Fifty-eight 50 ml samples of yeast cells (with an average of 4.2 x 108 cells) were harvested by centrifugation at 3,000 rpm for 5 minutes at 4 °C using a Sorvall Super T21 centrifuge. The supernatant was removed by vacuum aspiration, and the cell pellet was resuspended in 1 mL TRIzol reagent (an acidic solution containing guanidinium thiocyanate, phenol, and chloroform) and aliquoted in two tubes (each containing 0.5 ml sample). Samples were stored at −80 °C until further processing. One set of frozen yeast in TRIzol was sent to Novogene and the second set was processed at Stony Brook, see below.
Local (Stony Brook) Methods
For lysis, one set of frozen yeast in TRIzol was thawed on ice. Samples were transferred into 2 ml screw-cap tubes with an O-ring. Zirconium oxide beads (0.5 mm diameter) were added from a plastic weigh boat until the bead level reached approximately 1 mm below the liquid meniscus. Tubes were pre-chilled on ice to 0 °C, then loaded in a balanced configuration into an Omni International Bead Ruptor 12. Homogenization was performed at 3.1 m/s for 30 seconds.
After homogenization, tubes were placed on ice. A 1 µL sample (avoiding beads) was mixed with 6 µL of water on a microscope slide and inspected using phase-contrast microscopy to assess cell disruption. Typically, >85% of cells were lysed after a single cycle. If lysis is insufficient, a second 30-second cycle can be performed (Prolonged agitation is not recommended, as it may cause overheating.) No second cycle was needed for any of the samples considered here.
RNA Recovery and Processing
The cell lysate was transferred to a 1.5 mL Eppendorf tube using a pipette tip, leaving the zirconium oxide beads behind. Chloroform was added (20% of the TRIzol volume, i.e., 200 µL per 1 mL TRIzol), and the mixture was gently rocked for 3 minutes at room temperature. Samples were then centrifuged at 14,000 rpm for 15 minutes at room temperature in a tabletop centrifuge.
Phase separation resulted in three layers: a clear upper aqueous phase (containing RNA), a white interphase (containing DNA), and a red lower organic phase (containing proteins). The aqueous phase was carefully transferred to a 15 mL centrifuge tube, avoiding disruption of the interphase.
To the aqueous phase, 350 µL of Binding Buffer (from the Ambion RiboPure™ Yeast RNA Purification Kit, Cat. #AM1926) was added per 100 µL of solution. Ethanol (200-proof) was added to reach a final concentration of 70%, and the sample was then processed using the kit's filter cartridges, following the manufacturer's protocol. RNA was eluted in 2 × 25 µL of 10 mM Tris-HCl (pH 8.0).
RNA concentration was determined using a NanoDrop spectrophotometer.
Cell Lysis and RNA Recovery (Novogene)
The second set of frozen yeast in TRIzol was processed at Novogene. Per customer service at Novogene, their team used mechanical disruption using type C bead tubes from MN and a TissueLyser II by Qiagen operated at 30 Htz for 5 minutes per side and RNA may have been recovered using the RNeasy Mini kit-TRIzol. There was no mention of an assay for cell breakage.
Table 1 shows the number of cells in each sample, the theoretical amount of RNA in each sample, the amount of RNA recovered by Stony Brook or Novogene for each sample, the ratio of the amount of Novogene RNA to theoretical RNA; and the ratio of the amount of Novogene RNA to Stony Brook RNA.
Summary of Table 1 (per-sample data shown in the table): mean values across the 58 samples are reported as follows — Mean theoretical RNA: 143 µg; Mean Stony Brook (SB) RNA recovered: 49.4 µg; Mean Novogene RNA recovered: 0.71 µg; Mean Novogene/Theoretical ratio: 0.007; Mean Novogene/SB ratio: 0.0193.
"Cells (x 107)" is the number of cells in each sample.
"Theoret. RNA" is the theoretical amount of RNA in each sample, based on each cell containing 7.1 x 10-13 g RNA per cell [7]. However, this is likely a slight to moderate over-estimate in most of these cases, since cells in most of these samples were smaller than usual, due to the nature of the experiment that produced them.
"SB RNA" is the number of micrograms of RNA produced locally (i.e., at Stony Brook) from these samples. For SH8, SH30, and SH55, the sample was the residual cell suspension, much less than a full 0.5 ml sample.
"Novogene RNA" is the number of micrograms of RNA produced by Novogene from these samples.
"Novo/Theoret." is the ratio of the amount of RNA produced by Novogene to the theoretical amount of RNA.
"Novo/SB" is the ratio of the amount of RNA produced by Novogene to the amount of RNA produced at Stony Brook. For SH8, SH30, and SH55, this is "NA" because a full sample was not available at Stony Brook.
Since a yeast cell contains about 7.1 x 10-13 g RNA per cell [7], one expects an average of nearly 300 micrograms of total RNA per 50 ml sample. Since 0.5 ml was half the total sample, the theoretical expectation for the average amount of total RNA was about 150 micrograms (or somewhat less, given the small average cell size of elutriated cells).
Novogene reported an average recovery of only 0.71 micrograms of total RNA per sample, ranging from 0.03 micrograms to 2.63 micrograms (Table 1). This is only a small fraction of the ~150 micrograms of total RNA expected.
Our method yielded an average of 49.4 micrograms of total RNA per sample, compared to 0.71 micrograms of total RNA per sample at Novogene. When the yields in micrograms were compared between each sample by a paired t-test, the difference was significant with a p-value of 2 x 10-16.
Compared to our protocol, which achieved approximately 40% RNA extraction efficiency from S. cerevisiae, the method used by Novogene resulted in a substantially lower yield of around 1%.
This low efficiency could be because only about 1% of cells in the sample were lysed. Although there is no proof that inefficient lysis is the cause, there are two supporting arguments. First, the method possibly used by Novogene to lyse the cells would be expected to be inefficient (see Methods). Second, the yeast samples where Novogene was more successful (~ 3% efficiency) were samples that contained a relatively high proportion of large, aberrant cells, and the yeast samples where Novogene was less successful (~ 0.05% efficiency) were samples containing mainly small, round cells. Alternatively, the low efficiency could have been due to inefficient RNA recovery (see Methods).
Since only a few nanogram of RNA is enough to prepare library for RNA-Seq, so even if cell breakage was inefficient but completely random, then the ensuing RNA-Seq results could still be useful. However, if cell breakage occurred preferentially for large, aberrant cells, then the results from RNA-Seq would represent gene expression in these aberrant cells, and so would be misleading with regard to gene expression in the entire sample population. Thus it could be potentially misleading to sequence RNA samples when extraction efficiency is very low.
In the apparent absence of any quality control for the extraction step, it seems prudent for investigators to check the RNA yield for any kind of sample. This is a parameter that should be reported in publications.