (Granja, Popescu, Marques, Sa-Correia, & Fialho, 2007). In spite of the up- regulation of phosphoglucomutase, glucose-1p uridylyltransferase and UDP-glucose dehydrogenase, UDP-glucose 4-epimerase (spot17, OEOE_1402) and alpha-galactosidase (spot59,

OEOE_1781) involved in UDP-galactose metabolism were found to be down-regulated with most treatments in this study. A possible explanation for this might be that only 5‒10% O. oeni strains could consume galactose and the gene encoding proteins to export galactose was not identified until now (Cibrario, Peanne, Lailheugue, Campbell-Sills, & Dols-Lafargue, 2016).

Alternatively, glucose-1P could be converted into dTDP-4-oxo-6-deoxy-glucose and Dtdp-4- oxo-6-deoxy- by dTDP-glucose 4, 6-dehydratase (spot79, OENOO_59030) and dTDP-4- 14 dehydrorhamnose 3, 5-epimerase (spot63, OEOE_1448), respectively. These reactions were involved in the formation of dTDP-L-, which is used for the synthesis of cell wall linker

(Sen et al., 2011). The current study found that these two proteins were up-regulated, which was in agreement with earlier observations (Budin-Verneuil, Pichereau, Auffray, Ehrlich, & Maguin,

2007). It has been considered that dTDP-glucose 4, 6-dehydratase and dTDP-4-dehydrorhamnose 3,

5-epimerase related to cell envelope composition might be the key factors for intrinsic acid tolerance in Lactobacillus plantarum.

The second glycolytic intermediate, fructose-6p, could be converted from glucose-6p, fructose and mannose by glucose-6-phosphate isomerase (spot52, OEOE_0636), fructokinase (spot13,

OEOE_1708) and mannose-6-phosphate isomerase (spot53, OEOE_0249), respectively.

Subsequently, N-acetylglucosamine (GlcNAc) and UDP-GlcNAc could be formed from fructose-6p via amino sugar metabolism with the activities of glutamine-fructose-6-phosphate aminotransferase

(spot36, OEOE_0635, GlmS), glucosamine-6-phosphate deaminase (spot31, OEOE_0644, NagB),

GlcNAc-6-phosphate deacetylase (spot43, OEOE_1728) and UDP-GlcNAc-1- carboxyvinyltransferase (spot47, OEOE_1785). According to studies of Komatsuzawa et al. (2004) and Kawada-Matsuo et al. (2012), proteins GlmS and NagB played important roles in the allocation of sugars into the glycolysis pathway or amino sugar into cell envelope biosynthesis pathways, and the regulation directions of glmS and nagB was opposite in sugar metabolism in Streptococcus mutans. However, GlmS (spot36, OEOE_0635) and NagB (spot31, OEOE_0644) were both down- regulated with acid acclimation in this study. This result might be explained by the fact that freeze- drying conditions for O. oeni SD-2a were not supplemented with any sugars. As for GlcNAc-6- phosphate deacetylase, it converts GlcNAc-6p into GlcN-6p, which is the critical enzyme in production of amino-sugar precursors and influences the utilization of cell wall peptidoglycan 15 fragments (Park & Uehara, 2008). UDP-GlcNAc-1-carboxyvinyltransferase (spot47, OEOE_1785) was also found to be over-expressed. This enzyme could mediate the transfer of an enolpyruvate moiety from phosphoenolpyruvate (PEP) to UDP-GlcNAc to produce UDP-GlcNAc-enolpyruvate, which is part of peptidoglycan biosynthesis (Barreteau et al., 2008).

Additionally, the abundance of seven proteins involved in the pentose phosphate pathway was found to be changed with acid acclimation (Figure 6). Among the DAPs, some spots were predicted to be the same protein such as glucose-6-phosphate dehydrogenase (spot35 and spot68,

OEOE_0135) and phosphogluconate dehydrogenase (spot71 and spot76, OEOE_0892). As the separation of proteins by the 2-DE method was based on their isoelectric point (pI) and molecular weight (MW), it was difficult to avoid the existence of isoforms and post-translational modification

(Curreem, Watt, Lau, & Woo, 2012). In pentose phosphate pathway, one interesting finding is the up-regulation of -phosphate pyrophosphokinase (spot82, OEOE_0160), which transferred ribose-5p into phosphoribosyl pyrophosphate (PRPP). PRPP plays a central role in many biosynthetic pathways, including the biosynthesis of purine and pyrimidine nucleotides, amino acids (histidine and tryptophan) and co-factors (NAD and NADP). Ugbogu et al. (2016) found the impairment of cell wall integrity phenotype in the ribose-phosphate pyrophosphokinase (Prs) gene mutated or deleted yeast strains, and supposed the instrumental function of Prs polypeptides on yeast cell with stress. Among proteins involved in carbohydrate metabolism, lactate dehydrogenase

(spot80, OEOE_0413) converted pyruvate into lactate and was highly over-expressed as shown in

Figure 6. Similarly, lactate dehydrogenase was found to be up-regulated in Bifidobacterium longum with acid stress (Sanchez et al., 2007).

Apart from proteins involved in carbohydrate metabolism, there were other specific DAPs, which might be involved in the acid acclimation response to freeze-drying stress. Firstly, proteins related to 16 the arginine deiminase pathway (ADI pathway): arginine ABC transporter ATP-binding protein

(spot58, OENOO_41008) and ornithine carbamoyltransferase (spot20, OEOE_0258) were detected to be over-expressed with all acid acclimation treatments. Arginine ABC transporter ATP-binding protein, a transporter with high affinities for arginine, mediated the uptake of amino acids (Fleischer,

Wengner, Scheffel, Landmesser, & Schneider, 2005). Ornithine carbamoyltransferase has been proven to be pivotal for biosynthesis and catabolism of arginine in Streptococcus gordonii

(Jakubovics et al., 2015). Moreover, arginine could be degraded through the ADI pathway with the product of ammonia, which could help to neutralize the cytoplasm and improve cell acid resistance

(Papadimitriou et al., 2016). Secondly, most proteins related to ATP-binding were detected to be over- expressed, including ATP synthase subunit alpha (spot42, OEOE_0663), ATP-dependent Clp protease

(spot6, OEOE_0514), ATP-dependent Clp protease ATP-binding subunit (spot46, OEOE_0572), Clp protease ClpX (spot51, OEOE_0640) and chaperone protein DnaK (spot11, OEOE_1309). ATP synthase was associated with MLF. In our previous study, ATP synthase subunit alpha was also over- expressed under freeze-drying conditions (Yang et al., 2018). Besides, Liu et al. (2017) observed that the gene of ATP synthase subunit alpha was up-regulated with acid treatment. These indicated that

ATP synthase was the critical response to acid stress and freeze-drying process in O. oeni SD-2a. As for chaperone protein and Clp proteases, they were also over-expressed with freeze-drying treatment

(Yang et al., 2018). As reviewed by Papadimitriou et al. (2016), chaperones and proteases would be quickly induced under stress conditions to combat the potentially adverse aggregation of denatured proteins. Thirdly, proteins involved in DNA replication were over-expressed. This was opposite to the results of previous study (Yang et al., 2018), in which all proteins involved in DNA replication were inhibited with freeze-drying treatment. Taking these into consideration, we supposed that acid acclimation was helpful for DNA replication during the freeze-drying process. Finally, the expression 17 level of protein translocase subunit SecA involved in quorum sensing and the bacterial secretion system was found to be increased with acid acclimation treatment. According to the review of Du

Plessis et al. (2011), SecA played an important role in protein translation as well as translocation of hydrophilic domains of membrane proteins across the membrane. In addition, Bauer and Rapoport

(2009) supposed that SecA served as an ATP-driven molecular motor to drive the stepwise translocation of polypeptide chains across the membrane. As EPS were composed of polypeptide, polysaccharide and nucleic acids, it could thus be suggested that SecA was useful for EPS production.

5. Conclusion

In this research, FCM and SEM data corroborate the notion that the acid acclimation was beneficial for improving cell survival through modification of cell integrity and production of EPS.

Proteome analysis combined with bioinformatics confirmed the importance of carbohydrate metabolism during the freeze-drying process with acid acclimation pretreatments in O. oeni SD-2a, and highlighted the important role of amino sugar and nucleotide sugar metabolism during these processes. The up-regulation of proteins involved in production of sugar precursors for polysaccharides, cell wall biosynthesis, ADI pathway, ATP binding and the bacterial secretion system indicated that acid acclimation might enhance the freeze-drying resistance of O. oeni SD-2a mainly through improving cell wall composition, EPS and energy production. In the light of these results, new insights of molecular mechanisms regulated by acid acclimation improved our understanding of the biological characteristics involved in freeze-dried O. oeni.

Conflicts of interest

Authors declare no conflicts of interest regarding this manuscript. 18

Acknowledgments

The project was supported by the Natural Science Foundation of China (Grant No.31260371,

31560441, 31760381)，the Natural Science Foundation of Tibet (Grant No. XZ2017ZRG-26) and the program-Comprehensive utilization of the by-products of horticultural crops processing (Grant

No.201503142-10).

References

Antonioli, P., Bachi, A., Fasoli, E., & Righetti, P. G. (2009). Efficient removal of DNA from proteomic samples prior to

two-dimensional map analysis. Journal of Chromatography A, 1216(17), 3606-3612.

doi:10.1016/j.chroma.2008.11.053

Barreteau, H., Kovac, A., Boniface, A., Sova, M., Gobec, S., & Blanot, D. (2008). Cytoplasmic steps of peptidoglycan

biosynthesis. Fems Microbiology Reviews, 32(2), 168-207. doi:10.1111/j.1574-6976.2008.00104.x

Bastard, A., Coelho, C., Briandet, R., Canette, A., Gougeon, R., Alexandre, H., . . . Weidmann, S. (2016). Effect of

Biofilm Formation by Oenococcus oeni on Malolactic Fermentation and the Release of Aromatic Compounds

in Wine. Frontiers in Microbiology, 7. doi:10.3389/fmicb.2016.00613

Bauer, B. W., & Rapoport, T. A. (2009). Mapping polypeptide interactions of the SecA ATPase during translocation.

Proceedings of the National Academy of Sciences of the United States of America, 106(49), 20800-20805.

doi:10.1073/pnas.0910550106

Bensch, G., Ruger, M., Wassermann, M., Weinholz, S., Reichl, U., & Cordes, C. (2014). Flow cytometric viability

assessment of lactic acid bacteria starter cultures produced by fluidized bed drying. Applied Microbiology and

Biotechnology, 98(11), 4897-4909. doi:10.1007/s00253-014-5592-z

Boels, I. C., van Kranenburg, R., Hugenholtz, J., Kleerebezem, M., & de Vos, W. M. (2001). Sugar catabolism and its 19

impact on the biosynthesis and engineering of exopolysaccharide production in lactic acid bacteria.

International Dairy Journal, 11(9), 723-732. doi:10.1016/s0958-6946(01)00116-9

Budin-Verneuil, A., Pichereau, V., Auffray, Y., Ehrlich, D., & Maguin, E. (2007). Proteome phenotyping of acid stress-

resistant mutants of Lactococcus lactis MG1363. Proteomics, 7(12), 2038-2046. doi:10.1002/pmic.200600773

Candiano, G., Bruschi, M., Musante, L., Santucci, L., Ghiggeri, G. M., Carnemolla, B., . . . Righetti, P. G. (2004). Blue

silver: A very sensitive colloidal Coomassie G-250 staining for proteome analysis. Electrophoresis, 25(9),

1327-1333. doi:10.1002/elps.200305844

Chou, K. C., & Shen, H. B. (2008). Cell-PLoc: a package of Web servers for predicting subcellular localization of

proteins in various organisms. Nature Protocols, 3(2), 153-162. doi:10.1038/nprot.2007.494

Cibrario, A., Peanne, C., Lailheugue, M., Campbell-Sills, H., & Dols-Lafargue, M. (2016). Carbohydrate metabolism in

Oenococcus oeni: a genomic insight. Bmc Genomics, 17. doi:10.1186/s12864-016-3338-2

Costantini, A., Rantsiou, K., Majumder, A., Jacobsen, S., Pessione, E., Svensson, B., . . . Cocolin, L. (2015).

Complementing DIGE proteomics and DNA subarray analyses to shed light on Oenococcus oeni adaptation to

ethanol in wine-simulated conditions. Journal of Proteomics, 123, 114-127. doi:10.1016/j.jprot.2015.04.019

Curreem, S. O. T., Watt, R. M., Lau, S. K. P., & Woo, P. C. Y. (2012). Two-dimensional gel electrophoresis in bacterial

proteomics. Protein & Cell, 3(5), 346-363. doi:10.1007/s13238-012-2034-5

Dimopoulou, M., Bardeau, T., Ramonet, P.-Y., Miot-Certier, C., Claisse, O., Doco, T., . . . Dols-Lafargue, M. (2016).

Exopolysaccharides produced by Oenococcus oeni: From genomic and phenotypic analysis to technological

valorization. Food Microbiology, 53, 10-17. doi:10.1016/j.fm.2015.07.011

Dimopoulou, M., Raffenne, J., Claisse, O., Miot-Sertier, C., Iturmendi, N., Moine, V., . . . Dols-Lafargue, M. (2018).

Oenococcus oeni Exopolysaccharide Biosynthesis,a Tool to Improve Malolactic Starter Performance. Frontiers

in Microbiology, 9. doi:10.3389/fmicb.2018.01276

Dimopoulou, M., Vuillemin, M., Campbell-Sills, H., Lucas, P. M., Ballestra, P., Miot-Sertier, C., . . . Dols-Lafargue, M. 20

(2014). Exopolysaccharide (EPS) Synthesis by Oenococcus oeni: From Genes to Phenotypes. Plos One, 9(6).

doi:10.1371/journal.pone.0098898

Du Plessis, D. J. F., Nouwen, N., & Driessen, A. J. M. (2011). The Sec translocase. Biochimica Et Biophysica Acta-

Biomembranes, 1808(3), 851-865. doi:10.1016/j.bbamem.2010.08.016

Fleischer, R., Wengner, A., Scheffel, F., Landmesser, H., & Schneider, E. (2005). Identification of a gene cluster

encoding an arginine ATP-binding-cassette transporter in the genome of the thermophilic Gram -positive

bacterium Geobacillus stearothermophilus strain DSMZ 13240. Microbiology-Sgm, 151, 835-840.

doi:10.1099/mic.0.27591-0

Grandvalet, C., Assad-Garcia, J. S., Chu-Ky, S., Tollot, M., Guzzo, J., Gresti, J., & Tourdot-Marechal, R. (2008).

Changes in membrane lipid composition in ethanol- and acid-adapted Oenococcus oeni cells: characterization

of the cfa gene by heterologous complementation. Microbiology-Sgm, 154, 2611-2619.

doi:10.1099/mic.0.2007/016238-0

Granja, A. T., Popescu, A., Marques, A. R., Sa-Correia, I., & Fialho, A. M. (2007). Biochemical characterization and

phylogenetic analysis of UDP-glucose dehydrogenase from the gellan gum producer Sphingomonas elodea

ATCC 31461. Applied Microbiology and Biotechnology, 76(6), 1319-1327. doi:10.1007/s00253-007-1112-8

Jakubovics, N. S., Robinson, J. C., Samarian, D. S., Kolderman, E., Yassin, S. A., Bettampadi, D., . . . Rickard, A. H.

(2015). Critical roles of arginine in growth and biofilm development by Streptococcus gordonii. Molecular

Microbiology, 97(2), 281-300. doi:10.1111/mmi.13023

Kawada-Matsuo, M., Mazda, Y., Oogai, Y., Kajiya, M., Kawai, T., Yamada, S., . . . Komatsuzawa, H. (2012). GlmS and

NagB regulate amino sugar metabolism in opposing directions and affect Streptococcus mutans virulence. Plos

One, 7(3). doi:10.1371/journal.pone.0033382

Komatsuzawa, H., Fujiwara, T., Nishi, H., Yamada, S., Ohara, M., McCallum, N., . . . Sugai, M. (2004). The gate

controlling cell wall synthesis in Staphylococcus aureus. Molecular Microbiology, 53(4), 1221-1231. 21

doi:10.1111/j.1365-2958.2004.04200.x

Levander, F., Svensson, M., & Radstrom, P. (2002). Enhanced exopolysaccharide production by metabolic engineering

of Streptococcus thermophilus. Applied and Environmental Microbiology, 68(2), 784-790.

doi:10.1128/aem.68.2.784-790.2002

Liu, L. X., Zhao, H. Y., Peng, S., Wang, T., Su, J., Liang, Y. Y., . . . Wang, H. (2017). Transcriptomic analysis of

Oenococcus oeni SD-2a response to acid shock by RNA-Seq. Frontiers in Microbiology, 8.

doi:10.3389/fmicb.2017.01586

Briz-Cid, N., Pose-Juan, E., Rial-Otero, R., & Simal-Gándara, Jesús. (2016). Proteome changes in garnacha tintorera

red grapes during post-harvest drying. LWT - Food Science and Technology, 69, 608-613.

doi:10.1016/j.lwt.2016.02.026

Margalef-Catala, M., Araque, I., Bordons, A., Reguant, C., & Bautista-Gallego, J. (2016). Transcriptomic and Proteomic

Analysis of Oenococcus oeni Adaptation to Wine Stress Conditions. Frontiers in Microbiology, 7.

doi:10.3389/fmicb.2016.01554

Pan, H., Zhang, Y., He, G.-X., Katagori, N., & Chen, H. (2014). A comparison of conventional methods for the

quantification of bacterial cells after exposure to metal oxide nanoparticles. Bmc Microbiology, 14.

doi:10.1186/s12866-014-0222-6

Papadimitriou, K., Alegria, A., Bron, P. A., de Angelis, M., Gobbetti, M., Kleerebezem, M., . . . Kok, J. (2016). Stress

physiology of lactic acid bacteria. Microbiology and Molecular Biology Reviews, 80(3), 837-890.

doi:10.1128/mmbr.00076-15

Park, J. T., & Uehara, T. (2008). How bacteria consume their own exoskeletons (Turnover and recycling of cell wall

peptidoglycan). Microbiology and Molecular Biology Reviews, 72(2), 211-227. doi:10.1128/mmbr.00027-07

Sanchez, B., Champomier-Verges, M. C., Collado, M. D., Anglade, P., Baraige, F., Sanz, Y., . . . Zagorec, M. (2007).

Low-pH adaptation and the acid tolerance response of Bifidobactetium longum biotype longum. Applied and 22

Environmental Microbiology, 73(20), 6450-6459. doi:10.1128/aem.00886-07

Santivarangkna, C., Kulozik, U., & Foerst, P. (2008). Inactivation mechanisms of lactic acid starter cultures preserved

by drying processes. Journal of Applied Microbiology, 105(1), 1-13. doi:10.1111/j.1365-2672.2008.03744.x

Sen, M., Shah, B., Rakshit, S., Singh, V., Padmanabhan, B., Ponnusamy, M., . . . Sadhale, P. P. (2011). UDP-glucose 4,

6-dehydratase activity plays an important role in maintaining cell wall integrity and virulence of Candida

albicans. Plos Pathogens, 7(11). doi:10.1371/journal.ppat.1002384

Shevchenko, A., Tomas, H., Havlis, J., Olsen, J. V., & Mann, M. (2006). In-gel digestion for mass spectrometric

characterization of proteins and proteomes. Nature Protocols, 1(6), 2856-2860. doi:10.1038/nprot.2006.468

Spano, G., & Massa, S. (2006). Environmental stress response in wine lactic acid bacteria: Beyond Bacillus subtilis.

Critical Reviews in Microbiology, 32(2), 77-86. doi:10.1080/10408410600709800

Tourdot-Marechal, R., Fortier, L. C., Guzzo, J., Lee, B., & Divies, C. (1999). Acid sensitivity of neomycin-resistant

mutants of Oenococcus oeni: a relationship between reduction of ATPase activity and lack of malolactic

activity. Fems Microbiology Letters, 178(2), 319-326. doi:10.1016/s0378-1097(99)00377-8

Ugbogu, E. A., Wang, K., Schweizer, L. M., & Schweizer, M. (2016). Metabolic gene products have evolved to interact

with the cell wall integrity pathway in Saccharomyces cerevisiae. Fems Yeast Research, 16(8).

doi:10.1093/femsyr/fow092

Wang, C., Cui, Y. H., & Qu, X. J. (2018). Mechanisms and improvement of acid resistance in lactic acid bacteria.

Archives of Microbiology, 200(2), 195-201. doi:10.1007/s00203-017-1446-2

Yang, K., Liu, M., Wang, J., Hassan, H., Zhang, J., Qi, Y., . . . Zhang, G. (2018). Surface characteristics and proteomic

analysis insights on the response of Oenococcus oeni SD-2a to freeze-drying stress. Food Chemistry, 264, 377-

385. doi:10.1016/j.foodchem.2018.04.137

Zeidan, A. A., Poulsen, V. K., Janzen, T., Buldo, P., Derkx, P. M. F., Oregaard, G., & Neves, A. R. (2017).

Polysaccharide production by lactic acid bacteria: from genes to industrial applications. Fems Microbiology 23

Reviews, 41, S168-S200. doi:10.1093/femsre/fux017

Zhang, G., Fan, M., Lv, Q., Li, Y., Liu, Y., Zhang, S., & Zhang, H. (2013). The effect of cold, acid and ethanol shocks

on synthesis of membrane fatty acid, freeze-drying survival and malolactic activity of Oenococcus oeni.

Annals of Microbiology, 63(2), 477-485. doi:10.1007/s13213-012-0492-x

24

Figure captions

Figure 1 Cell damage responses of NAC and AC freeze-dried O. oeni SD-2a by FCM. A: NAC freeze-dried O. oeni

SD-2a (pH4.8); B‒D: AC freeze-dried O. oeni SD-2a (pH4.0, pH3.5 and pH3.2 respectively)

Figure 2 Scanning electron microscopy of freeze-dried O. oeni SD-2a with (B and F: pH4.0; C and G: pH3.5; D and H: pH3.2) or without (A and E: pH4.8) acid acclimation pretreatments. Resolution was 10k (A‒D) and 20k (E‒H) respectively.

Figure 3 2-DE images of proteomes in freeze-dried NAC and AC O. oeni SD-2a. A: 2-DE image marked with differential abundance proteins; B: 2-DE map of freeze-dried NAC O. oeni SD-2a (pH4.8); C‒E: 2-DE maps of freeze- dried AC O. oeni SD-2a (pH4.0, pH3.5, pH3.2 respectively)

Figure 4 Changes of differential abundance proteins in three comparisons (A‒C) and Venn diagram analysis of overlaps among the three comparisons (D, E). D: The numbers of differential abundance proteins with up-regulation; E: The numbers of differential abundance proteins with down-regulation.

Figure 5 Protein-protein interaction combined with enriched KEGG pathways in comparisons of (A) pH4.0/pH4.8, (B) pH3.5/pH4.8 and (C) pH3.2/pH4.8. Circle nodes represented proteins and filled color represented abundance fold change with gradient color from green (low) to red (high). Square nodes indicated KEGG pathways and filled color represented the value of –lg(P-value) with gradient color from yellow (low) to blue (high). Interactions between two proteins were presented with solid and dashed lines, the solid lines indicated the known interaction annotated in the

STRING database and the dashed lines between proteins indicated indirect interaction. The size of spot was related to the number of connected proteins, the more proteins it interacted with the bigger the spot was.

Figure 6 Proposed simplified carbohydrate metabolism pathways of DAPs with acid acclimation in freeze-dried O. oeni

SD-2a. The annotations were according to KEGG analysis (KEGG; http://www.kegg.jp/) as shown in Supplementary

Table S5. Color scale represented log2fold-change values and color of DAPs in comparisons of pH4.0/pH4.8, pH3.5/pH4.8 and pH3.2/pH4.8 were shown from left to right. P: phosphate, CoA: coenzyme-A, UDP: uridine- 25 diphosphate, dTDP: desoxythymidine diphosphate, Glu: glucose, Fru: fructose, Man: mannose, Glc: glucosamine,

NAc: N-acetyl, PRPP: 5-phosphoribosyl 1-pyrophosphate.

26

Highlights

 Acid acclimation improved cell integrity of O. oeni SD-2a.

 Proteomic data revealed acid acclimation on freeze-drying stress of O. oeni SD-2a.

 Acid acclimation significantly influenced carbohydrate metabolism of O. oeni SD-2a.

 Acid acclimation enhanced expression of proteins involved in cell wall composition.

27

28

29

30

31

32

33

34