Glenn K. Fu

Combinatorial labeling of single cells for gene expression cytometry

Single-cell expression analysis on a large scale To understand why cells differ from each other, we need to understand which genes are transcribed at a single-cell level. Several methods measure messenger RNA (mRNA) expression in single cells, but most are limited to relatively low numbers of cells or genes. Fan et al. labeled each mRNA molecule in a cell with both a cellular barcode and a molecular barcode. Further analysis did not then require single-cell technologies. Instead, the labeled mRNA from all cells was pooled, amplified, and sequenced, and the gene expression profile of individual cells was reconstructed based on the barcodes. The technique successfully revealed heterogeneity across several thousand blood cells. Science, this issue 10.1126/science.1258367 A simple approach allows gene expression analysis of a large collection of cells by stochastic barcoding of cells and sequencing. INTRODUCTION The measurement of specific proteins and transcripts in individual cells is critical for understanding the role of cellular diversity in development, health, and disease. Flow cytometry has become a standard technology for high-throughput detection of protein markers on single cells and has been widely adopted in basic research and clinical diagnostics. In contrast, nucleic acid measurements such as mRNA expression are typically conducted on bulk samples, obscuring the contributions from individual cells. Ideally, in order to characterize the complexity of cellular systems, it is desirable to have an affordable approach to examine the expression of a large number of genes across many thousands of cells. RATIONALE Here, we have developed a scalable approach that enables routine, digital gene expression profiling of thousands of single cells across an arbitrary number of genes, without using robotics or automation. The approach, termed “CytoSeq,#x201D; employs a recursive Poisson strategy. First, single cells are randomly deposited into an array of picoliter wells. A combinatorial library of beads bearing cell- and transcript-barcoding capture probes is then added so that each cell is partitioned alongside a bead. The bead library has a diversity of ~106 so that each cell is paired with a unique cell barcode, whereas the transcript barcode diversity is ~105 so that each mRNA molecule within a cell becomes specifically labeled. After cell lysis, mRNAs hybridize to beads, which are pooled for reverse transcription, amplification, and sequencing. Because cDNAs from all polyadenylated transcripts of each cell are covalently archived on the bead surface, any selection of genes can be analyzed. The digital gene expression profile for each cell is reconstructed when barcoded transcripts are assigned to the cell of origin and counted. RESULTS We applied CytoSeq to characterize complex heterogeneous samples in the human hematopoietic system by examining thousands of cells per experiment. In addition to surface proteins that are traditional cell type markers, we examined genes coding for cytokines, transcription factors, and intracellular proteins of various cellular functions that may not be readily accessible by flow cytometry. We demonstrated the ability to identify major subsets within human peripheral blood mononuclear cells (PBMCs). We compared cellular heterogeneity in resting CD3+ T cells versus those stimulated with antibodies to CD3 and CD28, as well as resting CD8+ T cells versus those stimulated with CMV peptides, and identified the rare cells that were specific to the antigen. Highlighting the specificity of large-scale single-cell analysis compared with bulk sample measurements, we found that the up-regulation of a number of genes in the stimulated samples originated from only a few cells (<0.1% of the population). CONCLUSION The routine availability of gene expression cytometry will help transform our understanding of cellular diversity in complex biological systems and drive novel research and clinical applications. The massively parallel single-cell barcoding strategy described here may be applied to assay other biological molecules, including other RNAs, genomic DNA, and the genome and the transcriptome together. Gene expression cytometry (CytoSeq). Massively parallel, stochastic barcoding of RNA content from single cells in a microwell bead array enables digital gene expression profiling of thousands of single cells simultaneously. Shown here is the principal component analysis for human PBMCs. Each point represents a single cell. Cells with correlated expression profiles are coded with similar colors. We present a technically simple approach for gene expression cytometry combining next-generation sequencing with stochastic barcoding of single cells. A combinatorial library of beads bearing cell- and molecular-barcoding capture probes is used to uniquely label transcripts and reconstruct the digital gene expression profile of thousands of individual cells in a single experiment without the need for robotics or automation. We applied the technology to dissect the human hematopoietic system and to characterize heterogeneous response to in vitro stimulation. High sensitivity is demonstrated by detection of low-abundance transcripts and rare cells. Under current implementation, the technique can analyze a few thousand cells simultaneously and can readily scale to 10,000s or 100,000s of cells.

Counting individual DNA molecules by the stochastic attachment of diverse labels

We implement a unique strategy for single molecule counting termed stochastic labeling, where random attachment of a diverse set of labels converts a population of identical DNA molecules into a population of distinct DNA molecules suitable for threshold detection. The conceptual framework for stochastic labeling is developed and experimentally demonstrated by determining the absolute and relative number of selected genes after stochastically labeling approximately 360,000 different fragments of the human genome. The approach does not require the physical separation of molecules and takes advantage of highly parallel methods such as microarray and sequencing technologies to simultaneously count absolute numbers of multiple targets. Stochastic labeling should be particularly useful for determining the absolute numbers of RNA or DNA molecules in single cells.

Molecular indexing enables quantitative targeted RNA sequencing and reveals poor efficiencies in standard library preparations

Significance RNA sequencing (RNA-Seq) is a common tool for measuring relative gene expression levels. However, as an absolute quantitative tool, the data are prone to various distortions due to biases from library preparation. We improve the quantitative aspects of RNA-Seq by barcoding individual cDNA molecules to correct for amplification bias, distinguish clonal replicates, and obtain absolute measurements of gene expression. We have also developed a set of barcoded synthetic RNAs that can be spiked into samples as easy-to-use quantitative controls. Additionally, we demonstrate the combined use of capture enrichment with molecular barcoding for the sequencing of targeted genes. These results demonstrate low library preparation efficiency leading to the stochastic loss of low-abundance transcripts, which cannot be overcome by simply increasing sequencing depth. We present a simple molecular indexing method for quantitative targeted RNA sequencing, in which mRNAs of interest are selectively captured from complex cDNA libraries and sequenced to determine their absolute concentrations. cDNA fragments are individually labeled so that each molecule can be tracked from the original sample through the library preparation and sequencing process. Multiple copies of cDNA fragments of identical sequence become distinct through labeling, and replicate clones created during PCR amplification steps can be identified and assigned to their distinct parent molecules. Selective capture enables efficient use of sequencing for deep sampling and for the absolute quantitation of rare or transient transcripts that would otherwise escape detection by standard sequencing methods. We have also constructed a set of synthetic barcoded RNA molecules, which can be introduced as controls into the sample preparation mix and used to monitor the efficiency of library construction. The quantitative targeted sequencing revealed extremely low efficiency in standard library preparations, which were further confirmed by using synthetic barcoded RNA molecules. This finding shows that standard library preparation methods result in the loss of rare transcripts and highlights the need for monitoring library efficiency and for developing more efficient sample preparation methods.

Human papillomavirus DNA in malignant and hyperplastic prostate tissue of black and white males

This study hypothesized that human papillomavirus (HPV) infection is associated with increased prostate cancer risk, and that the 40% higher incidence rate in blacks is attributable to a greater prevalence of oncogenic viral DNA in prostatic tissues. Viral L1 and E6 gene sequences were polymerase chain reaction (PCR) amplified in archival tissues from 56 prostate cancer cases and 42 hyperplastic controls. L1 amplimers were hybridized by dot blot to HPV L1 generic probes, as were E6 amplimers to E6 probes specific for HPV 6, 11, 16, 18, 31, 33, and 45. 12.5% of cases and 9.5% of controls were HPV positive by L1 hybridization (age/race adjusted odds ratio = 1.66, 95% confidence interval = 0.33, 8.37). Four of 52 (7.7%) blacks were HPV positive compared to 7 of 46 (15.2%) whites. However, none of the L1‐positive samples hybridized to the E6 type‐specific probes, and positive results were not replicable using a broader spectrum of PCR primers and probes. These data suggest that HPV infection is not a significant risk factor for prostate cancer and does not explain the excess cancer risk in blacks.

Digital Encoding of Cellular mRNAs Enabling Precise and Absolute Gene Expression Measurement by Single-Molecule Counting

We present a new approach for the sensitive detection and accurate quantitation of messenger ribonucleic acid (mRNA) gene transcripts in single cells. First, the entire population of mRNAs is encoded with molecular barcodes during reverse transcription. After amplification of the gene targets of interest, molecular barcodes are counted by sequencing or scored on a simple hybridization detector to reveal the number of molecules in the starting sample. Since absolute quantities are measured, calibration to standards is unnecessary, and many of the relative quantitation challenges such as polymerase chain reaction (PCR) bias are avoided. We apply the method to gene expression analysis of minute sample quantities and demonstrate precise measurements with sensitivity down to sub single-cell levels. The method is an easy, single-tube, end point assay utilizing standard thermal cyclers and PCR reagents. Accurate and precise measurements are obtained without any need for cycle-to-cycle intensity-based real-time monitoring or physical partitioning into multiple reactions (e.g., digital PCR). Further, since all mRNA molecules are encoded with molecular barcodes, amplification can be used to generate more material for multiple measurements and technical replicates can be carried out on limited samples. The method is particularly useful for small sample quantities, such as single-cell experiments. Digital encoding of cellular content preserves true abundance levels and overcomes distortions introduced by amplification.

Purification of the pets factor. A nuclear protein that binds to the inducible TG-rich element of the human immunodeficiency virus type 2 enhancer.

The peri-ets (pets) site is a TG-rich element found immediately adjacent to two binding sites for the ets family member Elf-1 in the human immunodeficiency virus type 2 (HIV-2) enhancer. Enhancer activation in response to T cell stimulation by phorbol myristate acetate, phytohemagglutinin, soluble or cross-linked antibodies to the T cell receptor, or antigen is mediated through this site in conjunction with its two adjacent Elf-1 binding sites, PuB1 and PuB2, and a κB site. Site-specific mutation of the pets element significantly reduces inducible activation of this enhancer but does not affect its transactivation by HIV-2 tat or other viral transactivators. Similar TG-rich sequences adjacent to ets-binding sites have also been found to be functionally important in the human T-cell leukemia virus type I and murine Moloney leukemia virus enhancers. As the cellular factor binding to the pets site plays a significant role in regulating the HIV-2 enhancer in both T cells and monocytes, we have purified this protein from bovine spleens and demonstrate that it is 43 kDa in size. In addition, using glycerol gradient centrifugation, Southwestern blotting, electrophoretic mobility shift assays employing purified protein eluted from a gel, and a new in solution UV cross-linking competitive assay, we show that the dominant protein binding to the pets site is 43 kDa in size. These results indicate that a nuclear protein of 43 kDa binds specifically to the pets site of the HIV-2 enhancer and may mediate transcriptional activation of this important human pathogen in response to T cell stimulation. As retroviruses generally expropriate important human regulatory proteins for their own use, the 43-kDa pets factor is also likely to play a significant role in signal transduction in T cells and in other cellular processes.