igh and eGFPlow hESCs was performed on a DiVa flow cytometer with DiVa software. Approximately 200.000 cells were sorted from each subpopulation and used for mRNA extraction. Re-analysis after cell sorting confirmed that the purity of the sorted populations was.95%. Semi-quantitative and quantitative polymerase chain reaction Total RNA was isolated from hESCs or embryoid bodies using the Genelute total RNA kit. RNA was digested with DNAseI and reverse transcribed to cDNA using Superscript III reverse transcriptase. Semi-quantitative PCR was performed with Taq-polymerase. Primers, cycle numbers and annealing temperatures for semi-quantitative PCR are listed in supplementary Text S1. PCR products were separated on agarose gels in presence of ethidium bromide and analyzed under UV light. For quantitative real-time PCR, reactions were performed with Platinum quantitative PCR SuperMix in the presence of SYBR green. PCR cycles on a 7900HT fast real time PCR system were performed as follows: 50uC/ 2min, 95uC/2min, 95uC/15sec, 60uC/25sec, 73uC/30sec. Gene expression levels were normalized to endogenous GAPDH expression and AG-1478 web quantified using the DDCt method. Primers used for quantitative realtime PCR are listed in supplementary Text S1. Gene expression arrays Gene expression arrays were performed with TaqMan Human Stem Cell Pluripotency Array plates according to the manufacturer’s instructions. Briefly, cDNA from FACS-sorted NANeG subpopulations was combined with gene expression mastermix and loaded onto array plates containing gene-specific primers and probes. Gene expression was measured on a 7900HT fast real time PCR system with the cycle conditions 50uC/ 2min, 95uC/10min, 95uC/15sec, 60uC/1min. Gene expression levels were normalized to the threshold cycle values of endogenous GAPDH and quantified using the DDCt method. Copy-number determination by quantitative PCR The copy number of the NANOG targeting vector in NANeG cells was determined by quantitative PCR following the guidelines for assay design, controls and evaluation given in. A fragment of the NANOG proximal promoter was amplified and quantified relative to the copy number of the single-copy reference genes GDF3 and FOXJ2. Primers used for qPCR reactions are shown in supplementary Text S1. PCR cycles on a 7900HT fast real time PCR system or a Step1Plus real time PCR system were performed as follows: 50uC/2min, 95uC/2min, 95uC/15sec, 60uC/25sec, 73uC/30sec. Measurements on all samples were repeated in at least three qPCR experiments which were performed in triplicates. A standard curve with serial dilutions of hESC genomic DNA ranging from 0.78 to 25ng/reaction in 2fold dilution steps was prepared by plotting mean threshold cycle values against log-transformed gDNA concentrations. A linear trend line was fitted into each standard curve and slopes and correlation coefficients were obtained in Microsoft Excel. Assay efficiencies were calculated using the formula E = 10. Melt curve analysis and absence of amplification products from water controls confirmed the specificity of the assays. Copy numbers of the NANOG promoter fragment in test samples were normalized to reference assays and compared to untransfected hESCs using the amplification efficiency-adjusted DDCt method with the formula: Results Gene targeting of the NANOG locus in hESCs The NANOG targeting vector used for this study was created by bacterial artificial 
chromosome recombineering . BAC recombineering is a fast and efficie
