Surf4 facilitates reprogramming by activating the cellular response to endoplasmic reticulum stress

Abstract Objectives Maternal factors that are enriched in oocytes have attracted great interest as possible key factors in somatic cell reprogramming. We found that surfeit locus protein 4 (Surf4), a maternal factor, can facilitate the generation of induced pluripotent stem cells (iPSCs) previously, but the mechanism remains elusive. Materials and Methods In this study, we investigated the function and mechanism of Surf4 in somatic cell reprogramming using a secondary reprogramming system. Alkaline phosphatase (AP) staining, qPCR and immunofluorescence (IF) staining of expression of related markers were used to evaluate efficiency of iPSCs derived from mouse embryonic fibroblasts. Embryoid body and teratoma formation assays were performed to evaluate the differentiation ability of the iPSC lines. RNA‐seq, qPCR and western blot analysis were applied to validate the downstream targets of Surf4. Results Surf4 can significantly facilitate the generation of iPSCs in a proliferation‐independent manner. When co‐expressed with Oct4, Sox2, Klf4 and c‐Myc (OSKM), Surf4 can activate the response to endoplasmic reticulum (ER) stress at the early stage of reprogramming. We further demonstrated that Hspa5, a major ER chaperone, and the active spliced form of Xbp1 (sXbp1), a major mediator of ER stress, can mimic the effects of Surf4 on somatic cell reprogramming. Concordantly, blocking the unfolded protein response compromises the effect of Surf4 on reprogramming. Conclusions Surf4 promotes somatic cell reprogramming by activating the response to ER stress.


| INTRODUC TI ON
Terminally differentiated somatic cells can be reprogrammed into a pluripotent state by Yamanaka factors (Oct3/4, Sox2, Klf4 and c-Myc). 1 This transition is accompanied by global and dramatic changes at the transcriptional, epigenetic and metabolic levels. [2][3][4] Although many cellular mechanisms have been revealed to date, the process of reprogramming is still inefficient, time-consuming and stochastic. 5 Somatic cell nuclear transfer provides a fast, relatively efficient and deterministic reprogramming model in which terminally differentiated somatic nuclei can be reprogrammed to a totipotent state by factors in the oocyte cytoplasm. 6 Therefore, the role of oocyte factors in somatic reprogramming has been widely studied, and an increasing number of studies have shown that oocyte factors can improve reprogramming efficiency. 7,8 We previously found that some oocyte-enriched proteins identified through mass spectrometry 9 can enhance somatic cell reprogramming. 10 In this study, we further explored the function and mechanism of Surf4, which is one of these maternal factors.
SURF4, known as endoplasmic reticulum (ER)-derived vesicles protein 29 (Erv29p) in Saccharomyces cerevisiae and surfeit locus protein 4 homolog (SFT-4) in Caenorhabditis elegans, is an integral ER membrane protein 11,12 and is required for packaging soluble secretory proteins into ER-derived transport vesicles. 11,13 Binding to the amino-terminal hydrophobic tripeptide motifs of soluble cargo proteins with different affinities, SURF4 enables prioritization of their exit from the ER. 14 SURF4 circulates between the ER/ER-Golgi intermediate compartment (ERGIC)/Golgi and mediates the anterograde or retrograde transport of cargo proteins. 13,[15][16][17][18][19] Disruption of Surf4 trafficking results in a reduction in the number of ERGIC clusters 20 and accumulation of cargo proteins in the ER compartment. 21 Dysregulation of ER-Golgi vesicle transport induces ER stress, 22 and in turn, when ER stress occurs, the expression of Erv29p significantly increases. 23 Deficiency of Surf4 in mice results in embryonic lethality after implantation. 21 A study of lipoprotein transport revealed that SURF4-mediated ER export of lipoproteins controls lipid homeostasis in mice and humans. 24 During mouse preimplantation development, Surf4 was highly enriched in MII oocytes 9,25 and early embryos before the two-cell stage. 25,26 In this paper, we demonstrate that Surf4 significantly promotes Yamanaka factor-mediated iPSC generation via activation of the response to ER stress.
The specific pathogen-free mice were housed in the animal facility of Tongji University. All our study procedures were consistent with the Tongji University Guide for the care and use of laboratory animals.  After colonies formation, the cells were cultured in ESM without Dow for further 2-3 days, and then the colonies were mechanically picked for establishing the iPS cell lines.

| Cell growth curve
The MEFs were plated onto 12-well plates at a density of 1.2 × 10 4 cells per well and were harvested every 48 or 72 h and counted in a haemocytometer. Each group contained three replicates.

| Alkaline phosphatase (AP) staining
Alkaline phosphatase staining kit (Beyotime, C3206, China) was used for AP staining according to the instructions of the manufacturer. In briefly, at the end of reprogramming, the cells were washed once by Dulbecco's Phosphate-Buffered Saline (DPBS), and fixed by 10% formaldehyde solution for 5 min at room temperature. Then, the cells were washed once by deionized water and stained by the reagent provide by the kit.

| RNA isolate and real time PCR
Total RNA was extracted using TRNzol Universal Reagent (Tiangen) and reverse transcribed using the 5× All-In-One RT MasterMix (ABM). Quantitative reverse-transcription PCR was performed with SYBR ® FAST Universal qPCR Kit (KAPA) and the ABI7500 Fast Realtime PCR system (Applied Biosystems) or QuantStudio5 (Applied Biosystems). The reactions were performed in triplicate and relative mRNA expression is normalized to β-actin as an endogenous control using the ΔCT method. Primer sequences are available in the Table S1.

| Immunofluorescence (IF) staining
Immunofluorescence staining was performed as previously described. 30 Cells growing on slides were fixed with 4% paraformal-

| Teratoma formation
The iPSCs were trypsinized and a total of 2-5 × 10 6 iPSCs were subcutaneously injected into the groin of SCID mice. Four to eight weeks post-injection, teratomas formed and were very palpable. The tumour samples were dissected and processed for haematoxylineosin staining.

| RNA-sequencing and data processing
Total RNA from independent biological replicates of MEFs, day All of the RNA-Seq sequencing reads were processed using BBDuk (version 38.34) to remove adapters and low-quality reads. 31 The filtered reads were mapped to the mouse reference genome using STAR (version 0.6.0) with the default parameters except for the 'quantMode GeneCounts' parameter. 32 Gene expression for each sample was quantified by FPKM using StringTie (version 1.3.3b).

| Statistical analysis
The statistical data are presented as the mean ± SEM of at least three independent experiments. Significance was calculated using Student's t tests.

| Surf4 can facilitate iPSCs induction
In our previous study, by mining proteomic data of preimplantation embryos, we found that several maternal factor candidates can facilitate somatic cell reprogramming. 10 In the present study, we aimed to explore the function and mechanism of one of the ma- As indicated by the Oct4-GFP signal, many more iPSC colonies appeared in the Surf4 OE group after day 9 ( Figure 1A). At the end of reprogramming, Surf4 caused an approximate 4-to 8-fold increase in iPSC colony number and up to a 20% increase in the percentage of Oct4-GFP + cells ( Figure 1B). Together with improving reprogramming efficiency, we also observed that OE of Surf4 reduced cell proliferation during the process ( Figure S1B). However, this phenomenon did not recur in MEFs ( Figure S1C), which suggested that the proliferation attenuation by Surf4 was dependent on reprogramming. Therefore, we monitored the reprogramming kinetics with or without the OEof Surf4 and found that OE of Surf4 significantly reduced the THY1 + cell population ( Figure S1D) and increased the SSEA1 + cell percentage ( Figure S1E) during reprogramming. The primary iPS colonies in the Surf4 OE group exhibited normal morphology with a multiplied colony number compared with the control group ( Figure 1D), as presented by AP staining ( Figure 1E). In contrast, knockdown of Surf4 (Surf4 KD) led to a decrease in the number of AP + or Oct4-GFP + colonies and the percentage of Oct4-GFP + cells ( Figures 1C and S1F) without influencing the morphology of the iPSCs ( Figure S1G).

Established iPS cell lines derived upon the OE of Surf4
(OSKM + Surf4-iPSCs) displayed typical dome-shaped morphology resembling embryonic stem cells (ESCs) ( Figure 1F). Most of the iPS cell lines possessed a normal karyotype ( Figure S1H). The pluripotent genes were activated at the RNA and protein levels ( Figure 1G,H).
We also evaluated the differentiation ability of these iPS cell lines in vitro and in vivo. The cells can differentiate into cells from all three germ layers through embryoid body (EB) formation ( Figure S1I). They could also form teratomas consisting of cells from three germ layers after subcutaneous injection into SCID mice ( Figure 1I). Thus, Surf4 can facilitate iPSC generation without influencing pluripotency.

| Global profile of the effects of Surf4 on reprogramming
To investigate the effect of Surf4 on reprogramming, we analysed the transcriptomes of reprogramming cells with or without Surf4 on day 3 and MEFs. Based on the correlation matrix ( Figure 2A) and PCA ( Figure S2A), the two reprogramming cell samples were distinct from MEFs. When compared to MEFs individually, the reprogramming cells with or without OE of Surf4 had 3880 and 4077 differentially expressed genes (DEGs), with more than one-half of these DEGs shared between both cell samples ( Figure 2B). These shared genes were related to reprogramming: The upregulated genes were mainly enriched in keratinization, suggesting that the mesenchymal-toepithelial transition (MET) occurred, and the downregulated genes were related to focal adhesion and development ( Figure S2B).
By comparing the three cell groups, we obtained 443 DEGs (fold change >1.5, FDR <0.05), which were clustered into six groups ( Figure 2C and Table S2). A large number of genes downregulated or mildly upregulated in the early stage of reprogramming (as the control group showed) were markedly upregulated by Surf4 OE (Cluster I, 50 genes; Cluster II, ~150 genes and Cluster III, ~110 genes). These genes were mainly enriched in vesicle-mediated transport and response to ER stress, which were closely related to the function and localization of SURF4. In addition, the expression of dozens of cell cycle genes that was upregulated at the early phase of reprogramming but was decreased in the Surf4 OE group (Cluster V and Cluster VI), which was consistent with the proliferation suppression function of Surf4 ( Figure 1C). The expression levels of the DEGs were also confirmed by qPCR (Figures S2C,D).  Figure S1 and Table S1

| Activation of the response to ER stress facilitates reprogramming
To determine the effect of protein transport and ER stress on reprogramming, we employed brefeldin (BFA), a specific inhibitor of protein trafficking, and found that it could enhance reprogramming in a dose-dependent manner ( Figure 3A). BFA is an ER-Golgi transport inhibitor that has been shown to cause protein accumulation in the ER and lead to ER stress. 34 Then, we tested two other ER stress inducers: tunicamycin (Tm), which inhibits N-linked glycosylation and disrupts protein maturation in the ER, 35 and thapsigargin (Tg), which inhibits sarcoplasmic and ER Ca 2+ -ATPase (SERCA), which subsequently depletes Ca 2+ stores in the ER. 36,37 Tm and Tg both promoted reprogramming at proper concentrations but suppressed reprogramming at high concentrations owing to impaired cell survival ( Figure S3A,B). However, when we tried an ER stress inhibitor, tauroursodeoxycholic acid (TUDCA), which functions as a chemical chaperone, reduces stress-induced aggregation of proteins and inhibits the PERK pathway to prevent unfolded protein response (UPR) dysfunction, 38,39 we found that it negligibly affected reprogramming efficiency ( Figure S3B). Thus, we speculated that the response to ER stress can promote reprogramming, and it may not through PERK pathway.
Upon ER stress, the UPR is triggered and mediated by the IRE1-XBP1, PERK-eIF2α or ATF6 pathways to activate downstream transcription factors to reduce global protein synthesis and enhance the cellular protein-folding capacity. Eventually, these factors relieve stress and re-establish ER homoeostasis or lead to apoptosis if they fail to recover. Then, we examined the effects of the main mediators of the response to ER stress on re-programming. OE of Hspa5 or the active spliced form of Xbp1 (sXbp1) dramatically increased the number of Oct4-GFP + or AP + iPSC colonies and the percentage of Oct4-GFP + iPS cells ( Figure 3C,D). Similar to Surf4, OE of Hspa5 and sXbp1 reduced cell proliferation during the process ( Figure 3E). In contrast,  Figure 3F). These data strongly suggest that the response to ER stress, especially the IRE1-XBP1 pathway, is required for reprogramming.

| Response to ER stress mediates the reprogramming facilitation by Surf4
To further investigate the relationship between Surf4 and the UPR in the ER (UPR ER ), we examined the expression level of ER  Figure S2, Tables S1 and S2 stress-induced effectors at day 3 during reprogramming with or without Surf4. These effectors, such as Ddit3, Hsp5a and Atf4, were boosted by exogenous Surf4 at the RNA and protein levels ( Figure 4A,B). In the whole reprogramming process, these ER stressrelated genes exhibited transient increases at the early phase and the surge appeared at day 6, while Surf4 caused earlier increases ( Figure S3E), as the expression level of Surf4 was upregulated during reprogramming ( Figure S3C,D). These results suggested that Surf4 might facilitate reprogramming by activating UPR ER at an early phase.
We speculated that Surf4 may improve the efficiency of reprogramming by temporarily activating the response to ER stress. To test our hypothesis, we introduced Surf4 and sXbp1-ΔDBD at the same time in our reprogramming system to examine their function in iPSC  Figure S3 and Table S1 generation. OE of sXbp1-ΔDBD blocked the activation effect of Surf4 on reprogramming ( Figure 4C-F). This result suggested that the ability of Surf4 to enhance reprogramming relies on the IRE-XBP1 pathway.
It was recently reported that two small-molecule modulators of the UPR, salubrinal (Sal) and azoramide (Azo), enhanced reprogramming. 40 Sal selectively inhibits eIF2alpha dephosphorylation 41 and activates the PERK/eIF2α branch of the UPR pathway, 42 while Azo improves ER protein-folding ability and stimulates the expression of ER chaperones. 40 Therefore, we evaluated whether they can reverse the reduction in reprogramming efficiency caused by Surf4 KD. As shown in Figure S4A-C, the impaired reprogramming elicited by Surf4 KD was not reversed by these two activators, suggesting that activation of the PERK-eIF2α pathway cannot rescue the reprogramming efficiency that was decreased upon Surf4 KD. Thus, these results suggested that downstream of ER stress, the IRE1-XBP1 pathway mediated the effect of Surf4 in facilitating reprogramming.

| DISCUSS ION
Surf4 is enriched in mouse MII oocytes and zygotes 9 Figure S4 and Table S1 embryonic development. 25,26 We had previously found this maternal factor can promote somatic cell reprogramming, but its mechanism had not been elucidated. In this study, we demonstrate that Surf4 promotes reprogramming by activating the response to ER stress.
This activation may cause a transient increase in the expression of UPR-related genes, and the blockade of XBP1 impaired the effect of Surf4 on reprogramming.
It has been reported that Erv29p (homologous gene of Surf4 in S. cerevisiae) is involved in the degradation of soluble ER quality control substrates and is upregulated transcriptionally in response to ER stress. 11 It is possible excessive Surf4 may disturb the balance of protein transport flow and protein folding and processing in the ER, and subsequently, triggered the UPR ER at the early stage of reprogramming. Such UPR ER activation adapted to the stress, eventually restored of ER homeostasis or programmed cell death to protect the remaining cells.
The role of UPR-related genes in reprogramming was consistent with a recent study, that reported transient activation of the UPR ER is required for the acquisition of pluripotency. 43 In our study, we also observed that exogenous expression of an appropriate amount of Hspa5 or sXbp1 contributed to reprogramming. These UPR ER effectors, as the downstream of Surf4, were transiently activated at the early stage of reprogramming. However, over-high concentrations of ER stress inducers did not promote reprogramming owing to excessively reduced cell mount during the process. 'Hyperactivated ER stress' led to a decrease in reprogramming efficiency as a strong inducer of cell death. 40 During reprogramming, transient activation of the UPR ER (we prefer to term it UPR ER surge) at early phase is necessary and sufficient to promote reprogramming to somatic cells to a pluripotent state. 43 In our reprogramming system, UPR ER surge occurred at day 6 in control group, and Surf4 brought such surge at day 3, which is probably why Surf4 facilitates reprogramming. Although the expression levels of sXpb1 and Ddit3 were significantly lower than those of the control at day 6 ( Figure S3E), the increase in their expression levels on day 3 was sufficient to promote reprogramming.
Although the expression level of most ER stress-relative genes were transiently upregulated by exogenous Surf4, not each of these genes overexpression can promote reprogramming efficiency. We have tried to overexpress Ddit3, but it did not facilitate somatic reprogramming (data not shown). In contrast, sXbp1 and Hspa5 can significantly increase reprogramming efficiency. Furthermore, the activation effect of Surf4 on reprogramming can be blocked by sXbp1-ΔDBD, a dominant negative form of Xbp1. This means that the IRE-XBP1 signal plays an important role in reprogramming.
The mechanism through which the activation of IRE-XBP1 pathway, increases reprogramming efficiency remains to be elucidated.
The UPR mainly alleviate ER stress by increasing the amount of molecular chaperones (such as HSPA5), ER luminal space and other folding catalysts to restore homeostasis, or to initiates apoptosis. 44 IRE1 mediated adaptive events, such as activation of XBP1s to upregulated expression levels of target genes, ER-associated degradation (ERAD) of unfolded proteins, and IRE1-dependent decay (RIDD) of cytosolic mRNAs. 45 These adaptive remodelling to ameliorate imbalances in ER proteostasis may benefit to the somatic signature turn off and allow pluripotent network to be set.
Previous studies have implicated Erv29p in ER quality control and are transcriptionally upregulated upon ER stress. 11 In this study, we found that overexpression of Surf4 in turn activates UPR at early phase in reprogramming. Although the exact mechanism of UPR ER activation by Surf4 still needs to be investigated, activation of UPR ER can promote reprogramming of human somatic cells to a pluripotent state. 43 We supported that overexpression of SURF4 may improve the reprogramming efficiency of human cells.
At the molecular level, SURF4 can interact with STIM1 in the ER to modulate store-operated Ca 2+ entry (SOCE). 46 In the present study, during reprogramming, SOCE was found to be reduced gradually and was further reduced by Surf4 at the early stage (data not shown). Whether SOCE is another barrier to reprogramming needs further investigation.
In the mature oocyte, many nucleic acids (mainly RNA) and proteins accumulate, which constitute the maternal material for early embryonic development. These factors not only drive sperm or somatic nuclei into totipotent embryos but also augment the efficiency

CO N FLI C T O F I NTE R E S T
The authors declare no competing interests.
supervised the study and contributed to writing.

DATA AVA I L A B I L I T Y S TAT E M E N T
The sequencing data sets have been deposited in NCBI's Gene Expression Omnibus (GEO) and are accessible through the GEO accession number GSE17 6177.