PropertyValue
is nif:broaderContext of
nif:broaderContext
is schema:hasPart of
schema:isPartOf
nif:isString
  • The OP6 cell line was cultured at 33 °C in Dulbecco's Modified Eagle's Medium (DMEM, Life Technologies) supplemented with 10% fetal bovine serum (Gibco), as described previously [29]. For RNA FISH, cells were seeded on 22cm2 coverslips coated with 0.1% gelatin (Sigma) in a 6 well plate at about 50% confluency and expanded for one day until near confluency. For colony RNA FISH, cells were seeded at ~2,000 cells per slide and grown for 7–8 days (50 cell colonies) or at ~10,000 cells per slide (4 cell colonies), and grown for 2–4 days. Long intron probes for some RNA FISH experiments were synthesized using Long Range polymerase (Qiagen) with sequence-specific primers (see S1 Table) and incorporation of DIG-16-dUTP (Sigma) into PCR products. We utilized intron probes for OR RNA FISH for three reasons: (i) we can design longer intron than exon probes (enhanced sensitivity); (ii) for genes (like ORs) expressed at low levels, unprocessed RNAs at the native locus are more spatially concentrated than processed RNAs in the cytoplasm (enhanced sensitivity); (iii) the one-spot (monoallelic) nuclear signal is an important validation of an OR signal (enhanced specificity). For most RNA FISH experiments, the long-intron PCR products (PCR primer sequences are provided in S1 Table) were cloned into pCRII-TOPO vector (Invitrogen); prepared plasmids were linearized prior to in vitro transcription using SP6 or T7 polymerases (Roche) for production of sense- or antisense-specific probes with incorporation of DIG-16-dUTP or Biotin-16-dUTP (Sigma). 100 ng of labeled probe was combined with 5 μg Cot1-DNA (Invitrogen) and 10mg salmon sperm DNA (Sigma) in a 2 ×SSC, 10% dextran sulfate solution, and heat denatured. For two-color RNA FISH colony experiments, 20ug of E. coli tRNA (Sigma) and 50ug of BSA (Sigma) were added to reduce background. Cells were permeabilized with 0.5% Triton-X in CSK buffer, fixed with 4% paraformaldehyde in PBS, and dehydrated in a 70%–80%–95%–100% ethanol series. Probe and cells were incubated overnight at 37°C in a humidified chamber. Following washes (maximum stringency = 50% formamide, 0.5 ×SSC at 37°C), samples were blocked for subsequent antibody incubations (4% BSA, 4 ×SSC, 0.2% Tween-20). DIG signals were visualized using sheep anti-DIG FITC (11207741910, Roche) and donkey anti-sheep FITC (sc-2476, Santa Cruz Biotech) antibodies, at a 1:100 dilution in 1% BSA, 4 ×SSC, 0.2% Tween-20. Biotin signals were visualized using avidin-DCS rhodamine (A2012, Vector Labs), followed by biotinylated anti-avidin antibody (Ab73235, Abcam) plus an additional incubation with the avidin-DCS rhodamine, each used at 1:100 dilution in 1% BSA, 4 ×SSC, 0.2% Tween-20. For OR re-selection assays, we used 6 long-intronic DNA probes against 40 small colonies for each probe (240 total colonies; 975 total cells were screened) and 6 larger colonies each (36 total colonies; 2,497 total cells were screened). In addition, we conducted two-color RNA FISH experiments using the Olfr920 (labeled with DIG) and Olfr57 (labeled with biotin) probes in order to investigate whether small colonies are able to activate more than one OR gene. For measuring expression frequencies in a well-defined lineage, we used sense/antisense RNA probes against 9 OR genes on 28 total cultures (2–5 replica cultures per probe). Images were acquired using a Deltavision RT imaging system (Applied Precision) adapted to an Olympus (IX71) microscope equipped with XYZ motorized stage. Each image was sectioned with 0.5 μm intervals to ensure complete coverage of the nucleus. ImageJ (Fiji) was used for analysis of positive cells. Method considerations for OR profiling: We considered various methodologies for profiling OR representation in OP6 cell populations. We estimate that OR mRNA yield from a positive OP6 cell is roughly in the 10–100 template range based on relative expression levels of abundant ORs from previous qPCR experiments. For example, if we assume that frequently represented OR genes are expressed in ~1–2% of OP6 cells (as observed in RNA FISH experiments herein, see text), then observed actin:OR cDNA ratios in full OP6 cell populations would be ~50-100-fold greater than the actin:OR cDNA ratio in single positive cells. The observed actin:OR ratio for a set of commonly expressed OR genes (e.g., Olfr920, Olfr544, and Olfr57) averaged ~6,500:1 (not shown), suggesting ~65-130-fold actin excess in single OR-positive cells. We approximate the actin transcript abundance at ~103−104 molecules per cell [34–36], and therefore estimate the OR transcript abundance to be between ~8 mRNAs (assuming ~103 actin mRNAs at ~130-fold excess relative to OR mRNA) and ~150 mRNAs (assuming ~104 actin mRNAs at ~65-fold excess relative to OR mRNA). These estimates are consistent with OR template numbers for relatively low-abundance genes in typical cells. With this low OR transcript abundance in mind, we decided against using RNA-seq for three reasons: (1) Assuming a positive OP6 cell contains at most ~100 OR transcripts, and estimating ~105 total RNA transcripts per cell [37, 38], we reasoned that the putative median OR in the population (expressed in ~1/1,000 OP6 cells) might be <1 transcript per million transcripts from an OP6 population (or, <1 FPKM). This is commonly the noise cutoff threshold used in RNA-seq to account for false-positives in alignment or other technical limitations [39, 40]. (2) In order to obtain a reliable number of sequence reads per OR (e.g., >100 hits per gene), we would need to sequence to a depth >100 million reads per sample, which was not financially feasible when considering our goal of characterizing numerous OP6 populations within several well-defined lineages. (3) The cDNA preparation protocols commonly utilized prior to sequencing involve competitive enrichment (e.g., polyA isolation, PCR-based amplifications), which potentially introduces methodological biases that are particularly skewed against low-copy number transcripts [41, 42]. Obviously, we did not want potential methodological biases to obscure our interpretation of apparent OR expression biases. To achieve the desired sensitivity while mitigating potential amplification bias, we developed a nested PCR strategy detailed in the following section. We optimized cDNA preparation and nested PCR protocols to identify a condition for each OR gene tested that (a) reliably and reproducibly gave robust products on gDNA diluted to <100 templates, and (b) reliably and reproducibly did not generate products in no-RT controls for various RNA preparations from OP6 populations. We were able to develop nested PCR assays that satisfy the above criteria for 21 OR genes. Each OR gene was investigated in cell populations across various OP6 cell lineages. Approximately 5x106 OP6 cells per culture were harvested and RNA was extracted using Trizol (Thermo Fisher/Life Technologies). Approximately 5 μg of RNA was treated with DNase (Thermo Fisher/Ambion) and further purified using the RNeasy Mini Kit (Qiagen). Approximately 500 ng of resulting RNA was subjected to first-strand cDNA synthesis by SmartScribe reverse transcriptase (Clontech), followed by PCR using multiplexed OR primer pairs for 20 cycles. A second nested PCR reaction was conducted for the 21 OR genes (individual reactions, not multiplexed) with cycle numbers optimized to report 50–100 OR templates from known gDNA quantities, while maintaining cleanly negative results in no-RT controls from various OP6 cell populations. All PCR primer sequences are provided in S1 Table. The <100 template sensitivity threshold was chosen because we estimate that this is approximately the OR template yield from one positive OP6 cell (see above). In our analyses of various OP6 cell populations, we input unamplified cDNA corresponding to the equivalent of ~400 and ~2,000 OP6 cells, thus reporting ORs that we estimate are expressed in >1/400 (>0.25%) and >1/2000 (>0.05%) cells, respectively. Therefore, these two sensitivity thresholds should permit surveying the middle portions and upper half of a normal distribution that models OR frequencies for the null hypothesis in which all ORs are represented by a completely stochastic selection process.
rdf:type