Archive for category ResearchBlogging
Saccharomyces cerevisiae (hereafter simply referred to as “yeast”) is one of the most intensively studied eukaryotic organisms. In fact, it is probably one of the oldest domesticated ones, being used for beer brewing already in Sumeria and Babylonia around 6000BC and in old Egypt for dough leavening.
In the 1930s, it was recognized as an ideal model system in which to study several aspects of eukaryotic cell biology, competing with the by then very popular Neurospora crassa, (a filamentous fungus which we have discussed before due to its importance in the history of molecular biology, see here), and is now at the forefront of experimental molecular biology.
Many of the widely used methods in the field have been pioneered in yeast (gene expression microarrays, for example) and it continues to be the major proving ground for large-scale studies and a wellspring of new discoveries.
Yeast, a single-celled hemiascomycetes fungus, combines several advantages that makes it particularly suitable for cell and molecular biology studies, including rapid growth, short generation time, tractable genetics, well-known physiology, a variety of established vector and transformation systems, high-efficiency homologous recombination (which in itself could be used as a tool for the study of gene function in other organisms, see here) and a collection of KO mutants (see below), just to name a few. Notably, yeast was the first eukaryotic organism to have its genome sequenced1 and this information, along with all the experimental advantages that yeast possesses, have proven instrumental for the advancement of the study of gene function not only in yeast, but in other organisms as well, including humans.
As mentioned, a lot of what we now know about yeast comes from high-throughput studies that have been pioneered in yeast, and notably, many of them have been later on applied to other systems.
The study of gene function, a logical step after genome sequencing, encompasses a combination of several complementary approaches, including subcellular localization, protein-protein interactions, genetic interactions, gene expression profiles and KO phenotyping (just to name a few), and all of them can be (and have been) scaled up to study gene function globally in this organism.
In this two-post series on yeast, I will discuss a very clever methodology developed for the high-throughput study of KO strains in Saccharomyces cerevisiae. In this first post, I will talk about how KO strains are generated and in the second one, how they are studied globally. Through these methodologies, it’s possible to study which genes are important for growth under a variety of different conditions in this unicellular organism, which has helped us understand the role of many genes in yeast that we knew nothing about. The information derived from these studies has also helped us understand gene function in other organisms, including humans, and has had wide implications, for example, for the study of drug action.
The global approach I will discuss, pioneered by the laboratory of Ron Davis2, is based on tagging each KO strain and then using that tag to evaluate how the KO strains are doing under different growth conditions over time. This brief explanation will make a lot more sense once I describe how the KO strains are generated and how the importance of each gene for growth under different conditions, is scored.
In the interest of brevity, I won’t do a historical perspective of the methodology, but just describe the way the mutants were generated as part of the Yeast Deletion project and how they have been used.
A general scheme of the methodology for KO generation is shown on figure 1, which represents a “PCR-based gene deletion strategy”.
The idea of the Yeast Deletion Project is to generate a start- to stop- codon deletion of each of the ORFs in the yeast genome, by replacing it with a kanMX module (which confers resistance to G418).
In order to do this, you’ll need to target the kanMX module specifically to each locus. How is this specificity attained?
This is when two rounds of PCR kick in.
In the first round, you will attach something to the common kanMX module. The primers for this are long (74mers) and are divided into functional regions.
Primers are specifically designed to flank the start and stop codons of each ORF in the genome. The forward primer contains 18 bases of homology to the region upstream of the target ORF and the reverse one, 18 bases of homology to the region downstream of the target ORF (these are depicted as the “ATG” and “TAA” regions in the primers in figure 1). Appended to these sequences, there are unique 20-base sequence tags (labelled UPTAG and DNTAG in figure 1) and common priming sites (U1-U2 in the forward primer and D1-D2 in the reverse one). The importance of these sequences will be discussed below.
Note that the common tag priming sites U2 and D2 are homologous to the 5′ and 3’ regions of the kanMX4 module, respectively. This allows for the primers to anneal to the cassette for Round 1 of PCR. The importance of the U1 and D1 sequences will be discussed in the next post of the series.
After that round of PCR, you are left with a kanMX cassette that is flanked by specific tags and by sequences that flank a particular ORF in the genome.
A second round of PCR comes next, aimed at extending the cassette’s homology region to the yeast genome. This greatly increases the specificity in the homologous recombination step (see figure 1 and below).
By the end of the second round of PCR, you are left with a kanMX cassette flanked by specific tags and by sequences that will allow its specific integration in the genome.
Now, we are ready to transform yeast. The cassette is transformed into diploid yeast cells where, by homologous recombination, will replace one copy of the WT gene (the possibility of replacing both alleles in the same cell in one transformation event is extremely low).
Transformation is done into diploid strains in case the gene being knocked out is essential. Growth in media containing G418 is used to select for transformants and correct integration is later checked by PCR.
The resulting transformants are heterozygous diploid deletion strains (harboring one WT allele and one disrupted one). How do we get the homozygous deletion strains?
Simple. We just sporulate the diploids (induce diploids to go through meiosis).
The KO procedure described, when applied to a non-essential gene, produces four viable haploid spores: two contain the intact ORF and two contain the cassette.
Homozygous diploid strains, which are also available, are constructed from the mating of two independently isolated haploid mutants.
What about essential genes? Applying the KO procedure to these genes and inducing sporulation, yields only two viable spores, each containing the wild-type ORF (i.e. yeast haploid cells with the disrupted allele will die). In any case, this tells us something about that gene that we didn’t know: it is essential under those growth conditions. The methodology for essential genes then, only proceeds up to the heterozygous diploid stage
Note that “non-essential does not mean “without a function” or “non-important”: deletion of “non-essential” genes can also have a negative impact on fitness (resulting for example, from the imbalance of gene product), but under the conditions of selection used in the project, their deletion does not lead to lethality (although they could be sick) and homozygous deletion strains for these genes can be obtained.
The Yeast Knockout collection thus contains deletion strains in four different backgrounds: haploids of each mating type (there are two of them, “a” and “alpha”), homozygous diploids (for non-essential genes) and heterozygous diploids.
In this way, more than 90% of the ORFs larger than 100 amino acids as well as verified shorter ORFs have been disrupted, providing the research community with an invaluable tool for studying gene function.
As I mentioned before, yeast is particularly amenable to high throughput studies and many of the global approaches used in other organisms have been pioneered in yeast. How can we then study gene function globally in this organism? How can we evaluate which genes are important under different growth conditions on a high-throughput manner? One approach, which we will discuss in the next post of this series, makes use of the Yeast Deletion Collection, particularly, of the unique tag each strain contains.
1 Goffeau A, et al. (1996) Life with 6000 genes. Science 274 (5287): 546-567.
2 Shoemaker DD et al. (1996) Quantitative phenotypic analysis of yeast deletion mutants using a highly parallel molecular bar-coding strategy. Nat Genet. 14(4):450-6.
3 Chu AM, & Davis RW (2008). High-throughput creation of a whole-genome collection of yeast knockout strains. Methods in molecular biology (Clifton, N.J.), 416, 205-20 PMID: 18392970
The “Direct Connection” section at “The MolBio Hut” includes blog posts discussing primary research articles in the field, but these posts are written by the authors themselves. This allows them to discuss the background, results and implications of their work with a wider audience and in a more relaxed format. We hope that this direct link between the authors and the scientific community (hence its name), promotes discussion and interaction with scientists in other fields.
The intestinal epithelium has the highest self-renewal rate in our body (and in mammals in general). It has been estimated that the entire intestinal epithelial lining is renewed every 5-7 days (which essentially means that around 5 grams of cells are discarded every day). Tissue regeneration in the intestine is ultimately sustained by actively proliferating intestinal stem cells (ISCs) that reside at the base of mucosal invaginations called crypts of Lieberkühn (see Figure 1).These cells give rise to all intestinal cell lineages.
Colorectal cancer (CRC) is the second cause of death by cancer. It has been known for over 20 years that mutations that activate the Wnt signaling pathway, essential for intestinal homeostasis, can give rise to benign lesions called adenomas. Adenomas are not dangerous per se, but are the substrate for further mutations, a process that can lead to the development of CRC and eventually metastasis, which is the main cause of death in cancer patients.
The current therapeutic strategy for most CRC patients consists of surgical extraction of the tumor followed by preventive chemotherapy. Clinicians have been aware, however, for a very long time, that several patients relapse (cancer recurs in 30%–50% of all cases) and eventually develop tumors, usually in the form of metastasis.
The fact that there is disease relapse, means that there are cells that are able to survive for extended periods of time, even after chemotherapy, and resume growth and lead to the development of new (metastatic) tumors, which resemble the primary tumor.
Given the fact that these “relapse-causing cells” are long lived and able to regenerate whole tumors, we sought out to find whether there was a relationship between cancer recurrence and intestinal stem cells. In this “Direct Connection”, I will describe our work entitled “The Intestinal Stem Cell Signature Identifies Colorectal Cancer Stem Cells and Predicts Disease Relapse”, published earlier this year on Cell Stem Cell1.
The first thing we did was analyze the transcriptome of the various populations of intestinal cells in order to study the differences between the ISCs and their progeny.
By using fluorescence-activated cell sorting (FACS), we were able to separate the different populations of cells based on the surface expression of EphB2, a tyrosine-kinase receptor whose expression is highest in the cells at the base of the crypt (in ISCs) and gradually decreases towards the villi (where differentiated cells are, see Figure 1). We then globally analyzed gene expression on these populations by using microarrays and determined genes whose expression was specifically restricted to ISCs, proliferating cells or differentiated cells.
Using these “gene signatures”, we analyzed several cohorts of colorectal cancers and basically found two main things:
- Late stage cancer has an expression profile more similar to ISCs than early stage cancer.
Tumors are divided in stages which reflect how advanced they are (based on size and invasiveness). This classification is called the American Joint Cancer Committee [AJCC] staging system. We found that tumors that are in the latter stages (most invasive and dangerous) display high expression of an ISC-like program, while early stage tumors do so to a much lower extent.
2. Tumors that relapse have a higher expression of ISC-specific genes than tumors that don’t.
Further, tumors that display expression profiles resembling the ones from proliferating cells have a lower incidence of relapse (which makes sense, since chemotherapy is highly effective against these cells).
Basically, the data appeared to indicate that most aggressive CRCs expressed high levels of ISC-specific genes and that relapsing tumors had an overall ISC-like phenotype.
What we didn’t know, however, was whether the ISC gene expression program was characteristic of all cells in the tumor or restricted to specific subpopulations. As Lgr5, a G protein-coupled receptor whose function has just been reported2, is to date the best ISC marker, we used it to identify the localization of tumor cells with an ISC-like phenotype. We also investigated the expression of intestinal differentiation markers in CRC samples by analyzing the expression of Krt20, a widely described marker of intestinal cell differentiation. Notably, we found that both Lgr5+ and Krt20+ cells were present in tumors, but their expression patters were mutually exclusive (i.e tumor cells expressed either Lgr5 or Krt20).
Further, as we found these cell types in very close proximity to one another, we considered unlikely that they were derived from independent tumor subpopulations. With this in mind, we postulated that advanced CRCs are organized in a hierarchical fashion, reminiscent of the normal intestinal epithelium, in which we can find cells with two main mutually exclusive phenotypes: ISC-like or differentiated cell-like phenotypes.
To functionally validate the relevance of these different phenotypes, we once again turned to EphB2. We investigated whether EphB2 expression could distinguish between ISC-like and differentiated-like cells in CRCs, just as it does in the normal intestinal mucosa. Indeed, we saw that tumor EphB2-expressing cells were also enriched in the expression of ISC genes, while the majority of cells expressing markers of differentiation (i.e Krt20+ cells) were EphB2 negative.
To further test our model, we sought to determine the tumor forming capacity of ISC-like and differentiated-like tumor cells. We purified epithelial tumor cells expressing high, medium, or low surface EphB2 levels and injected them into immune-deficient mice. Notably, we found that EphB2 positive populations retained the capacity to generate tumors with high efficiency (i.e. are enriched in tumor initiating cells), whereas EphB2 negative populations displayed reduced or null tumorigenic capacity. In case you are wondering, the EphB2med cells showed an intermediate behavior. Further, the EphB2 derived tumors recapitulated the organization of the tumor of origin. We concluded that “ISC-like tumor cells hold high tumor-initiating potential as well as display long-term self-renewal and differentiation capacity”.
Overall, we described the transcriptional landscape of normal intestinal populations and showed that the risk of developing recurrent CRC is proportional to the expression of ISC-specific genes. Furthermore, we showed that colorectal tumors are organized in a hierarchical structure, similar to that present in normal crypts, and that the ISC-like cells within these tumors have tumor-initiating and self-renewal capacity.
With all this evidence, we postulate that CRC follows a “cancer stem cell” model of growth (See “Cancer Stem Cells: the root of all evil?“).
As with any research project, there are several open questions of which I’ll only highlight three:
Is the enrichment in late stage cancer of the ISC signature due to increased expression of these genes by a small subset of cells? Or is there an expansion of the ISC-like population?
Which ISC genes are actually relevant for tumor behavior? Which are only markers?
Will the targeting of cancer stem cells actually improve the treatment of cancer? (I think this is the most relevant question in the field of cancer stem cells)
Finally, I´d like to thank my PhD supervisor Eduard Batlle and the post-doc who led this project, Anna Merlos-Suárez, for the opportunity to work on this project and for all the support and help during my PhD.
I hope that before my PhD is over I’ll get to write another article for the “Direct connection” section explaining the project I’m currently working on.
-Francisco M. Barriga
1Merlos-Suárez A, Barriga FM, Jung P, Iglesias M, Céspedes MV, Rossell D, Sevillano M, Hernando-Momblona X, da Silva-Diz V, Muñoz P, Clevers H, Sancho E, Mangues R, & Batlle E (2011). The intestinal stem cell signature identifies colorectal cancer stem cells and predicts disease relapse. Cell stem cell, 8 (5), 511-24 PMID: 21419747
2de Lau, W., Barker, N., Low, T.Y., Koo, B.-K., Li, V.S.W., Teunissen, H., Kujala, P., Haegebarth, A., Peters, P.J., van de Wetering, M., et al. (2011). Lgr5 homologues associate with Wnt receptors and mediate R-spondin signalling. Nature advance online publication. PMID: 21727895