Defined conditions for propagation and manipulation of mouse embryonic stem cells

ABSTRACT The power of mouse embryonic stem (ES) cells to colonise the developing embryo has revolutionised mammalian developmental genetics and stem cell research. This power is vulnerable, however, to the cell culture environment, deficiencies in which can lead to cellular heterogeneity, adaptive phenotypes, epigenetic aberrations and genetic abnormalities. Here, we provide detailed methodologies for derivation, propagation, genetic modification and primary differentiation of ES cells in 2i or 2i+LIF media without serum or undefined serum substitutes. Implemented diligently, these procedures minimise variability and deviation, thereby improving the efficiency, reproducibility and biological validity of ES cell experimentation.


INTRODUCTION
Mouse embryonic stem (ES) cells are cell lines derived from the preimplantation epiblast of mouse embryos (Boroviak et al., 2014;Brook and Gardner, 1997;Evans and Kaufman, 1981;Martin, 1981). Under appropriate culture conditions they retain the properties of their tissue of origin and can re-enter normal development when introduced into morula-or blastocyst-stage embryos (Bradley et al., 1984). Moreover, ES cells can readily be genetically modified and clonally expanded. Advances such as genome editing using CRISPR/Cas9 have expanded the opportunities for multiplex and/or complex genome engineering in ES cells (Andersson-Rolf et al., 2017;Yang et al., 2013). Their properties make ES cells uniquely powerful tools, both for generating genetically modified mice, and for experimental dissection of fate choice in pluripotent cells. These attributes depend entirely, however, upon the genetic and phenotypic fidelity of ES cells during propagation.
ES cells were first derived in 1981 by culturing early mouse embryos in conditions optimised for teratocarcinoma stem cells (Evans and Kaufman, 1981;Martin, 1981). The capacity of ES cells to contribute to chimaeras, colonise the germline and engender healthy offspring was demonstrated in 1984 (Bradley et al., 1984), establishing that they are non-transformed and, in all essential aspects, genetically normal. This finding provoked a major effort to introduce targeted genetic modifications into mice by implementing homologous recombination in ES cells. It was not until the 1990s, however, that the technology became relatively routine (for a review, see Capecchi, 2005). A major reason for the time gap is that ES cell cultures were frequently found to be aneuploid, particularly after clonal selection. This problem gradually diminished as appreciation spread of the relatively fastidious demands of ES cell culture compared with other cell types (Robertson, 1987). In particular, ES cells must be passaged frequently to avoid any overgrowth, which confers advantage to genetically abnormal cells.
The original culture conditions for ES cells comprised co-culture with a feeder layer of mitotically arrested embryonic fibroblasts and medium containing carefully screened foetal calf serum (FCS) (Robertson, 1987). This effective but complex system was simplified with the discovery that a major contribution of feeders is to provide the cytokine leukaemia inhibitory factor (LIF) (Smith et al., 1988;Williams et al., 1988). Addition of LIF increases the robustness of ES cell cultures on feeders and this remains a widely used system. LIF can also support ES cell derivation and propagation without feeders in either serum or bone morphogenetic protein (BMP) (Nichols et al., 1990;Ying et al., 2003a). In these conditions, however, the cultures are morphologically heterogeneous. Moreover, the cells show fluctuating expression of several transcription factors known to be expressed in the pre-implantation epiblast and downregulated during peri-implantation development (Chambers et al., 2007;Hayashi et al., 2008;Toyooka et al., 2008). Thus, ES cells in these culture conditions do not correspond to a discrete stage of embryonic pluripotency and the developmental relevance of metastable gene expression remains unclear (Filipczyk et al., 2015;Nichols and Smith, 2012;Smith, 2017). In contrast, ES cells appear morphologically and molecularly relatively homogeneous when maintained in defined medium in which the Erk1/2 signalling pathway is blocked and glycogen synthase kinase 3 is partially inhibited (Wray et al., 2010;Ying et al., 2008). Under this dual inhibition, known as 2i, ES cells exhibit transcriptome similarity to pre-implantation epiblast (Boroviak et al., 2014). Importantly, unlike other conditions, use of 2i (or its predecessor 3i) enables reliable derivation of authentic ES cells from different strains of mice and also from another species, the rat (Buehr et al., 2008;Kiyonari et al., 2010;Li et al., 2008;Nichols et al., 2009a).
We have proposed that the defined culture system using the highly specific 2i inhibitors corrals ES cells in a stable 'ground state' . The 2i platform has been widely exploited to study signalling, gene regulation and network control of naive pluripotency, and the transition path from pluripotency to lineage commitment (Blaschke et al., 2013;Carey et al., 2014;Dunn et al., 2014;Ficz et al., 2013;Habibi et al., 2013;Hackett et al., 2017;Hayashi et al., 2012Hayashi et al., , 2011Kalkan et al., 2017;Kumar et al., 2014;Leitch et al., 2013a;Marks et al., 2012;McEwen et al., 2018 preprint;Mulas et al., 2017;Murakami et al., 2016;Semrau et al., 2017). Importantly, male ES cells maintained in 2i can retain a euploid karyotype and germline chimaera competency over multiple passages with similar efficiency to cells cultured in serum Nichols et al., 2009a;Kiyonari et al., 2010;Leitch et al., 2010Leitch et al., , 2013bLeeb et al., 2012;Morgan et al., 2013;Jakubczik et al., 2016;Kalkan et al., 2017;Zhang et al., 2018). Interestingly, ES cells show lower global DNA methylation in 2i than in serum (Ficz et al., 2013;Habibi et al., 2013;Leitch et al., 2013a). Female ES cells are vulnerable to severe hypomethylation (Choi et al., 2017a) and loss of methylation imprints in either 2i or serum, which can reduce their chimaera contribution (Yagi et al., 2017). Male ES cells on the other hand can retain relatively normal methylation imprints ( Fig. 1), although loss of specific differentially methylated regions (DMRs) has also been reported in both serum and 2i cultures (Choi et al., 2017b;Dean et al., 1998). Genetic background may be one component influencing loss of imprinted DMRs. Another important factor may be levels of culture stress or stimulation due to differences in media composition and cell handling between laboratories, even when using apparently similar conditions. As with any cells in vitro, ES cells are inherently liable to acquire genetic and epigenetic adaptations or abnormalities if subjected to untoward selective pressures. Vulnerability to environmental stress may be enhanced in serum-free culture. Consequently, apparently minor variations in media formulation or culture procedures may give rise to discrepant findings. Reports of subpopulations and cell cycle heterogeneity in 2i culture (Kolodziejczyk et al., 2015) may likewise be influenced by particular culture practices.
The field could benefit, therefore, from a stringent methodology for ES cell culture. To that end, we detail here standardised media composition and cell handling procedures for robust propagation and genetic manipulation of mouse ES cells using 2i in defined medium.

RESULTS AND DISCUSSION Propagation of ES cells without serum factors or feeders Aim
The aim of this procedure is to expand undifferentiated cells in 2i or 2i/LIF . The anticipated outcome is actively growing colonies of uniform size that are evenly distributed across the dish, with no signs of differentiation, and that can be passaged by enzymatic dissociation and expanded continuously. Note that in defined medium cells are generally less tolerant to suboptimal media conditions or environmental perturbations than in the presence of serum, serum substitutes such as KnockOut Serum Replacement (KSR), or feeders. This susceptibility may manifest as spontaneous differentiation and/or cell death. Cells should not be kept out of the incubator any longer than necessary or exposed to varying incubator environments.

and 37°C
Protocol: routine passage of ES cells from a 6-well plate 1. Coat plates or wells with 0.1-0.2% gelatine in the incubator for a minimum of 15 min. 2. Pre-warm necessary volume of 2i (±LIF) or other culture medium, wash medium and Accutase to 37°C. Note: avoid leaving 2i medium or Accutase at 37°C for too long. 3. Remove all gelatine from plates or wells and replace with warm medium. Return to the incubator to pre-equilibrate (not necessary but might help survival, especially for ES cells of non-permissive strains, compromised mutants or when plating at low density). Drying the plates is not necessary. 4. Aspirate medium gently but completely from cells and add 0.5 ml of Accutase per well. Avoid drying the cells. 5. Incubate with Accutase (see Table 1) at room temperature for 4-6 min until colonies decompact and detach. Tap plate to ensure detachment and initial dissociation. 6. Add 1 ml of wash buffer and pipette up and down (without touching the bottom of the dish) 10-20 times in order to obtain a single cell suspension. Try to minimise the formation of bubbles. Tip: check under the microscope to ensure single cell suspension. 7. Transfer cell/Accutase/wash suspension to a Falcon tube containing 6 ml of wash medium (see Table 1). 8. Centrifuge cells at 300 g for 3.5 min. 9. Aspirate supernatant with care, removing as much liquid as possible without disturbing the pellet. 10. Re-suspend cell pellet by pipetting up and down 10-15 times in 0.5-2 ml of 2i (±LIF) medium, ensuring a single cell suspension is obtained. 11. Count number of cells/ml and plate appropriate number of cells (see Table 1). Slide plate back and forth across a flat surface to distribute cells evenly, then place carefully in incubator. Do not disturb for several hours.

Notes
Cell density will have a significant effect on metabolism, cell cycle and differentiation kinetics, amongst other factors. This is, therefore, a key parameter to monitor. We have observed that culture at high density or colony overgrowth can compromise ability to differentiate and that this effect may become irreversible. Overgrowth of cells in FCS/LIF conditions is also associated with impaired capacity of differentiation and karyotypic instability, but in defined conditions ES cells must be passaged at smaller colony sizes than in FCS-or KSR-containing medium. A single period of overgrowth may induce a permanent change, even if undifferentiated morphology is retained. Specifically, even if overgrown cells regain refractile domed morphology after passaging, they may be compromised. Two key parameters should be kept relatively constant: (1) the density at which cells are plated after each split; (2) the frequency with which cultures are split. Representative images of cultures ready to be passaged are shown in Fig. 2.
The following cell concentrations are routinely used: for maintenance of cells, 1.5-3.0×10 4 cells/cm 2 ; for most experiments: 1.5×10 4 cells/cm 2 (see below for differentiation). See also Table 1. It is highly recommended to count cell numbers at every split to avoid overgrowing cells and to monitor growth rate (see Fig. 2 for representative images).
In routine culture, cells should be split every 2-3 days (Fig. 2). Healthy cultures double every 12-14 h (Carey et al., 2014). ES cells in 2i have a substantial proportion of cells in G1 phase ( Fig. 3) (Huurne et al., 2017). This is in contrast to FCS cultures, in which ∼70% of cells are in S phase (Huurne et al., 2017). Even if plated at lower/clonal density, colonies should not be allowed to grow for more than 4-5 days before passaging.
If passaging is delayed beyond 2 days, medium should be renewed on day 3, and any day thereafter.
The quality of N2B27 must be monitored. Signs of suboptimal N2B27 include flattening of colonies, cells detaching, reduced proliferation rate, or increased cell death.
ES cells from certain genetic backgrounds (e.g.C57BL/6) require 2i/LIF in order to be stably propagated long term, whereas others, such as 129 strains, can be propagated in 2i alone or in single inhibitor with LIF. LIF invariably increases colony formation after low-density plating. The addition of LIF alters the kinetics of differentiation, however, delaying the process by ∼12 h (Nett et al., 2018). Certain mutant cell lines (e.g. those carrying Etv4/5 gene deletions), are sensitive to MEK inhibition and show more robust proliferation in CH/LIF (LIF containing CHIR99021) (Kalkan et al., 2019). To convert cells between 2i, CH/LIF and PD/LIF (LIF containing PD0325901), two passages over 4-6 days are sufficient. PD/LIF cells typically show a more flattened morphology than those grown in conditions containing CH. Because culture in 2i/LIF is most robust, cells can be kept in 2i/LIF for routine passaging, and transferred to other conditions (e.g. 2i or PD/LIF) for one or two passages before experimental analyses. The cell densities reported above work well for all medium conditions. Cell lines can either be derived in defined conditions (see below) or adapted from conventional FCS/LIF (with or without feeders) conditions. When converting cells from FCS/LIF, we recommend plating cells first in FCS/LIF and changing media to 2i or 2i/LIF after 24 h. Cell death, mainly associated with elimination of differentiating cells, is often observed for the first few days, but stable homogeneous cultures can be established within one or two passages.
Cells should be routinely maintained without any antibiotics, in keeping with good tissue culture practice (Freshney, 1994). Antibiotics can mask low level microbial infections and may have unknown effects on cell metabolism and gene expression.
Mycoplasma testing should be carried out frequently and on all newly generated or obtained lines using PCR assays and commercial kits. If mycoplasma is detected, the cultures should be discarded (Markoullis et al., 2009).
It is important to maintain two types of cell line stock. Primary stocks are the lowest passage cells available [ passage ( p) 4-8 for embryo-derived cell lines], that have been genotyped, mycoplasma screened, have at least ∼80% diploid cells (40XY or 40XX) and show uniform morphology. These are kept in liquid nitrogen for long-term storage. Secondary stocks are more numerous at slightly higher passage number ( p10-15). These stocks can be kept short term at −80°C and are used for routine experiments. Cells should not normally be passaged beyond 30. However, we have generated high contribution chimeras that give germline transmission after multiple passages of mouse ES cells ( p30+), embryonic germ cells ( p15+), and haploid ES cells ( p20+) in 2i/LIF. Accutase is routinely used because it is gentler than trypsin, results in a single cell suspension and does not require serum inactivation. TrypLE reagents may also be used, although these occasionally result in incomplete dissociation. Trypsin can be used, but we recommend adding ∼5% FCS to the wash medium (see Table S1) to ensure complete inactivation, or doubling the volume of wash medium, or increasing the BSA concentration.
Attachment of poorly adherent cells, a feature of some mutants and non-129 strains (e.g. CBA, NOD, DBA, etc.), can be improved by plating onto laminin-coated plates. For this, coat plates with laminin solution (∼10 µg/ml in PBS) for a minimum of 2 h at 37°C, before aspirating the solution and adding the culture medium. DMEM/F12 and Neurobasal buffer can be used at 5-10% CO 2 . We use 7% CO 2 for compatibility with FCS cultures maintained in GMEM. In the presence of B27 we have not observed any advantage of low O 2 compared with ambient atmosphere.

Colony-formation assay Aim
The colony-formation assay is a key functional test in ES cell biology. Cultures are dissociated to a single cell suspension and plated at very low (clonal) density, and the number of colonies that form is determined after several days. This assay can be used as a diagnostic tool for media quality control, and experimentally for determining the proportion of undifferentiated cells in a differentiation or reprogramming study or following a genetic/ chemical perturbation, or in a modified culture environment. See Fig. 3 for an example.

Materials
Accutase 10 µg/ml laminin in PBS Wash medium 2i/LIF (see Table S1 for media formulation and suggested suppliers) FCS+LIF Haemocytometer Tissue culture treated plates Falcon tubes Centrifuge Humidified incubator at 7% CO 2 and 37°C Alkaline phosphatase staining kit Protocol 1. Coat 12-well plates ( provide three wells per sample) with laminin (∼10 µg/ml in PBS) for a minimum of 2 h at 37°C. At clonal density, laminin helps colonies remain attached for the duration of the experiment. 2. Pre-warm Accutase, wash medium, and culture medium (e.g. 2i/LIF). 3. Before starting, aspirate laminin and add 1 ml of 2i/LIF per well. Return plate to the incubator in the meantime to equilibrate medium. 4. Aspirate medium from cells and quickly add appropriate volume of Accutase. 5. Incubate with 1 volume of Accutase at room temperature (RT) for 4-6 min (1 min after the colonies have detached). 6. Add 1 ml of wash medium and pipette up and down 10-20 times in order to obtain a single cell suspension. Tip: check under the microscope to ensure single cell suspension. 7. Transfer cell/Accutase/wash suspension to a Falcon tube containing 5 volumes of wash medium. 8. Centrifuge cells at 300 g for 3.5 min. 9. Aspirate supernatant with care, removing as much liquid as possible without disturbing the pellet. 10. Re-suspend cell pellet in 0.5-2 ml of culture medium (e.g. 2i/LIF). 11. Count number of cells/ml. 12. Plate 400 cells/well in prepared medium. If necessary, carry out two sequential 1:10 dilutions of the cell to improve accuracy, and aim to pipette ∼50 µl of cell suspension per well. 13. Slide plate back and forth across a flat surface to distribute cells evenly, then place carefully in incubator. 14. After 4-5 days, perform alkaline phosphatase staining according to manufacturer's instructions, let dry, image plate and count the number of colonies.

Notes
The use of 12-well or 6-well plates is advisable. Plate 400 cells in 12-well plate, and 800 in 6-well plates. It is recommended to perform at least three technical replicates per condition as high variability can occur when plating small numbers of cells.
Fluorescence-activated cell sorting of cells directly onto plates can also be used to reduce technical variability. Colonies can detach easily; therefore, add fixative gently. Leaving cells for >5 days will increase the likelihood of colony detachment.
We also recommend performing colony-formation assays in parallel FCS/LIF medium and analysing plates by alkaline phosphatase staining on day 3/4. Monitor cultures to make sure colonies do not merge. In FCS/LIF conditions, differentiating cells, which would not survive in 2i/LIF, will persist and contribute to mixed or wholly alkaline phosphatase-negative colonies. This experiment can serve as a control for plating efficiency and differentiation potential. Fig. 4 shows representative images of colony assays in 2i/LIF and FCS/LIF.
Typically, we expect wild-type cells maintained in 2i or 2i/LIF to have a clonogenic capacity of ∼70-80% when plated in 2i/LIF in optimal media (all alkaline phosphatase positive) with minimal differentiation.
It is recommended to analyse multiple time points. Typically, ∼25-36 h after withdrawal of self-renewal conditions offers a time window during which both delayed and accelerated transition can be evaluated, as ∼50% of the wild-type cell population should remain responsive to 2i/LIF and positive for Rex1GFPd2. The exact timing for each experiment will depend on the starting culture condition (e.g. 2i versus 2i/LIF), the batch of N2B27, cell plating density, etc.

Neural differentiation Aim
Monolayer neural differentiation is a simple and wellcharacterised system (Ying et al., 2003b) and, under the right conditions, an efficient one. Therefore, it can be used to determine competence for differentiation and to examine gene expression dynamics.
2. Aspirate laminin (do not wash) and add appropriate volume of N2B27. Return to the incubator to pre-equilibrate. Note: pre-equilibration is not necessary but it helps survival. 3. Split cells, re-suspend in N2B27 and count them. 4. Plate directly in N2B27 onto laminin-coated plate at a density of 1.0×10 4 cells/cm 2 or 1.2×10 4 cells/cm 2 for flow-sorted cells. Optional: adding 1 µg/ml laminin directly to N2B27, before plating cells in laminin-coated plates, can help with adhesion. 5. Change media to fresh N2B27 on day 2 and every day thereafter.
Alternative protocol for poorly adherent cells 1. Coat plates with laminin (∼10 µg/ml in PBS) overnight at 37°C. 2. Aspirate laminin (do not wash) and add appropriate volume of 2i (+LIF). Return to the incubator to pre-equilibrate. Note: pre-equilibration is not necessary but it helps survival. 3. Split cells as late as possible in the day and plate in 2i or 2i/LIF onto laminin-coated plate at a density of 1.0×10 4 cells/cm 2 or 1.2×10 4 cells/cm 2 for sorted cells. Optional: adding 1 µg/ml laminin directly into N2B27 can help with adhesion. 4. As early as possible on the next day, gently wash cells with PBS before changing the media to N2B27. 5. Change media to fresh N2B27 on day 2 and every day thereafter.

Notes
Little or no death should be observed until cells become confluent (Fig. 6B). Significant death at day 3 is a sign of poor-quality media (see 'Batch testing N2B27' section) or incorrect plating density. Certain lines may require further optimisation of plating density to minimise cell death. ES cells derived from certain strains might show more cell death compared with 129 or mixed strains. When troubleshooting differentiation, we recommend plating cells at three different cell densities (e.g. 0.75×10 4 cells/cm 2 , 1.0×10 4 cells/cm 2 and 1.5×10 4 cells/cm 2 ), monitoring cell death at day 3-4 and quantifying Sox1-positive cells at day 4-5. Efficient conditions typically yield ∼90% Sox1-positive cultures on day 5 with little non-neural differentiation (Fig. 6C). The efficiency and timing of differentiation can be assessed by using the Sox1GFP reporter cell line (Ying et al., 2003b) or by measuring the expression of different neural markers such as Sox1 protein by immunofluorescence over time (Fig. 6A,C). When starting from 2i (no LIF) cells, the following changes in transcription factor expression can be anticipated: day 1, predominantly Oct4 (Pou5f1) positive, Sox2 positive, Sox1 negative; day 3, predominantly Sox2 positive (cells should be either Oct4 or Sox1 positive, as their expression is mutually exclusive), 30-50% Sox1GFP-positive cells; day 4-5, mostly Sox1positive cells and 80-90% Sox1GFP-positive cells. Tuj1 (Tubb3)positive cells should appear on day ∼6. Differentiation of cells cultured in 2i+LIF will be delayed by ∼1 day.
Low cell density is essential for efficient differentiation (Fig. 6A). If starting from 2i+LIF or if cells have to be pre-plated before initiating differentiation, there will be more proliferation before exiting pluripotency. Cell density can be lowered to 0.8×10 4 cells/cm 2 .
The quality of N2B27 should be monitored. Batches should be specifically tested for neural differentiation. Certain mouse strains benefit from high insulin N2B27 (see Table S1).
Failure to change media regularly, or plating cell density higher than 1.2×10 4 cells/cm 2 , can result in mixed differentiation (most obviously indicated by cells undergoing spontaneous contractions) and in persistence of undifferentiated ES cells.

CRISPR/Cas9 mutagenesis Aim
CRISPR/Cas9 (Cong et al., 2013;Jinek et al., 2012) allows for fast and efficient targeted mutagenesis of ES cells. As per routine ES cell culture, it is important to avoid overgrowing cells as they might develop phenotypes unrelated to the genetic perturbation. It is advantageous to use clones that go through the targeting process but that have not been edited as control cell lines. Depending on the plasmid used, transfected cells can be selected either using drug selection or by sorting for the expression of GFP.

Materials
Accutase 0.1-0.2% gelatine Wash medium FuGene gRNA vectors FCS (optional) N2B27, 3 µM CHIR99021, LIF (hereafter referred to as CH/LIF; alternatively use 2i/LIF; see Table S1 for media formulation and suggested suppliers) Penicillin and Streptomycin (PenStrep) Puromycin (depending on the strategy) Haemocytometer Tissue culture treated 3.5 cm diameter plates or 12-well plates and 96-well plates for single clone picking Falcon tubes Centrifuge Humidified incubator at 7% CO 2 and 37°C Cell sorter (depending on the strategy)  9. Ensure even distribution of cells by sliding the dish across a flat surface vertically and horizontally. 1. Design gRNAs and clone into nickase plasmid [ pSpCas9n(BB)-2A-Puro (PX462) V2.0; Addgene plasmid #62987]. The distance between the gRNA target sequences should be kept <200 bp for efficient insertion. 2. Different approaches exist for designing targeting vectors. In our hands, homology arms of 1.5 kb for the 3′ end, and 3 kb for the 5′ end work efficiently, but shorter homology arms might be used. Both gRNA target sequences should be absent in the targeting vector sequence to avoid cleavage. The amount of targeting vector should be minimised to avoid random integration. 3. A 6-well plate format should be used with a cell concentration of 1-2×10 5 cells and a final volume of 1.6 ml before adding transfection reagent. 4. Transfection mixes should be prepared as follows. Mix A: 900 ng gRNA 1, 900 ng gRNA 2, 200 ng targeting vector, 400 μg CH/LIF; Mix B: 6 μl FuGene, 400 μl CH/LIF.

Single colony picking and expansion
Typically, 12-24 clones are picked per transfection and this is sufficient to obtain at least two targeted lines. Each transfection will require 300,000 cells per well (round dish with 3.5 cm diameter and a low edge to facilitate colony picking). Include one extra well as a control.
Transfections in N2B27/CHIR99021/LIF (CH/LIF) result in better attachment and improved survival during selection when targeting some genes but in general 2i/LIF is effective. The addition of 1% FCS can help cell adhesion, especially after picking clones, but it is not necessary for most lines.
For efficient knockout, we routinely use two gRNA constructs per gene and FuGene HD transfection reagent (Promega, E2311). Two genes can be targeted at the same time by transfecting four gRNA constructs. In such cases, the amount of FuGene HD can be doubled.
Include a well containing only FuGene and no gRNA as a control for drug selection or cell sorting.
Do not let primary colonies overgrow as this will increase the chances of cell lines becoming compromised.

siRNA knockdown in ES cells Aim
Efficient transfection of siRNA for knockdown of single or multiple genes in cells grown in serum-free conditions.

Accutase
Wash medium 2i (±LIF) (see Table S1 for media formulation and suggested suppliers)

Notes
Transfect Rex1::GFPd2 cells with control or GFP siRNA and analyse by flow cytometry after overnight incubation. An acceptable transfection efficiency is >90% (Fig. 7). Note that this control is sensitive because GFPd2 protein is destabilised with a half-life of ∼2 h. PenStrep-or serum-containing media cannot be used as this inhibits transfection.
Use the same media to prepare the transfection mix and culture cells (best results in 2i, without LIF).
Do not leave the cells in the transfection media for more than ∼16 h or they will die. Cells must be actively replicating; using confluent cells (large colonies) will decrease the transfection efficiency.

ES cell derivation Aim
Derivation of ES cell lines from individual embryos, in serum-free conditions Nichols and Jones, 2017;Nichols et al., 2009b).

Materials
KSOM medium (Millipore, MR-106-D) or Blast (Origio, 83060010) M2 medium (Millipore MR-015-D) 2i/LIF medium in N2B27 (see Table S1 8A) and place into pre-equilibrated 2i/KSOM for 1-2 days until embryos reach the blastocyst stage (Fig. 8B). Note: Blast media (Origio) can be used as an alternative, but N2B27 is not suitable until after blastocyst cavitation. 2. Prepare and pre-equilibrate a fresh organ culture dish containing 2i/LIF in N2B27 (LIF improves success rate, but is not essential) in the central well and PBS in the outer well. 3. Transfer embryos (now at blastocyst stage) to the new dish. 4. Incubate for 1 or 2 more days, depending on when the embryos were transferred to N2B27. A total of 3 days in culture is optimal (Fig. 8C). 5. On day 3 of culture, pre-equilibrate an organ culture dish containing N2B27+20% anti-mouse serum and three dishes containing 400 µl N2B27 in the incubator. Also, gelatinise a 96-well plate. After 20 min, aspirate gelatine and add 200 µl 2i/LIF in N2B27. Leave in the incubator to pre-equilibrate. 6. If the embryos have not hatched, remove the zona pellucida by use of acid Tyrode's solution (Fig. 8D). 7. Place a drop (∼300 µl) of Tyrode's solution on a sterile flat dish. 8. Transfer embryos still contained within the zona pellucida to Tyrode's solution with minimal carry over and monitor  Cell survival during differentiation Although most batches will enable self-renewal of ES cells, differentiation requires specific attention. 1. Follow neural differentiation protocol above and examine cells on day 3-4. 2. Good batches of N2B27 show little cell death until cells reach confluence. Death on day 2-3 is evident in poor batches.

Troubleshooting tips
Some useful troubleshooting tips are presented in Table 2.

Conclusions
In this paper, we have aimed to provide an accessible and comprehensive set of step-by-step protocols, including all media formulations, to facilitate robust and standardised manipulation of mouse ES cells in defined conditions. We provide guidelines for quality control and troubleshooting tips. The indicated reporter cell lines for calibration are available from the authors. We highlight key parameters that may perturb ES cell behaviour and genetic integrity. In particular, we emphasise the requirement for timely passaging to avoid overgrowth of colonies. Feedback on these protocols and recommendations for further improvement are welcome. Residual dissociation media present. After spinning cell suspension with Accutase and wash buffer, make sure no residue remains. Alternatively, increase wash buffer volume. If using trypsin, add 5% FCS to wash buffer. Cells were overgrown/underfed; mistreatment for one passage is enough to affect cells permanently.
It is recommended to thaw a new vial of cells.
Increase gelatine concentration to 0.2% or add small amount to BSA to medium.
Colonies are of different sizes Incomplete dissociation of colonies. Ensure single cell suspension when splitting cells. Verify under the microscope after pipetting to dissociate and before transferring to tube with 5× wash buffer. Merging of colonies due to poor distribution.
Ensure cells are evenly distributed before allowing them back in the incubator. Avoid circular motions of the plate and open/close incubators gently until cells have settled.
Flattened/differentiated colonies Cells grown at too high a density or starved. Although morphology can be recovered after plating at single cell density, the damage to cells is normally irreversible. The main effect of overgrowing cells is that they can no longer differentiate efficiently. Thawing a new vial of cells is recommended. Poor quality N2B27.