Intestinal DCs global gene expression dynamics affected by recombinant baker’s yeasts saccharomyces cerevisiae

Background The baker’s yeast, saccharomyces cerevisiae, has been widely used throughout our daily life in diverse aspects for thousands of years. The saccharomyces cerevisiae was found to specically target the dendritic cells (DCs) in mammalian with a manner of antigen-receptor interaction as described previously. It is necessary to investigate the effect of the baker’s yeasts on global gene expression dynamics of intestinal DCs and explore the possibilities of using baker’s yeast as gene delivery vehicle to modulate animal’s immune functions Results with a murine oral delivery model in vivo, we conrmed the feasibility of using budding yeast as gene delivery vehicle to the intestinal DCs using the Western blots. We then examined the transcriptome prole of the mouse intestinal DCs upon yeast stimulus. The enrichment analysis of unique transcripts indicated the benecial role of yeast in modulating the DC-mediated adaptive immunity. Compared with previous study, we also found that a large fraction of the regulated genes is coincident with the response induced by other fungus, suggesting that the budding yeast induces a similar tailored unique genetic re-programming of DCs. Another analysis of transcriptome prole indicated that expression of β-catenin gene signicantly changes DCs gene expression related to inammatory response and cell adhesion. Conclusions Here, we dened the role of budding yeast on global gene expression of intestinal DCs, and conrmed the important role of β-catenin gene on the DCs-related inammatory response, which provides a framework for the development of mucosa yeast-based DNA vaccine.

of costimulatory molecules, and subsequent stimulation of native T cells in lymphoid organs [15]. The extensive reprogramming of DCs during maturation prompted us to measure the corresponding changes in gene expression. Thereafter, through diverse methods (e.g. oligonucleotide microarrays), many reports have discovered the interaction mechanisms between many DCs subsets (bone marrow-derived DCs) and many pathogens, including some fungi species (e.g. Candida albicans) [16,17]. But how intestinal DCs respond appropriately to the baker's yeasts S. cerevisiae they encounter daily remains ambiguous in genetic level.
Several studies have also shown that DCs are involved in the initiation of both innate and adaptive immunity [18]. The importance of DCs in initiating immune responses led us to investigate at a genetic level how intestinal DCs interact with yeasts. To further understand how yeast, affect intestinal DCs gene expression and to con rm the feasibility of using budding yeast cells as gene delivery vehicle to the intestinal DCs, we used a murine oral delivery model in vivo to pursue a whole genome transcriptomic analysis of the response of intestinal DCs to different recombinant yeasts.
We show that intestinal DCs suffer extensive genetic regulation upon diverse yeasts exposure, comprising 1442 transcripts (JMY-P group) and 2078 transcripts (JMY-Pcat group) compared with PBS group. A large fraction of these is similarly regulated at both exposures studied, comprising 1097 common unique transcripts and some of them have functions potentially related to T cell activation and adaptive immunity. And the JMY-Pcat group displayed another 944 DEGs other from the JMY-P group, which were mainly involved in "in ammatory response", "myeloid leukocyte activation", "leukocyte migration", "cellcell adhesion" and "extracellular structure organization". A comparison of yeast-regulated genes with a set of genes previously identi ed as the common DC transcriptomic response to pathogens [16] revealed about 77.70% (108/139) genes expressions were in accordance with the former study, indicating a similar gene cluster related to DCs-derived immune response in diverse fugus. Only a small fraction of the regulated genes is reverse with the response induced by other fungus, suggesting that the budding yeast induces a tailored unique genetic re-programming of DCs. After oral administration, the isolated DCs was utilized to test the possibility of application of oral gene delivery platform. Up to now, this is the rst report to systematically explore how saccharomyces cerevisiae modulate intestinal DCs-mediated immune response and successfully veri ed the feasibility of using budding yeast as gene delivery vehicle to modulate intestinal immune functions.

Vector validation in HEK293T cells
We used Xho and Not enzymes (NEB, UK) to digest the JMB84-CMV-β-catenin-polyA vector, which was in accordance with the prediction (two bands: 5969 bp and 3030 bp). We then digested JMB84-CMV-βcatenin-HA-polyA vector with EcoR and Nhe enzymes (NEB, UK) (two bands: 7917 bp and 1118 bp). And we also used EcoR and Nhe enzymes (NEB, UK) to digest the JMB84-CMV-β-catenin-HA-T2A-GFP-polyA vector (two bands: 7917 bp and 1901 bp) (data not shown). We also determined the sequence accuracy of these vectors with sequencing, in consistent with the released sequence in NCBI. Then we transfected the JMB84-CMV-β-catenin-HA-T2A-GFP-polyA vector into the HEK293T cell. 48 hours after transfection, clear uorescence was observed which proved the GFP expression in cells (Fig. 1A), which proved the utility of this vector in subsequent experiments.
Detection of HA-tagged fusion protein expression in the CD11c+ DCs from orally administrated mice After separating CD11c+ DCs from the orally administrated mice, we observed the representative HAtagged fusion protein expression from the JMY-Pcat group ( Fig. 1B(a)), while the control group was negative. We con rmed the feasibility of using budding yeast cells as gene delivery vehicle to the intestinal DCs. Notably, as displayed in the Fig. 1B(b), the mice fed with recombinant yeast containing JMB84-CMV-β-catenin-HA-polyA for ve days showed the highest relative expression calibrated with reference protein (β-actin), indicating the time point with highest expression of target protein is 5-day after oral administration, which provides a foundation for development of oral DNA vaccine in the future.
After 5-day orally administration, we separated intestinal CD11c+ DCs and isolated RNA, then fragmented RNA to synthesize cDNA and connected short fragments for the nal sequencing. To investigate the genes and signaling pathways that are associated with DC responses to yeast stimuli, we performed a differential expression analysis between different groups. Treatment with different yeasts resulted in transcriptional changes in all DCs examined, albeit to differing degrees (Fig. 1C). According to the analysis result (Additional le 6, Table S3), a total of 18410 genes were screened, including 9488 induced genes and 8922 repressed genes between control and JMY-P group; a total of 18942 genes were screened, including 10807 induced genes and 8135 repressed genes between control and JMY-Pcat group; a total of 19188 genes were screened, including 11103 induced genes and 8084 repressed genes between JMY-P group and JMY-Pcat group.
Using signi cance analysis of multiclass testing type using a cut-off of P < 0.05, after excluding redundant DEGs, we found that JMY-P yeast and JMY-Pcat regulate a high number of unique genes or DEGs (1442 and 2078) at the RNA level at least twofold, considering both induced and repressed transcripts (Additional le 7, Table S4). We segregated the 1442 differentially regulated transcripts between those that are: (i) induced upon JMY-P treat (1062 genes, 73.6%) and (ii) repressed upon JMY-P treat (380 genes, 26.4%); And we also found that 2078 differentially regulated transcripts between those that are: (i) induced upon JMY-Pcat treat (1747 genes, 84.1%) and (ii) repressed upon JMY-Pcat treat (331 genes, 15.9%) ( Fig. 2A). The intersection of these groups (JMY-P and JMY-Pcat) responses revealed a common set of 1097 highly regulated genes, which further proved the signi cant modulating role of yeast on CD11c+ DCs. According to the comparison based on selected 1097 genes between these two groups, we found that the corresponding trend of regulated 1091 genes (99.5%) in intestinal CD11c+ DCs upon different yeasts treats is very similar at these two stimuli. Yeast preferentially regulates DCs genes involved in T cell activation, cytokine production, adaptive immune response, positive regulation of immune response, Cytokine-cytokine receptor interaction, Natural killer cell mediated cytotoxicity and Hematopoietic cell lineage Using the Gene Ontology systems of classi cation and the Kyoto Encyclopedia of Genes and Genomes (KEGG), we analyzed gene annotation for the total unique genes or ESTs that were regulated after yeast treatment. As might be expected, the host response is enriched for ESTs genes (1097) that are associated with the immune response, particularly those that encode cytokines and chemokines. In consideration of the p-value of different class ( Fig. 2B and 2D), we found that the main biological process under yeast stimuli are: "T cell activation" (126 genes), "cytokine production" (126 genes), "adaptive immune response" (103 genes), "positive regulation of immune response" (113 genes), "myeloid leukocyte activation" (53 genes) and "leukocyte migration" (67 genes) ( Fig. 2B and 2C; Additional le 9, Table S6). These class induced by the yeast stimuli, indicated the important boosting in uence of yeast on the DCsparticipated immune response. From a series of heatmaps for many terms (Additional le 2, Fig S2A), we found that many diverse genes dramatically increased upon the yeast treat, further demonstrating the important in uence of yeast on the DCs-mediated immune response.
Using the Kyoto Encyclopedia of Genes and Genomes (KEGG) analysis ( Fig. 2B and 2D) and from a series of heatmaps for many terms (Additional le 2, Fig S2C), we found that yeast preferentially regulates the following metabolic pathways: "Cytokine-cytokine receptor interaction" (50 genes), "Natural killer cell mediated cytotoxicity" (32 genes) and "Hematopoietic cell lineage" (27 genes) (Additional le 9, Table S6), further demonstrating the yeast's important in uence on cytokine or chemokine signaling pathway which actively participated into the immune response and Hematopoietic cell lineage-related metabolism.
Altogether, the data on functional annotation of yeast regulated transcripts of DCs show that the yeast exerts pressure on many cell functions including diverse metabolic pathways, but the major target was concentrated on Cytokine-cytokine receptor interaction and Chemokine signaling pathway, which plays very important roles in regulating the DCs-mediated immune response and T cell activation.
JMY-Pcat yeast regulates 944 genes in CD11c+ DCs compared with the JMY-P group We then used the JMY-P group as control group and conducted signi cance analysis of multiclass testing type using a cut-off of P < 0.05, after excluding redundant DEGs, we found that JMY-Pcat yeasts regulate a high number of unique genes or ESTs (944) at the RNA level at least twofold, considering both induced and repressed transcripts. We segregated the 944 differentially regulated transcripts between those that are: (i) induced compared with JMY-P treat (802 genes, 85.0%) and (ii) repressed compared with JMY-P treat (142 genes, 15.0%) ( Fig. 2A). Individual genes following different expression pro les and complete lists of the transcripts in each group are shown (Additional le 1, Fig S1B; Additional le 10, Table S7). Such a large-scale change in gene expression and the Circos plots (Fig. 3A) demonstrated that β-catenin truly expressed in intestinal DCs and β-catenin modulates a marked gene expression change on the intestinal DCs.
In addition, we found that the intersection of the two different comparisons (PBS vs JMY-P and JMY-P vs JMY-Pcat) revealed a common set of 388 highly regulated genes (Additional le 1, Fig S1C). We also found that the 18 selected genes expressions of JMY-Pcat group is relatively higher than other two groups within the 388 ESTs ( Fig. 3E), which further proved the expression of β-catenin in DCs and the signi cant modulating role of β-catenin on CD11c+ DCs function.
JMY-Pcat yeast preferentially regulates DCs genes involved in in ammatory response, myeloid leukocyte activation, leukocyte migration, cell-cell adhesion, extracellular structure organization, positive regulation of cytosolic calcium ion concentration, and Cytokine-cytokine receptor interaction Using the Gene Ontology systems of classi cation and the Kyoto Encyclopedia of Genes and Genomes (KEGG), we analyzed gene annotation for the total DEGs that were regulated in the comparison between JMY-P and JMY-Pcat treatments ( Fig. 3B and 3C; Additional le 11, Table S8). We found that the main biological process under JMY-Pcat yeast regulation are: "in ammatory response" (70 genes), "myeloid leukocyte activation" (31 genes), "leukocyte migration" (39 genes), "cell-cell adhesion" (60 genes), "extracellular structure organization" (34 genes) and "positive regulation of cytosolic calcium ion concentration" (33 genes). From a series of heatmaps for many terms (Additional le 3, Fig S3), we found that many diverse genes dramatically increased upon the JMY-Pcat treat, further demonstrating the important in uence of β-catenin on the DCs-mediated immune response, especially the in ammatory response and leukocyte-related functions. Besides these categories, it should not be ignored that βcatenin participated into the "cell-cell adhesion", "extracellular structure organization" and "positive regulation of cytosolic calcium ion concentration". Together with the high number of genes with metalloproteinase activity, it suggested that JMY-Pcat treat had increased boosting role on regulating permeability of ion channels and activity of ion pumps. JMY-Pcat yeast also preferentially regulates the following metabolic pathways: "Cytokine-cytokine receptor interaction" (29 genes) (Additional le 11, Table S8).
Altogether, the data on functional annotation of JMY-Pcat yeast regulated transcripts of DCs show that the β-catenin exerts many promoting roles on many DCs functions including diverse metabolic pathways differ from the JMY-P yeast, but the major targets were concentrated on in ammatory response and leukocyte-related functions, which plays very important roles in regulating in ammation.
Cluster analysis of DEGs among the three comparative groups Finally, we conducted the cluster analysis of DEGs among three comparative groups using the Metascape (http://metascape.org/). And it was found that every two groups shared many DEGs ( Fig. 4A, 4B, 4C, 4D and 4E). Using the Gene Ontology systems of classi cation and the Kyoto Encyclopedia of Genes and Genomes (KEGG), these DEGs mainly concentrated on following categories: "T cell activation", "in ammatory response", "cytokine production", "positive regulation of immune response", "mononuclear cell migration" and "leukocyte activation involved in immune response". It further demonstrated the yeast and β-catenin could in uence the DC-mediated adaptive immunity and in ammatory response.
Yeasts-mediated regulation of genes expression by DCscon rmation using quantitative real-time polymerase chain reaction Transcriptomic analysis shows that DC expression of many genes such as CD80, CD6 and CD40LG is induced with yeast treatment. To con rm the accuracy of RNA-seq, we performed quantitative real-time polymerase chain reaction (PCR) analysis (Fig. 5). For up-regulated genes, the relative expression levels in JMY-P group was signi cantly higher than that in the PBS group (including TLR9, CD6, CD80, IL12RB1, Gas1, ITGAE, Serpina3g, Cxcr3, CD3d, Gzmk, Serpinb6b, Ly6c2, IL10RA, Fcrl1, Rnase6, CD40lg, Klra7, Rhoh, Arhgap15 and Lair1), while expression levels in JMY-Pcat group was relatively higher than that in JMY-P group (including IL10RA, Rnase6, Tff2 and Cfd). For down-regulated genes, the relative expression levels in JMY-P group was signi cantly lower than that in the PBS group (Gsdmc3), while the expression level of JMY-Pcat group is lower than the JMY-P group (such as NeuroD2 and Spata17). Above results reinforced the accuracy of the RNA-seq.

Discussion
No speci c or detailed investigation on the modulation of intestinal DCs gene expression and function by baker's yeast saccharomyces cerevisiae has been addressed, despite the importance of DCs for initiation and maintenance of the protective immunosurveillance [19]. To ll this gap, we comparatively examined responses generated in intestinal CD11c+ DCs upon encounter with different yeasts, in consideration of the supposed, differential modulation role of the two yeasts in immune response and other pathways.
We used the murine model to accomplish oral administration of the baker's yeast saccharomyces cerevisiae. Abundant yeasts can guarantee the enough interaction between the DCs and yeasts in intestine area. 5-day after administration, we separated the CD11c+ DCs and conducted the RNA-seq to examine the different response in the transcriptional level.
As might be expected, the common host response upon encounter with the yeast is enriched for genes that are associated with the immune response, particularly those that encode cytokines and chemokines.
We identi ed and grouped several functional groups of gene products according to the part in which they function. Functional groups of genes are described in more detail below.
Gene that mediates DCs activation and maturation. There is a known fact that APCs (Macrophages and DCs) play diverse important roles in adaptive immunity [14,20]. We found that the phagocytosis of yeast by intestinal DCs can induced highly-expressed CD19 providing the rst signal for activating the T leukomonocyte and enhance the antigen-presenting ability, and some marker genes displayed higher expressions, such as ICAM-1, ICAM-3, CD80 and CD40, further indicating the maturation and activation of DCs after engul ng the yeast. Activated APCs can use pattern recognition to distinguish the "self" or "nonself" to initiate different adaptive immunity (humoral immunity or cellular immunity) [21] and the immune system can increase some effectors expression of the APCs to enhance the capture and presentation abilities [22]. Our results further proved the enhancing role of yeast on the activation and maturation of DCs, which can provide a foundation for explaining the boosting role of yeast on T cell activation and cytokine production.
Gene that activates the adaptive immunity. The group of genes that is most strongly and most consistently upregulated consists almost entirely of those encoding cytokines or chemokines. These processes mainly concerned on the DCs-Lymphocytes interactions. As we all know, lymphocytes (B cells or T cells) are activated by antigen-speci c receptors on the cell surface [23], which causes the cells to proliferate and differentiate into speci c effector lymphocytes. For example, activated B cells can produce antibodies-producing cells, and some activated T cells become cytotoxic T cells. [24,25]. In this study, for the APCs-T cells interaction, the DEGs include CD2, CD3, CD4, CD6, CD8α1, CD8β1, CD40L, CD72, CTLA-4, IL-1β, IL-16 and ICOS; for the APCs-B cells interaction, the DEGs include CD19, CD40, CD79α, CD79β and CD80. In addition, there are some other chemokines being expressed highly, which participate into the proliferation and migration of T helper cells [26], such as Cxcl1, Cxcl2, Ccl3, Ccl4, Ccl5 and Ccl22; some chemokines receptors were also found much higher expressed in activated DCs, including Cxcr3, Cxcr4, Cxcr5, Cxcr6, Ccr2, Ccr5, Ccr6, Ccr7, Ccr9 and Cx3cr1. All these increased expressions indicated that the cooperation of chemokine and the coupling receptors may initiate the directional migration of target cells. Additionally, the immunological synapse plays an important role in participating the interaction between DCs and T cells [27], which can promote the binding ability of TCR and MHC complex, the signal transduction of T cells and the effector function of T cells. Herein, CD3, ICAM-1, ICAM-5, LCK, ZAP70, LAT, VAV1, Was, Rac2, Fyn and Fyb, are the most DEGs participating into the formation of immunological synapse upon encounter with yeast. There are also some increasingly expressed transcription factors involving in the development and activation of leukomonocytes, including NFAT (Nfatc1 and Nfatc1), Relt, Stat4, ETS-1, PAX5, BATF, IRF4, GATA-2 and GATA-3. Induction of signaling genes and transcription factors may be involved in the process of allowing DCs to receive regulatory signals from lymphatics and lymph nodes. In a word, the yeast plays the bene cial role of in modulating the adaptive immunity through diverse pathways.
Gene that mediates in ammation. The strongest, most persistently up-regulated set of genes consists almost entirely of genes encoding cytokines called in ammatory/chemokine cytokines. This set contains genes encoding pro-in ammatory mediators, such as TNFAIP3, IFNγ, IL1β and IL18RAP, and PTGS2 enzyme (epoxidase-2), involved in the production of prostaglandins.
Other common host-response-gene. This set of genes includes genes encoding OAS (2',5'-oligo adenylate synthetase)1h, OAS2 and IFIM1. As expected, similar with Candida albicans exposure [16], some same genes are upregulated in DCs stimulated with yeast, which further describing the important role of yeast in regulating IFN-stimulated function of the DCs. What the fact is equally important is the dramatically increased expressions of the TLR family genes upon yeast exposure, suggesting an intensive capacity of DCs to communicate with other cells of both the innate and adaptive arms of the immune system. A class of cytokine receptors (IL10RA, IL12RB1, IL12RB2, IL1R2, IL21R, IL22RA2, IL27RA, IL2RB, IL2RG,  IL7R, and IL9R) were also induced. The expression of these receptors may allow DCs to respond to lymphocyte-derived interleukins within the lymph node.
To better understand the overlap of the total transcriptome regulation of DC cells after yeast stimulation with changes caused by other fungus, and the similarity of regulation of DC cell maturation, we compared a series of orthologs genes that are regulated by a fungus in human monocyte-derived DCs [16] to the yeast-regulated genes in this study. About 77.70% (108/139) genes expressions were in accordance with the former study (Additional le 12, Table S9), indicating a similar gene cluster related to DCs-derived immune response in diverse fugus. More importantly, we found 38 genes had a reverse expression trend combined with this report, and it further validated that the DCs detect diverse microorganism and induce tailored speci c immune responses.
We then con rmed the DEGs expressions of separated DCsin JMY-Pcat group with the comparison analysis between the two groups: JMY-P and JMY-Pcat. As we all know, β-catenin is a bifunctional protein involved in the regulation and coordination of cell-cell adhesion and gene transcription [28], which is a known subunit of the cadherin protein complex and functions as a signal transducer in the Wnt signaling pathway [29]. Mutation and overexpression of β-catenin is associated with the development of many cancers, including hepatocellular carcinoma, colorectal cancer, lung cancer, malignant breast tumors, ovarian cancer, and endometrial cancer [30]. Changes in β-catenin localization and expression levels are also associated with a variety of heart diseases [31], including dilated cardiomyopathy. In the process of adherens junction, β-catenin molecules are also recruited by cadherins onto their intracellular regions. β-catenin, in turn, associates with another important protein, α-catenin that directly binds to the actin laments [32]. The β-catenin-α-catenin complex can thus physically bridge cadherins with the actin cytoskeleton [33]. Based on our analysis, we also identi ed and grouped several functional groups of gene products according to the part in which they function upon encounter with the JMY-Pcat yeast. Functional groups of genes are described in more detail below.
Genes that mediate in ammatory response. Core stimulated genes are as follows: Ccl1, Ccl3, Ccl4, Ccl7, Ccr3, Chi3l3, Cxcl2, Cxcl3, Cxcl5, Cxcl9, Fcer1a, Il13, Il1a, Il1b, Il23a, Il27, Il4, Il5ra, Itih4, Nfkbid, Tlr11, Tlr8, Siglece and so on. These core immune-related DEGs again proved the importance of β-catenin in modulating the immune response, especially the "in ammatory response". As we all know, the in ammatory response has a tight relationship with the leukocyte activation and leukocyte migration. Based on our analysis, for the "leukocyte migration", the signi cant reproduction, and tissue remodeling. These regulated gene families further proved the important role of βcatenin on actins arrangement for DCs, which was hoped to be involved in "leukocyte migration", "cell-cell adhesion", "extracellular structure organization", "positive regulation of cytosolic calcium ion concentration". Other from these genes, there are some gene families which should be targeted. Alpha-actinins (ACTN2 and ACTN3), which helps to anchor the myo brillar actin laments; Pcdhga5, Pcdhga7, Pcdhgb1, Pcdhgb4 and Pcdhgb5 are integral membrane proteins that mediate calcium-dependent cell-cell adhesion. But the speci c and detailed network of this process remains unclear and need more efforts to discover it.
Much more than this, as described previously, saccharomyces cerevisiae can be an effective oral DNA vaccine vehicle targeting for intestinal cells and provoke strong immune response in different animal species [38][39][40]. Our results further proved the utility of oral yeast-based gene delivery vehicle targeting for intestinal DCs directly, which lays a solid foundation for the development of yeast-based DNA vaccine.

Conclusions
In a word, we con rmed the bene cial role of yeast on modulating the intestinal DCs functions and proved the feasibility of using budding yeast cells as gene delivery vehicle to the intestinal DCs directly.
Moreover, we further validated the important modulating role of β-catenin on the DCs-mediated immune response and other functions.

Vectors construction and veri cation
Based on the pcDNA3.1(-) vector, we ampli ed the CMV promotor and polyA fragment using the following primers: CMV-F/CMV-R and PA-F/PA-R; then referred on the β-catenin sequences (NM_001098209.1) from NCBI database, we ampli ed the β-catenin fragment with special primers: B-F/B-R templated on cDNA of HEK293T cells. Next, we conducted overlap PCR to get the CMV-β-catenin-polyA cassette, which was templated on CMV, β-catenin and polyA fragments. Finally, CMV-β-catenin-polyA cassette and JMB84 base vector (maintained in our lab) were digested with SacI /NotI and Xho /Not , separately; then we conducted gel extraction and ligated them with T4 DNA ligase to get the vector JMB84-CMV-β-catenin-polyA. In the following step, to get the β-catenin-HA fragment, we made two continuous-step PCR reaction using primers Beta-F/ HA-R1 and Beta-F/ HA-R2 based on the JMB84-CMV-β-catenin-polyA vector.
To insert the GFP reporter gene into the JMB84-CMV-β-catenin-HA-polyA vector, we rst ampli ed the βcatenin-HA fragment templated on the JMB84-CMV-β-catenin-HA-polyA vector using special primers Beta-F/Beta-R; meantime, based on the pLenti-T2A-GFP vector, we ampli ed the T2A-GFP fragment with primers: T2A-GFP-F/T2A-GFP-R. Then overlap PCR was conducted to get the β-catenin-HA-T2A-GFP cassette using β-catenin-HA and T2A-GFP fragments as templates. Finally, we digested this cassette and JMB84-CMV-β-catenin-polyA vector with EcoR and Nhe enzymes; then we made the gel extraction of above cassettes and ligated them with T4 DNA ligase to get JMB84-CMV-β-catenin-HA-T2A-GFP-polyA vector. All above vectors were digested by special enzymes and determined by sequencing (data not shown). All primers sequences were displayed in Additional le 4, Table S1. We then transfected the vector JMB84-CMV-β-catenin-HA-T2A-GFP-polyA into the HEK293T cells to verify the successful expression of β-catenin.
Mice and oral administration For optimization of oral administration formula and con rm the feasibility of using budding yeast as gene delivery vehicle, mice were randomly divided into ve groups, and each group contained six female individuals. Four groups were orally administrated with 1×10 8 yeast [42] containing JMB84-CMV-βcatenin-HA-polyA vector suspended in 100ul PBS buffer for 3, 4, 5 and 7 days separately once a day. The control group was fed with the same volume of yeast containing naked vector. Before oral administration, mice need fasting treatment for 12h. After limited feeding time, mice were sacri ced for intestinal CD11c+ DCs separation post-anesthesia [39]. Then the best formula was determined via the Western blots through detecting the amount of HA-tagged fusion proteins.
Once determined the best formula, another batch of mice were randomly divided into three groups, and each group contained six female individuals. Mice from different groups were orally administrated with Euthanasia/sacri ce methods After 5 days' administration, mice were sacri ced for intestinal CD11c+ dendritic cells separation postanesthesia, the detailed method was as following: according to previous reports [43], we chose intraperitoneal injection to perform this experiment. This is usually the most ideal method when it can be performed without causing fear or pain to the animal. When appropriately administered, an acceptable injection of the euthanasia agent can result in a smooth loss of consciousness before the heart and/or respiratory function ceases, minimizing pain and distress in the animal. The injectable euthanasia agent is Nembutal ie sodium pentobarbitone and it was administered at 150 mg/kg to accomplish the anesthesia. Overdose using this anaesthetic solution at 3-5 times. After that, we ensured the death by cervical dislocation or by opening the chest to collapse lungs.

Separation and puri cation of intestinal CD11c+ DCs
According to previous report [39], After incubation at 4°C for 30 min, intestinal DC cells were subjected to magnetic cell sorting from mixed cells using a MACS separation column (Miltenyi Biotec, Bergisch, Germany).

Western blots
After separating the intestinal CD11c+ DCs, total proteins were extracted by using the RIPA lysis buffer (Thermo, MA, USA). After the total RNA quality detection and DNase I treatment, we used magnetic beads (Invitrogen, CA, USA) to isolate mRNA. After mixing with the buffer, the mRNA is divided into short fragments. Then we synthesized the cDNA using the mRNA fragment as a template. The short fragment was then puri ed and resolved with EB buffer and single nucleotide A (adenine) was added for end repair.
In addition, we use speci c adapters to connect short segments together. After agarose gel electrophoresis, the appropriate fragment was selected as a template for PCR ampli cation. In the QC step, the quality and quantity of the sample library was identi ed using the Agilent 2100 Bioanaylzer and the ABI StepOnePlus Real-Time PCR system. At last, all DNA fragments was sequenced using Illumina HiSeqTM 2000 (LC Sciences, USA). High quality clean data obtained by quality control from raw data, were aligned to the reference sequence using SOAPaligner/SOAP2. According to the distribution and coverage of the reads, quality of reads was con rmed. Gene expression levels were normalized by considering the RPKM value (reads per kilobase of the exon model per million mapped reads). Based on the expression levels, the signi cant unigenes that were differentially expressed (DEGs) between every two experiment groups were identi ed "|fold change|≥2" and "P-value< 0.05" as threshold to judge [44].

GO enrichment analysis and Pathway enrichment analysis
Based on the gene ontology database (http://www. geneontology.org/), this part of the experiment analyzes the function of DEGs between different groups through GO analysis. GO analysis is a common method for large-scale gene function enrichment research. In this study we used the R package clusterPro ler for GO enrichment analysis. The p value < 0.05 and the FDR value < 0.05 were used as the cutoff criterion for GO enrichment analysis [45].
KEGG is a commonly used bioinformatics database that includes information about biochemical pathways. In this study, the R package clusterPro ler was used to perform KEGG pathway enrichment analysis, and some pathways with signi cant differences were screened. The p-value < 0.05 was used as the cutoff criterion for KEGG enrichment analysis. Network enrichment analysis was performed using Metascape (http://metascape.org/).