RNA integrity and RNA fragmentation

After RNA isolation and before following experiments, I know that RNA quality so called RNA integrity is so important. So why there is a RNA fragmentation kit? Do we need RNA intact or fragmented? (RNA will be used for biotinylation and RT-qPCR afterwards)

This paper uses a similar technique to you by purifying total RNA, synthesis of double-stranded cDNA containing a T7 promoter sequence, cDNA purification, in vitro transcription in the presence of biotin-labeled ribonucleotides and this transcription is possible with T7 RNA polymerase since the cDNA contains the T7 promoter region, and finally fractionation of RNA. This then allows target biotin-labeled RNA for short (~ 25 nucleotides) oligonucleotide arrays. The T7 polymerase performs a linear amplification, and the in vitro RNA obtained reflects the abundance of each transcript in the initial RNA.

Hence I assume you follow a similar protocol although I'm not sure why you do RT-qPCR on the fragmented RNA since that is unlikely to work if you are doing it using primers for specific regions as everything is fragmented. So I think you do a RT-qPCR before the fragmentation to test the expression levels of certain (known) genes (to see if you are getting the expected results) before you go ahead with fragmentation and look at the global RNA expression. If my assumptions are wrong, then it would be great if you added further details of your experimental protocols with the different steps and the exact orders of the steps followed and the results assessed and the experiments performed so that I can modify my response.

Integrity of SRP RNA is ensured by La and the nuclear RNA quality control machinery

The RNA component of signal recognition particle (SRP) is transcribed by RNA polymerase III, and most steps in SRP biogenesis occur in the nucleolus. Here, we examine processing and quality control of the yeast SRP RNA (scR1). In common with other pol III transcripts, scR1 terminates in a U-tract, and mature scR1 retains a U4-5 sequence at its 3' end. In cells lacking the exonuclease Rex1, scR1 terminates in a longer U5-6 tail that presumably represents the primary transcript. The 3' U-tract of scR1 is protected from aberrant processing by the La homologue, Lhp1 and overexpressed Lhp1 apparently competes with both the RNA surveillance system and SRP assembly factors. Unexpectedly, the TRAMP and exosome nuclear RNA surveillance complexes are also implicated in protecting the 3' end of scR1, which accumulates in the nucleolus of cells lacking the activities of these complexes. Misassembled scR1 has a primary degradation pathway in which Rrp6 acts early, followed by TRAMP-stimulated exonuclease degradation by the exosome. We conclude that the RNA surveillance machinery has key roles in both SRP biogenesis and quality control of the RNA, potentially facilitating the decision between these alternative fates.

© The Author(s) 2014. Published by Oxford University Press on behalf of Nucleic Acids Research.


Mutations in TRAMP and exosome…

Mutations in TRAMP and exosome components lead to accumulation of truncated scR1. (A–F)…

Altering Lhp1p expression affects scR1…

Altering Lhp1p expression affects scR1 levels in cells deficient in SRP core proteins,…

The scR1 3' end is altered in the absence of Lhp1 and in…

Trf4, Rrp6 and Rrp44 associate…

Trf4, Rrp6 and Rrp44 associate with scR1. Sequence reads mapped to scR1 were…

ScR1, but not SRP proteins,…

ScR1, but not SRP proteins, accumulate in the nucleolus of cells over-expressing Lhp1.…

ScR1 accumulates in the nucleolus…

ScR1 accumulates in the nucleolus of cells lacking TRAMP or exosome activities. Yeast…

Interactions of Lhp1 and RNA…

Interactions of Lhp1 and RNA surveillance factors with scR1. A model integrating findings…

Fragmentation of tRNA in Phytophthora infestans asexual life cycle stages and during host plant infection

Background: The oomycete Phytophthora infestans possesses active RNA silencing pathways, which presumably enable this plant pathogen to control the large numbers of transposable elements present in its 240 Mb genome. Small RNAs (sRNAs), central molecules in RNA silencing, are known to also play key roles in this organism, notably in regulation of critical effector genes needed for infection of its potato host.

Results: To identify additional classes of sRNAs in oomycetes, we mapped deep sequencing reads to transfer RNAs (tRNAs) thereby revealing the presence of 19-40 nt tRNA-derived RNA fragments (tRFs). Northern blot analysis identified abundant tRFs corresponding to half tRNA molecules. Some tRFs accumulated differentially during infection, as seen by examining sRNAs sequenced from P. infestans-potato interaction libraries. The putative connection between tRF biogenesis and the canonical RNA silencing pathways was investigated by employing hairpin RNA-mediated RNAi to silence the genes encoding P. infestans Argonaute (PiAgo) and Dicer (PiDcl) endoribonucleases. By sRNA sequencing we show that tRF accumulation is PiDcl1-independent, while Northern hybridizations detected reduced levels of specific tRNA-derived species in the PiAgo1 knockdown line.

Conclusions: Our findings extend the sRNA diversity in oomycetes to include fragments derived from non-protein-coding RNA transcripts and identify tRFs with elevated levels during infection of potato by P. infestans.


Sample preparation and data basis

Total RNA was obtained from various tissues and different organisms mainly human, rat, and mouse. All samples were analyzed on the Agilent 2100 bioanalyzer. For the development of the algorithm a large set of data files was kindly provided by the Resource Center for Genome Research [6] as well as by Agilent. The overall number of samples in the database totals 1208. About 30% of the samples are of known origin from human, mouse and rat extracted from liver, kidney, colon, spleen, brain, heart and placenta. The origin of the remaining samples was not traceable, but is known to be of mammalian cells or cell lines. For the development of the algorithm, it was important to include samples of all degradation stages into the database. The final data set included many intact as well as almost completely degraded RNA samples (cf. table 1 for the distribution of samples). Partially degraded RNA samples were less common but still sufficient in number. Furthermore, the data set comprised different sample concentrations and different extraction methods. To some extent anomalies were found in the data set as well. This provided a realistic collection of input data containing a representative basis for all stages of RNA degradation.

Applying our method described below to the data basis yields a sorted list of features, which was used to construct feature spaces for training regression models. Furthermore, results are given for models based on features proposed in the literature. Finally, we show the correlation of the RIN with the outcome of real-time PCR experiments.

Feature selection

The total RNA ratio was selected as first feature covering 79% of the entropy of the categorical values. The next two features contribute information about the 28S-region: 28S peak height and the 28S area ratio.

The fourth feature compares the 18S and 28S area to the area of the fast region. Feature 5 is the value of a linear regression at the end point of the fast region, whereas the next feature reflects the amount of detected fragments in the fast region. Then, the presence or absence of the 18S peak is selected, which enables the model to distinguish between weaker and stronger degradation. The last feature gives the relation of the overall mean value to the median value. Since the mean value is sensitive to large peaks it carries information about totally degraded RNA or about abnormalities like spikes. Table 2 summarizes the results of the feature selection process. An interpretation of these features from a biological point of view is given in the discussion and an overview of all features is given in the additional files [see Additional file 2].

Model training

Based on the sorted list of features we trained neural networks as regression models and systematically increased the number of hidden neurons from 0 to 8, until the model evidence decreased clearly. Furthermore, we varied the feature space as described in the previous section. We observed maximal model evidence using 5 to 7 features with 2 to 5 neurons in the hidden layer. The values are averaged over the results of a 10-fold cross-validation procedure (Fig. 3).

Evidence-based model selection. Dependency of model evidence on a logarithmic scale from the number of features used and from the degree of non-linearity in the hidden layer. The values are average values over a 10-fold cross-validation procedure. The highest evidence is reached for models with 5 to 7 input features and 2 to 5 hidden neurons respectively. All these models have a low generalization error below 0.25.

As expected, the model evidence is strongly negatively correlated with the generalization error (ρ = -0.93), which shows that the model evidence is a sensible model selection criterion (Fig. 3 and 4). We selected the topology using 5 features and 4 hidden neurons as the most probable model and performed the final training on the whole training data set. The log value for the evidence of the final model was slightly higher compared to the values during cross validation (-74 vs. -100), whereas the generalization error was stable (MSE of 0.26). The cross validation error was observed to be a good estimate for the generalization error on the test data.

Generalization errors. Dependency of the generalization error of the model from the number of features used and from the degree of non-linearity in the hidden layer. The values are average values over a 10-fold cross-validation procedure. Models with highest evidence (5 to 7 input features and 2 to 5 hidden neurons respectively) have a low generalization error below 0.25.

The feature selection procedure provides in each step the local optimal additional feature, which will not necessarily lead to the globally best combination. In the later iteration steps, several candidate features provide the same gain in information about the target and there is some randomness in the final selection. Explorative searching for the best combination is intractable because of the computational costs of the combinatorial search. In an additional, manual optimization step application knowledge was used to substitute some features by plausible alternatives. Feature 3 and 4 were replaced by the area ratio in the fast region (fast.area.ratio). Additionally the marker height was selected. In the normalized electropherogram, the marker height allows to detect strongly degraded samples, because it is the only part of the signal which differs from the background noise. This combination also has a relative MI value of 0.83, but the best model with 5 hidden neurons had a log value for the evidence of -42. It reaches a cross-validation error of 0.25 and a test error of 0.25, which is slightly better compared to the results from figure 4. Both models perform equally well, the later one was chosen for the final implementation in the expert software for the sake of simplicity.

Finally, we evaluated regression models for a subset of 400 samples on two different feature spaces: the 28S/18S ratio, and the feature computed by the degradometer software [1]. Table 3 shows that the RIN model is based on a feature space, with higher information content than the other two models. Model evidences indicate that using a single feature results in a lower posterior probability of the model. This is again consistent with the generalization performance of the models. The error of RIN model is forty times lower compared to 28S/18S-model and about twenty times lower compared the degradometer based model if all samples are considered.

If the samples that are labeled BLACK by the degradometer software are removed from the data set (N = 186, 42%), the relative MI value increases to 0.63, the evidence reaches a value of -121, whereas the cross validation error is at 0.60, which is still four times higher than for the RIN model.

Model evaluation

If a model is supposed to extract a relationship from experimental data, it is helpful for the model evaluation to explore the data in the most important two dimensions, as well as to check for large error values which correspond to categorical misclassifications. Furthermore, the model prediction can be cross-checked against control parameters of follow-up experiments, like RT-PCR.

Visualization of decision boundaries

The 2D-visualization of total RNA ratio and 28S peak height shows, that we can clearly separate high integrity values from low integrity values. The categories form clusters in this space. However, as mentioned in the previous section, the borderline between categories is not sharp, which is due to the fact that degradation is a continuous process.

Categorical misclassifications

Simple features like the ribosomal ratio which cover only one aspect of the degradation process tend to have larger errors for certain groups of experiments. That is, they cannot distinguish very well between the categories. It is very useful to check, that only a few experiments are interchanged over more than one categorical border, that is, the model covers all aspects of the degradation process. Misclassifications are measured by Receiver Operating Characteristics (ROC, cf. [7]) for distinguishing electropherograms from different groups of categories, whereas the value of the area under the ROC-curve (AUC: area under the curve) is a balanced measure for the classification error.

We briefly and informally describe how a ROC curve is constructed. The electropherograms are sorted into a ordered list according to the integrity measure estimated by the model. For a fixed classification task, a ROC curve is constructed as exemplarily described in the following for the task of distinguishing electropherograms of category 10 from all other categories: walk through the sorted list in descending order. In each step check if the considered item is in category 10 or not according to the original expert label. If it is true, draw a line fragment in vertical direction, if it is false draw it in horizontal direction. Perfect separation of category 10 from the others would imply that the ROC curve shoots up vertically on the y-axis to the maximal value, before the first horizontal step is taken. Random assignment of electropherograms would result in a ROC curve that corresponds to a diagonal line from the origin to the right top corner. A ROC curve gives a balanced measure of the model performance by integrating over all possible classification borders. Each border corresponds to a specific ratio of sensitivity to specifity, i.e. a specific point on the ROC curve.

Several electropherograms are interchanged between the adjacent categories 9 and 10 (AUC 0.96), which is natural due to the noise in label assignment step. Very rarely are assignments from electropherograms from category 10 to category 8 or less (AUC 0.98). Only 1 electropherogram is interchanged between category 7 and category 10 (AUC 0.999, cf fig. 5).

Receiver operating characteristics of categorical misclassifications. The figure shows the Receiver Operating Characteristics for distinguishing electropherograms of category 10 against the set union of other categories. The area under the curve (AUC) gives a measure of classification performance. Random assignment is equal to an area of 0.5, whereas perfect assignment is equal to an area of 1.0. Only few experiments are exchanged over more than one categorical border.

Computing the AUC value for all other sensible groups of categories shows that categorical misclassifications are seldom observed. The average AUC-value is 98.7 with a standard deviation of 1.4. Table 4 summarizes the categorical errors over all possible sets of experiments.

Correlation with the outcome of experiments

Correlation of RIN values with downstream experiments is of critical importance. On the one hand, a good correlation will demonstrate the validity of this approach. On the other hand, it allows determination of threshold values in order to get meaningful downstream data. For two-color microarray experiments, this could mean for example that the two input samples should not differ by more than a given number of RIN classes, while the lowest acceptable RIN can be determined as well.

In the present study, RIN values as well as ribosomal ratios were correlated with real-time PCR data. A detailed description of the sample types and extraction methods as well of the entire experimental setup has been published previously [5]. In short, a gene score was calculated based on the average apparent expression level of 4 different housekeeping genes (GAPDH, KYNF, NEFL, β 2M) as measured by real-time PCR. Please note that in this experiment, differences in the apparent gene expression levels are induced by progressing degradation of the RNA string material. Figure 6 shows the plot of the average apparent gene expression on a logarithmic scale against the RIN. Immediately 2 cluster of data appear corresponding to high gene expression (intact RNA) with a high RIN value and low gene expression (degraded RNA) with a low RIN value. On the other hand, the ribosomal ratio exhibits only a weak correlation with the experimentally observed gene expression level (RNA integrity). The RIN allows for a straightforward separation in positives and negatives, whereas the ribosomal ratio would reject many more experiments than necessary. The historical value of 2.0 would reject about 40 experiments of good quality and a value of 1.75 results in about 15 false negatives.

Correlation between RNA integrity and rt-PCR experiments. The figure shows the correlation between RNA integrity values and the outcome of an real-time-PCR experiment, i.e. the average expression values of 4 housekeeping genes (GAPDH, KYNF, NEFL, β 2M). The vertical line is a meaningful threshold value for RIN classification, while the horizontal separates acceptable from unacceptable real-time PCR results, a) The RIN shows a strong correlation (0.52) to the expression value of the house keeping genes. A straightforward separation into positives and negatives is possible. b) The ribosomal ratio shows a poor correlation (0.24) to the expression value of the house keeping genes.


Degradation of solid-state RNA in the presence or absence of a moist atmosphere

We first evaluated the stability of two RNA species at room temperature. First, encapsulated β-galactosidase mRNA samples were left for up to 6 months at room temperature and run on an electrophoresis gel (Figure 1a). On the basis of the measurements of the fluorescence intensity of the bands, one can conclude that no significant degradation occurred over this time period (Figure 1a). Thereafter, we compared the degradation kinetics of rRNA when fully protected from air and when exposed to air. As seen in Figure 1b, RNA exposed to air exhibited a clear degradation: after 92 weeks at room temperature, no intact 28S rRNA molecule could be seen and the RIN value dropped from 7.3 to 2.0. In contrast, when protected from air, the RIN number slightly dropped from 7.2 to 6.8 after 23 months at room temperature.

RNA degradation at room temperature in the presence or absence of a moist atmosphere. β-Galactosidase mRNA samples (a) and total RNA (bd) were dried and encapsulated under anoxic and anhydrous atmosphere. After incubation at room temperature, at −20 °C or at 90 °C, aliquots corresponding to 500 ng of the initial RNA amounts were denatured for 3 min at 75 °C and run on agarose gel. (a) Degradation of HPLC-purified β-galactosidase mRNA. The proportion of undegraded mRNA in minicapsules stored at −20 °C is given under each lane (the experiments were performed in triplicate, but a single representative gel is shown). ‘−80 °C’: RNA stored at −80 °C for 24 weeks. ‘−20 °C’: RNA in minicapsules stored at −20 °C for 24 weeks. (b) Degradation of total RNA at room temperature. Half of the capsules were opened and exposed to air. The ratios of the fluorescence intensities of 28S rRNA and 18S rRNA as well as the RIN values are given for each sample under the respective lanes. (c) Total RNA degradation at 90 °C in the absence of air (minicapsule conditions). Total RNA samples were dried and stored under anoxic and anhydrous atmosphere in minicapsules. They were heated at 90 °C under 50% relative humidity (RH). Two kinetics were run for 8 h and 240 h. Five hundred nanograms of total RNA samples were analyzed on electrophoresis gels. The proportion of intact 28S molecule, measured by the Bio1D software, is indicated for each time point (the experiments were performed in triplicate a single representative gel is shown). (d) Same as in (c), but with RNA exposed to air under 50% RH (the experiments were performed in triplicate a single representative gel is shown).

In order to quantify the degradation rate difference of RNA preserved in the presence or absence of a moist atmosphere, RNA aging was accelerated by heating. Degradation of total RNA was performed at 90 °C either in closed capsules or in opened capsules, and placed in a 50% relative humidity atmosphere (relative humidity control was necessary to maintain RNA in the same hydration state as it is at room temperature).

Figures 1c and d show that, after 8 h of exposure at 50% relative humidity, the amount of 28S rRNA band was no longer measurable, whereas, when protected from air, about 90% of the 28S band was still present. The degradation rate constants at 90 °C, k90 °C, as determined by exponential curve fitting were 7.7 × 10 −10 /nt/s and 1.9 × 10 −8 /nt/s for RNA protected or unprotected from moist air, respectively. This corresponds to a 25-fold increase in the degradation rate constant at 90 °C.

Determination of the room temperature chain-breaking rate of air-protected solid-state RNA

As the RNA degradation rate at room temperature in the minicapsules was slow, it was not possible to directly measure it over the time of the study. However, it can be estimated by using the Arrhenius model on data drawn from kinetics run at different temperatures. In order to obtain a wide variety of samples, we used RNA samples that differed in their origin (bacteria and human cell lines) and the nature of the resuspension buffer used at the end of extraction (10 m M Tris-HCl pH 7, 10 m M Tris-HCl 1 mM EDTA pH 8 and RNAse-free water pH 7.2). Seventy-two kinetics were run at temperatures ranging from 50 to 130 °C (not shown). The proportion of two intact RNA species (human 28S rRNA and E. coli 23S rRNA in total RNA samples) was monitored over time. Supplementary Figure 1, Supplementary Material, is a representative example of these kinetics.

R software was used for determining the strand-breaking rate constant kT from each of the kinetics. The kT values and the confidence intervals for all RNA species (28S and 23S) were averaged. The log10 of these values gave a straight line when plotted as a function of 1/T (K −1 ) (Figure 3). The temperature dependency of the RNA strand-breaking rates followed the Arrhenius model. No discrepancies were found between the 28S and the 23S strand-breaking rate constants. Thus, the degradation rate constant at room temperature in the absence of both atmospheric water and oxygen could be determined by linear extrapolation and was estimated to 3.2 10 −13 /nt/s at 25 °C with a 95% confidence interval of (2.3–4.2) 10 −13 /nt/s. This experiment also confirmed the decrease in stability induced by exposure to moisture (Figure 3, drop data point).

Size dependency of RNA degradation rate in the minicapsules

The degradation rate constants at 90 °C were calculated for five mRNA species by linear regression of the plots of Cq values as a function of time (data not shown). These rate constants appeared to be proportional to the distance between the 5′ of the terminal poly(A) and the 5′ end of the amplicon (Figure 2) (y= 6.10 −10 * (size in nt), r 2 = 0.84). This suggests that, in the minicapsules, the RNA degradation rate depends mainly on its size and can be considered as being independent of its sequence and of the presence of local structures.

Degradation rates as a function of RNA length on several mRNA species. Total RNA samples in minicapsules were heated at 90 °C to accelerate aging. Reverse transcription reactions were set up using (dT)18 oligonucleotides. The qPCR reactions were run using a SybrGreen fluorescent dye and transcript-specific oligonucleotides of TBP, β2M, GAPDH, TUBA1B and PSMB6 as described in Materials and Methods (cf Table 1). For each time point, the Cq value was measured and plotted as a function of time t. The relationship Cq=f(t) was linear for each mRNA species (not shown) and allowed the calculation of the degradation rate constants at 90 °C, k. k was plotted as a function of the distance between the 3′ end of the mRNA species (+18 nucleotides) and the 5′ end of the forward primer. The straight line is a linear fit through the data points.

At last, the normalized calculated k values for each mRNA species were reported on the Arrhenius graph (Figure 3, small circles). These values are in agreement with the other data summarized on the Arrhenius graph.

Temperature dependence of the degradation rate constant of RNA stored in the minicapsules plotted according to the Arrhenius model. Degradation rates (k) of 23S rRNA resuspended after purification in water (open squares), Tris-HCl-EDTA (open diamonds) or Tris-HCl only (open circles) and those of 28S rRNA resuspended in water (closed diamond) were determined at temperatures ranging from 50 to 130 °C with R software (28S=5070 nt 23S=2904 nt, mRNA=3313 nt). Degradation rates of five mRNA species were calculated through RT-qPCR reactions (Figure 2). Log10(k) value was plotted as a function of 1/T. Confidence interval (95%) was calculated with R software as described in Methods. As a comparison, the log10(k) corresponding to the 23S rRNA incubated in opened minicapsules under 50% RH at 90 °C (drop) was also plotted.

QPCR validation experiments on minicapsule-stored RNA

As RT-qPCR measurements are widely used, in particular for gene expression analysis, it was necessary to run this type of experiment on aged RNA samples. Encapsulated human total RNA was heated at 90 °C for increasing periods of times. The rehydrated RNA samples were submitted to a randomly primed reverse transcription, and qPCR reactions were run on eight RNA species that have very different levels of expression in HeLa cells. As controls, RT-qPCR reactions for TUBA1B and 18S amplicons were run on RNA solutions stored at −80 °C. Figure 4 shows either no significant changes or a slight increase in the Cq values for EEF1a1 and β2M qPCR reactions.

Evolution of Cq as a function of time at 90 °C. RT-qPCR reactions and Cq calculations were performed as described in Materials and Methods on amplicons from the following transcripts: TUBA1B (open diamonds), PSMB6 (open squares), PPIE (closed circles), TBP (open circles), EEF1a1 (open triangles), β2M (closed squares), GAPDH (closed diamonds) and 18S rRNA (closed triangles). The slopes of the linear regressions are indicated for each amplicon.

The PCR efficiencies of the qPCR reactions remained stable over the kinetics for each amplicon (SDs of the mean efficiency for each amplicon were between 0.01 and 0.17 and did not correlate with the degradation evolution over time). These data agree with previous studies showing no change in PCR efficiency with RNA degradation. 25

No difference in Cq values was observed between the control solution stored at −80 °C and the encapsulated RNA (TUBA1B: Cq= 20.55 and Cq=20.54, and 18S rRNA: Cq= 7.7+/−0.15, n= 4). These data as well as others (independent evaluations, to be published) suggest that there is no bias in RT-qPCR outcome between RNA stored in the minicapsules and RNA stored in solution at −80 °C.

In conclusion, RNA remained compatible with analyses by RT-qPCR when preserved in minicapsules (see Discussion).

MRIN for direct assessment of genome-wide and gene-specific mRNA integrity from large-scale RNA-sequencing data

The volume of RNA-Seq data sets in public repositories has been expanding exponentially, providing unprecedented opportunities to study gene expression regulation. Because degraded RNA samples, such as those collected from post-mortem tissues, can result in distinct expression profiles with potential biases, a particularly important step in mining these data is quality control. Here we develop a method named mRIN to directly assess mRNA integrity from RNA-Seq data at the sample and individual gene level. We systematically analyse large-scale RNA-Seq data sets of the human brain transcriptome generated by different consortia. Our analysis demonstrates that 3' bias resulting from partial RNA fragmentation in post-mortem tissues has a marked impact on global expression profiles, and that mRIN effectively identifies samples with different levels of mRNA degradation. Unexpectedly, this process has a reproducible and gene-specific component, and transcripts with different stabilities are associated with distinct functions and structural features reminiscent of mRNA decay in living cells.


Figure 1. Impact of presumptive mRNA degradation…

Figure 1. Impact of presumptive mRNA degradation on global gene expression profiling.

Figure 2. mRIN effectively measures mRNA integrity…

Figure 2. mRIN effectively measures mRNA integrity from RNA-Seq data.

( a ) Schematic illustration…

Figure 3. Comparison of mRIN with RIN…

Figure 3. Comparison of mRIN with RIN and other QC metrics.

Figure 4. Comparison of mRIN and RIN…

Figure 4. Comparison of mRIN and RIN in predicting altered gene expression profiles.

Figure 5. Gene-specific degradation and their functional…

Figure 5. Gene-specific degradation and their functional and transcript structural features.


A substantial fraction of exRNAs are not encapsulated inside EVs, yet the extracellular nonvesicular RNAome has not been studied in a comprehensive manner yet. By inhibiting extracellular RNases, our results highlight that the nonvesicular exRNA fraction is highly dynamic. This experimental approach enabled us to obtain exRNA profiles with an unprecedented level of detail and with temporal resolution. Furthermore, we succeeded in stabilizing extracellular full-length tRNAs and ribosomes, which have not been identified before outside EVs due to their susceptibility to extracellular ribonucleases. In contrast, some of their fragments were found to be highly stable and they collectively define the nonvesicular RNAome under standard conditions, especially in the presence of serum. These results have profound implications on the way we understand the mechanisms responsible for RNA release.

The presence of ribosomal aggregates in the extracellular EV-depleted fraction is further supported by the co-isolation of rRNAs, ribosomal proteins and polyA+ mRNAs from the same chromatographic fractions. Extracellular ribosomes were described in the 70’s in the blowfly Calliphora vicina ( 39) but subsequently linked to an experimental artefact ( 40) and have received little attention since then. However, we have demonstrated that extracellular ribosomes exist at least transiently in the media of cultured mammalian cells and possibly also in body fluids. In support of the latter, a modified small-RNA sequencing method has been recently reported that permits the identification of mRNA fragments in blood plasma or serum ( 41). Strikingly, the authors found that the distribution and length of reads mapping to mRNAs was reminiscent of ribosome profiling, suggesting that the sequenced fragments could be the footprints of ribosomes circulating in biofluids.

There is an increasing body of evidence showing that EVs actually contain more full-length ncRNAs than microRNAs or ncRNA fragments. For instance, Bioanalyzer's peaks corresponding to intact 18S and 28S rRNAs have been identified in purified EVs ( 32, 42–45), while full-length YRNAs and other ncRNAs have been identified by sequencing, RT-qPCR and/or Northern blot ( 17, 46). The use of thermostable group II intron reverse transcriptases (TGIRT-seq) has allowed the identification of full-length tRNAs in EVs, which greatly outnumber tRNA-derived fragments ( 47–49). Our results are consistent with these reports, and clearly show the presence of tRNAs and 7SL RNAs in EVs purified by buoyant density flotation in linear iodixanol gradients. At the level of sensitivity achievable by DIG-based Northern blotting, tRNA-derived fragments were not detectable in EVs (Figure 4H). Therefore, TGIRT-seq does not seem to be biasing results toward full-length tRNAs ( 48), and gives a picture of the RNA content of EVs which is in good agreement with our Northern blot results. This can change, however, when EVs are purified from stressed cells. Considering stressed cells upregulate tRNA-derived fragments ( 23, 24) and that changes in intracellular RNA profiles are mirrored in EVs ( 50), it is reasonable to speculate that EVs coming from stressed cells could contain higher payloads of stress-induced tRNA halves ( 51).

Nonvesicular exRNAs have received very little attention ( 15) until recently ( 32). In the past, we have compared the small RNA content between EVs and 100,000 x g supernatants of cell-conditioned medium and found that the EV-depleted fraction was highly enriched in 5′ tRNA halves of precisely 30 or 31 nucleotides and almost exclusively derived from glycine or glutamic acid tRNAs ( 16). Similar results were obtained by other groups working on primary cultures of glioblastoma ( 17). Furthermore, glycine tRNA halves are predominantly found outside EVs in serum ( 20) and are ubiquitous in many biofluids including serum, urine, saliva and bile ( 18). We are now showing that these fragments are not directly released from cells in vitro. Instead, they are generated in the extracellular space (Figure 6). Enrichment of these fragments, especially when found in the EV-depleted fraction, is probably a consequence of their differential extracellular stability rather than their preferential or selective secretion. This is further supported by the recent observation that circular RNAs, which are known to be highly stable, are increased in nearly all human biofluids when compared to matched tissues ( 52).

Proposed model. (A) Cells in culture release tRNAs, ribosomal subunits or ribosomes to the extracellular space, even outside EVs. When the CCM is analyzed by SEC, these RNAs define the P0 and P1 peaks, respectively (i). However, their detection is only possible after addition of RI to the medium. Active secretion (e.g., autophagy-dependent) might contribute to nonvesicular exRNA profiles, but damaged or dead cells with compromised plasma membrane integrity may be quantitatively more important. Extracellular RNases degrade nonvesicular RNAs and generate stable fragmentation products like glycine tRNA halves (ii), which can assemble into dimers and elute in the chromatographic P1 peak even in the absence of RI. The P2 peak is probably composed of rRNA-derived fragments (rRFs) forming tightly bound dsRNAs which are not amenable to standard small RNA sequencing techniques. While full-length tRNAs and YRNAs are not detected in the non-EV fraction in the absence of RI, those which are present inside EVs are protected from degradation. Thus, EVs are probably the only source of full-length ncRNAs in RNase-rich extracellular samples. Overall, this diagram represent the remarkable differences between what is sequenced in the extracellular space in the absence of RI and what cells actually release, as revealed by RI-SEC-seq. (B) A diagram explaining possible biogenetic routes for extracellular, nonvesicular tRNA Gly GCC 5’ halves.

Proposed model. (A) Cells in culture release tRNAs, ribosomal subunits or ribosomes to the extracellular space, even outside EVs. When the CCM is analyzed by SEC, these RNAs define the P0 and P1 peaks, respectively (i). However, their detection is only possible after addition of RI to the medium. Active secretion (e.g., autophagy-dependent) might contribute to nonvesicular exRNA profiles, but damaged or dead cells with compromised plasma membrane integrity may be quantitatively more important. Extracellular RNases degrade nonvesicular RNAs and generate stable fragmentation products like glycine tRNA halves (ii), which can assemble into dimers and elute in the chromatographic P1 peak even in the absence of RI. The P2 peak is probably composed of rRNA-derived fragments (rRFs) forming tightly bound dsRNAs which are not amenable to standard small RNA sequencing techniques. While full-length tRNAs and YRNAs are not detected in the non-EV fraction in the absence of RI, those which are present inside EVs are protected from degradation. Thus, EVs are probably the only source of full-length ncRNAs in RNase-rich extracellular samples. Overall, this diagram represent the remarkable differences between what is sequenced in the extracellular space in the absence of RI and what cells actually release, as revealed by RI-SEC-seq. (B) A diagram explaining possible biogenetic routes for extracellular, nonvesicular tRNA Gly GCC 5’ halves.

Live cells can release a representative fraction of their cytoplasm by mechanisms such as cytoplasmic extrusion ( 53) or amphisome fusion with the plasma membrane ( 32), and these mechanisms could be involved in the release of nonvesicular ribosomes and tRNAs to the extracellular space. However, a few events of cellular damage or death might be quantitatively more important in defining exRNA profiles as has been discussed above. In support of this, it has been shown that extracellular rRNA levels correlate with extracellular lactate dehydrogenase (LDH) activity, which is widely used as a marker of cell death ( 54). Even though exRNA analysis derived from dead cells can be considered as an artifact of cell culture, there are situations where nonapoptotic, immunogenic cell death (ICD) occurs at abnormal frequencies in an organism. These situations include aging ( 55), trauma ( 56), ischemia-reperfusion injury ( 57), infectious diseases and cancer. In the latter, ICD can occur because of the hypoxic inner mass characteristic of solid tumors or following treatment with cytotoxic agents ( 58). In all cases, dying cells release intracellular components which are sensed by innate immune cells and interpreted as damage-associated molecular patterns (DAMPs). Furthermore, the therapeutic activity of several anticancer drugs eliciting ICD involves an autocrine and paracrine circuit which depends at least in part on the release of self RNAs by stressed or dying cancer cells ( 59). Because rRNAs and tRNAs are highly abundant intracellularly and they are exposed in the extracellular space in cases of damage, and considering RNAs are actively sensed by the innate immune system ( 60, 61), we hypothesized that exRNA-containing nonvesicular complexes could be endowed with immunomodulatory abilities. The high turnover of these complexes as a consequence of extracellular RNases could prevent activation under physiological conditions.

As for the ribonuclease responsible for extracellular, nonvesicular tRNA cleavage, it is clearly a serum-derived ribonuclease in FBS-containing samples (probably RNase A itself). When serum is not present or highly diluted, such as after thoroughly washing cells with serum-free media or buffers, it is possible that endogenous secreted RNases are responsible for shaping nonvesicular exRNA profiles. Stressed cancer cells secrete enzymes to perform some metabolic reactions in the extracellular space and then uptake the enzymatic products to fuel cellular energetics ( 62). By analogy, we are tempted to speculate that secreted RNases such as ANG could play a role in extracellular RNA metabolism, preventing the toxicity associated with their intracellular activity in nonstressed cells ( 63). Although the function of ANG in tRNA cleavage seems to be partially redundant ( 36, 64), its implications in extracellular RNA cleavage under physiological conditions remains to be elucidated. Redundancy might be lower in serum-free environments as the nervous system, where several mutations in ANG have been functionally linked to neurodegenerative diseases ( 65). We have provided preliminary evidence suggesting an involvement of extracellular, nonvesicular RNAs or RNPs in immune surveillance. Thus, a link between mutations in ANG and deregulated extracellular RNA fragmentation patterns is feasible in diseases such as ALS whose etiology or evolution is deeply connected to inflammation ( 66).

Bacterial rRNA and tRNAs induce Toll-like receptor (TLR)-dependent DC maturation and IL-1β secretion, and are therefore considered pathogen-associated molecular patterns. However, to elicit such a response, addition of the purified RNAs with cationic lipids seems to be essential ( 67). In contrast, we have obtained high extracellular levels of IL-1β when incubating BMDC with approximately one microgram of RNA obtained from the P0 peak of MCF-7 cells (composed mainly of ribosome particles) in the absence of any transfection reagent. Strikingly, this effect was lost when incubating DCs with RNase A-pretreated P0. It remains to be elucidated whether RNA itself or any potentially associated RNA-binding proteins are responsible for these effects. We address some limitations in our experimental design, including the incubation of human-derived RNAs with murine dendritic cells. Thus, these results should be interpreted with caution. Follow-up studies will confirm whether or not extracellular nonvesicular RNAs are active players in immune surveillance.

Although extracellular ribosomes are not predicted to resist extracellular RNases they might still achieve functionally relevant concentrations in vivo in extracellular microenvironments. The identification of a RNase-resistant peak (P2) derived from partial fragmentation of P0 (and possibly P1 as well) suggests that, similarly to what we have shown for 30–31 nt tRNA halves, rRNA-derived fragments may accumulate in the extracellular space and their extracellular concentration may increase in situations of abnormal cell death. A new method has been recently described enabling RNA sequencing from a few microliters of human serum ( 68). With this method, almost perfect separation between normal and breast cancer patients was possible based on rRNA or mitochondrial tRNA sequences.

In conclusion, ribonuclease inhibition dramatically shapes extracellular RNA profiles and uncovers a population of extracellular ribosomes, tRNAs and other coding and noncoding RNAs. These RNAs, which are not protected by encapsulation inside EVs, are rapidly degraded by extracellular RNases. However, some of their fragments resist degradation and can accumulate in cell culture media and in biofluids. This dynamic view of exRNAs impacts our understanding of RNA secretion mechanisms and may offer a window to new molecules with biomarker potential. These include intrinsically stable ncRNA fragments and extracellular RNPs stabilized by addition of RI immediately upon collection of samples.

Alcohol and the Nervous System

Greg T. Sutherland , . Jillian J. Kril , in Handbook of Clinical Neurology , 2014

Gene expression

The other major spin-off from the knowledge and technology developed around the human genome project is genomewide expression or transcriptomic studies. The ability to investigate the total gene expression of a cell or tissue and then compare individuals with and without a disease gives us an unbiased platform both to test existing hypotheses and discover hitherto unknown pathobiology. Again, the original probe hybridization-based array technology is currently being replaced by next-generation sequencing, called RNA-Seq, that promises a more accurate and comprehensive analysis of neurodegenerative diseases ( Sutherland et al., 2011 ). However, the success of both techniques is highly dependent on the quality of the starting template, ribonucleic acid (RNA).

RNA quality or integrity is commonly measured using an algorithm that combines multiple features from an RNA electrophoresis gel scan and represents them as a total RNA integrity number (RIN) ( Schroeder et al., 2006 ). It was originally thought that RNA degradation was largely due to postmortem factors, but it is now known, for postmortem brain tissue in particular, that it is the premortem or agonal period that is critical for the quality of the RNA that is extracted ( Monoranu et al., 2009 Durrenberger et al., 2010 ). RIN is highly correlated with brain pH ( Monoranu et al., 2009 ) and brain pH and RIN decrease with the length of the agonal period and the number of adverse, often hypoxic, events within that period, such as hospitalization, coma, artificial respiration, or respiratory illness ( Durrenberger et al., 2010 ). In addition, the agonal period is associated with novel gene expression patterns that are not directly related to the disease ( Li et al., 2004 ). A major issue for case-control transcriptomic studies is that patients with neurologic disorders generally have a more protracted and averse agonal period and subsequently a lower pH and RIN compared with controls who have often died suddenly ( Yates et al., 1990 Monoranu et al., 2009 ). The removal of low-RIN samples is a standard practice in most studies using postmortem brain tissue but it remains prudent to evaluate the potential effects of RIN and other confounders in transcriptomic data before concluding that there are disease-specific differences.

Notwithstanding these technical issues, genomewide expression studies of brain tissue from alcoholics without comorbidities such as liver disease or WKS have revealed very subtle and largely divergent data ( Flatscher-Bader et al., 2006 ). An exception has been the dysregulation of the myelin-associated genes, although the individual genes themselves and their direction of change have varied between studies ( Lewohl et al., 2000 Iwamoto et al., 2004 Li et al., 2004 ). Other major themes have been DNA repair ( Mayfield et al., 2002 Iwamoto et al., 2004 ), the ubiquitin-proteasome system, and immune function ( Sokolov et al., 2003 Liu et al., 2004 ).

The proinflammatory or neuroimmune-activating effects of alcohol may initially involve cytokine production by resident microglia and astrocytes via the innate immune system, Toll-like receptors (TLR) ( Blanco et al., 2005 Blanco and Guerri, 2007 ). This in turn induces the proinflammatory transcription factor NF-κB (nuclear factor of kappa light polypeptide gene enhancer in B cells) transcription, with increased levels of the cytokines interleukin 1beta, tumor necrosis factor-alpha, and interleukin-6 ( Alfonso-Loeches et al., 2010 ). Research from experimental animals has shown that the inhibition of TLR4-mediated NF-κB signaling attenuates both the toxic ( Alfonso-Loeches et al., 2010 ) and acute behavioral ( Wu et al., 2012 ) actions of alcohol. Gene expression studies suggest that the activation of NF-κB transcription with acute alcohol exposure eventually gives way to downregulation, with subsequent cycles of intoxication and withdrawal ( Okvist et al., 2007 ). In keeping with the known regional loss of neurons in ARBD, the genes with NF-κβ elements were generally downregulated in the prefrontal cortex of chronic alcoholics but not the motor cortex ( Okvist et al., 2007 ).

The common comorbidity HE is another important potential confounder for transcriptomic analysis of ARBD. Like hypoxic episodes in the agonal period HE is associated with a metabolic acidosis, and alcoholics with HE have a significantly lower brain pH and RIN than those without HE ( Sheedy et al., 2012 ). Our own work comparing the white-matter transcriptome of alcoholics with HE and non-HE alcoholics with controls also found that transcriptomic changes were largely confined to the HE group ( Sutherland et al., 2014 ). Furthermore, like hypoxia, HE was associated with a downregulation of energy metabolism pathways, and the additive effect of HE and agonal hypoxic events on lowering brain pH, RIN, and the magnitude of gene downregulation could be seen. HE was also characterized by an upregulation of proinflammatory pathways consistent with the increasing realization that both systemic inflammation and elevated brain ammonia levels probably act in combination to cause this condition ( Shawcross et al., 2005, 2011 Rama Rao et al., 2010 ). The term systemic is important here as the resident microglia, although activated, do not appear to be a prominent source of inflammatory cytokines in HE ( Zemtsova et al., 2011 ).

Two relatively recent paradigm shifts in our knowledge of gene expression regulation have been recognition of the roles played by epigenetics and non-coding RNA species, along with interactions between the two. Ponomarev et al. (2012) showed that the alcoholic prefrontal cortex was characterized by global DNA hypomethylation and postulated that this was due to alcohol-induced deficiencies in folate and its derivative, S-adenosyl methionine, the major substrate for DNA methylation. Interestingly, the GC content of genes appeared to be the major factor that determined whether alcohol affected their expression rather than their membership of a specific cellular pathway.

The interest in non-coding RNA species has evolved with the realization that transcription is pervasive and not limited to just the protein-coding region of the genome ( Clark et al., 2011 ). Non-coding RNAs come in many forms and sizes, including the relatively well-known microRNAs, small oligonucleotides, 22 bases in length, that are able to control protein expression by restricting access to translational machinery or augmenting transcript degradation ( Fabian et al., 2010 ). Lewohl et al. (2011) found that 35 microRNAs were upregulated in chronic alcoholic brains and their putative targets coincided with the downregulated mRNAs from their previous microarray data ( Liu et al., 2006 ). MicroRNAs can modulate TLR (and NF-κB) signaling and they display regulatory reciprocity with epigenetic factors ( Nunez and Mayfield, 2012 ). A special property of microRNAs is their promiscuity, which means they are able to bind to and downregulate multiple gene targets simultaneously. This is not only ideal for coordinating complex gene expression networks in organisms but it also raises the possibility that microRNAs may be highly effective therapeutic targets for complex disorders such as ARBD ( Nunez and Mayfield, 2012 ).


The RNA Integrity database (RINdb) is a freely accessible collection of records containing electropherograms, RIN and other sample metadata. Scientists can query the database to prescreen methods for a new experiment, compare RIN thresholds and validate own results. All users can contribute records of their own.

RNA Integrity Database (RINdb)- Bioanalyzer RNA Profiles
Marc Valer, Microfluidics Program Manager, Genomics, Agilent Technologies, Santa Clara.

Synopsis: Thousands of users today trust the bioanalyzer 2100 for qualifying total RNA samples, looking for integrity, purity, recovery, and consistency information for sample preparation methods, assisting them to determine the optimal conditions for their experimental design. Marc Valer describes the use of Agilent's new RNA Integrity Database (RINdb) to assist in the development of experimental protocols, particularly sample preparations, by providing a means for researchers to contextualize RNA profiles.

RNA integrity and RNA fragmentation - Biology

Procedures for Quality Control of RNA Samples for Use in Quantitative Reverse Transcription PCR
Tania Nolan 1,2 & Stephen Bustin 2,3
1 Sigma Aldrich,Cambridge UK 2 Eureka Biotechnology, Cambridge UK 3 Institute of Cell and Molecular Science, Barts and the London Queen Mary’s School of Medicine and Dentistry, London, UK
Essentials of Nucleic Acid Analysis: A Robust Approach (2008) editors: Jacquie T. Keer, Lyndsey Birch

Book chapter 9 introduction - The quality of any scientific data is directly proportional to that of theoriginal starting samples, or simply ‘garbage in, garbage out’. In most circumstances it is logical to work with the highest quality material possible. However, for some experiments the highest quality possible is still a serious compromise from perfection. The degree to which the standard of input material influences final quantitative reverse transcription polymerase chain reaction (qRT-PCR) data and, potentially, the resulting scientific conclusion, is outlined in this chapter.
Preservation of fine-needle aspiration specimens for future use in RNA-based molecular testing.
Ladd AC, O'Sullivan-Mejia E, Lea T, Perry J, Dumur CI, Dragoescu E, Garrett CT, Powers CN.
Cancer Cytopathol. 2011 Apr 25119(2): 103-110
Department of Pathology, School of Medicine, Virginia Commonwealth University, Richmond, Virginia

BACKGROUND: The application of ancillary molecular testing is becoming more important for the diagnosis and classification of disease. The use of fine-needle aspiration (FNA) biopsy as the means of sampling tumors in conjunction with molecular testing could be a powerful combination. FNA is minimally invasive, cost effective, and usually demonstrates accuracy comparable to diagnoses based on excisional biopsies. Quality control (QC) and test validation requirements for development of molecular tests impose a need for access to pre-existing clinical samples. Tissue banking of excisional biopsy specimens is frequently performed at large research institutions, but few have developed protocols for preservation of cytologic specimens. This study aimed to evaluate cryopreservation of FNA specimens as a method of maintaining cellular morphology and ribonucleic acid (RNA) integrity in banked tissues.
METHODS: FNA specimens were obtained from fresh tumor resections, processed by using a cryopreservation protocol, and stored for up to 27 weeks. Upon retrieval, samples were made into slides for morphological evaluation, and RNA was extracted and assessed for integrity by using the Agilent Bioanalyzer (Agilent Technologies, Santa Clara, Calif).
RESULTS: Cryopreserved specimens showed good cell morphology and, in many cases, yielded intact RNA. Cases showing moderate or severe RNA degradation could generally be associated with prolonged specimen handling or sampling of necrotic areas.
CONCLUSIONS: FNA specimens can be stored in a manner that maintains cellular morphology and RNA integrity necessary for studies of gene expression. In addition to addressing quality control (QC) and test validation needs, cytology banks will be an invaluable resource for future molecular morphologic and diagnostic research studies.
Maintaining RNA integrity in a homogeneous population of mammary epithelial cells isolated by Laser Capture Microdissection.
Bevilacqua C, Makhzami S, Helbling JC, Defrenaix P, Martin P.
BMC Cell Biol. 2010 Dec 611:95.
INRA, UMR1313 Unité Génétique Animale et Biologie Intégrative, équipe Lait, Génome & Santé F-78350 Jouy-en-Josas, France

CONCLUSIONS: RNAs isolated from MEC in this manner were of very good quality for subsequent linear amplification, thus making it possible to establish a referential gene expression profile of the healthy MEC, a useful platform for tumor biomarker discovery.
Impact of RNA degradation on gene expression profiling.
Opitz L, Salinas-Riester G, Grade M, Jung K, Jo P, Emons G, Ghadimi BM, Beissbarth T, Gaedcke J.
BMC Med Genomics. 2010 Aug 93: 36.

BACKGROUND: Gene expression profiling is a highly sensitive technique which is used for profiling tumor samples for medical prognosis. RNA quality and degradation influence the analysis results of gene expression profiles. The impact of this influence on the profiles and its medical impact is not fully understood. As patient samples are very valuable for clinical studies, it is necessary to establish criteria for the RNA quality to be able to use these samples in later analysis.
METHODS: To investigate the effects of RNA integrity on gene expression profiling, whole genome expression arrays were used. We used tumor biopsies from patients diagnosed with locally advanced rectal cancer. To simulate degradation, the isolated total RNA of all patients was subjected to heat-induced degradation in a time-dependent manner. Expression profiling was then performed and data were analyzed bioinformatically to assess the differences.
RESULTS: The differences introduced by RNA degradation were largely outweighed by the biological differences between the patients. Only a relatively small number of probes (275 out of 41,000) show a significant effect due to degradation. The genes that show the strongest effect due to RNA degradation were, especially, those with short mRNAs and probe positions near the 5' end.
CONCLUSIONS: Degraded RNA from tumor samples (RIN > 5) can still be used to perform gene expression analysis. A much higher biological variance between patients is observed compared to the effect that is imposed by degradation of RNA. Nevertheless there are genes, very short ones and those with the probe binding side close to the 5' end that should be excluded from gene expression analysis when working with degraded RNA. These results are limited to the Agilent 44 k microarray platform and should be carefully interpreted when transferring to other settings.

RNA integrity in post-mortem samples: influencing parameters and implications on RT-qPCR assays.
Koppelkamm A, Vennemann B, Lutz-Bonengel S, Fracasso T, Vennemann M.
Int J Legal Med. 2011 Jul125(4): 573-580
Institute of Legal Medicine, Freiburg University Medical Center, Albertstrasse 9, 79104, Freiburg, Germany.

Analysis of Circulating MicroRNA - Pre-analytical and Analytical Challenges.
McDonald JS, Milosevic D, Reddi HV, Grebe SK, Algeciras-Schimnich A.
Clin Chem. 2011 Jun57(6):833-40. Epub 2011 Apr 12.
Departments of Laboratory Medicine and Pathology

Reliable gene expression measurements from degraded RNA by quantitative real-time PCR depend on short amplicons and a proper normalization.
Antonov J, Goldstein DR, Oberli A, Baltzer A, Pirotta M, Fleischmann A, Altermatt HJ, Jaggi R.
Lab Invest. 2005 Aug85(8): 1040-1050.
Department of Clinical Research, University of Bern, Bern, Switzerland.

The effect of RNA degradation on the diagnostic utility of microRNA has not been systematically evaluated in clinical samples. We asked if the microRNA profile is preserved in degraded RNA samples derived from mouse and human tissue. We selected tissue-specific microRNA candidates from published human microarray data, and validated them using quantitative reverse transcription polymerase chain reaction (QRTPCR) analyses on flash-frozen, normal mouse liver, pancreas, and stomach tissue samples. MiR-122a, miR-1, and miR-200b were identified as tissue-specific, and the 3-microRNA-based QRTPCR could predict the tissue origin for mouse tissue samples that were left at room temperature for 2 h with an accuracy of 91.7%. When we applied this 3-microRNA predictor to clinical specimens with various degree of RNA degradation, the predictor differentiated degraded RNA samples from liver, pancreas, and stomach with an accuracy of 90% (26/29). Expression levels of miR-122a, miR-1, and miR-200b were modestly changed after the extended (2-4 h) storage at room temperature, but the magnitudes of expression changes were small compared to the expression differences between various tissues of origin. This proof-of-principle study demonstrates that RNA degradation due to extended storage at room temperature does not affect the predictive power of tissue-specific microRNA QRTPCR predictor.
A robust RNA integrity-preserving staining protocol for laser capture microdissection of endometrial cancer tissue.
Cummings M, McGinley CV, Wilkinson N, Field SL, Duffy SR, Orsi NM.
Anal Biochem. 2011 Sep 1416(1): 123-125
Gynaeimmunology and Oncology Group, YCR and Liz Dawn Pathology and Translational Sciences Centre, Leeds Institute of Molecular Medicine, St. James's University Hospital, Leeds LS9 7TF, UK.

Laser capture microdissection of frozen tissue sections allows homogeneous cell populations to be isolated for expression profiling. However, this requires striking a balance between retaining adequate morphology for accurate microdissection and maintaining RNA integrity. Various staining protocols were applied to frozen endometrial carcinoma tissue sections. Although alcohol-based methods were superior to aqueous stains for maintaining RNA integrity, they suffered from irreproducible staining intensity. We developed a modified alcohol-based, buffered cresyl violet staining protocol that provides reproducible staining with minimal RNA degradation suitable for tissues with moderate to high levels of intrinsic RNase activity.
Effects of RBC removal and TRIzol of peripheral blood samples on RNA stability.
Kang JE, Hwang SH, Lee JH, Park DY, Kim HH.
Clin Chim Acta. 2011 Jun 17
Department of Laboratory Medicine, Pusan National University Hospital, School of Medicine, Pusan National University, Busan, South Korea.

BACKGROUND: Purification of mRNA from stored specimens is very important because results from RT-PCR and microarray analyses are largely affected by the quality of mRNA. Moreover, many preanalytical factors during collection, processing, and storage may affect mRNA quality and the expression of peripheral blood mononuclear cells (PBMC). In this study, we evaluate the effects of RBC removal techniques and TRIzol on RNA quality in blood samples.
METHODS: We obtained EDTA-blood samples from 50 adult volunteers, and made 10 pools of buffy coats for comparison between protocols and also evaluated RNA quality of clinical samples in biobank. Use of TRIzol and RBC removal (RBC lysis or cell separation) were evaluated their effect on the quality of mRNA from the stored blood samples.
RESULTS: RNA integrity with TRIzol was significantly better than that without TRIzol (RIN 4.5 vs. 9.2, respectively P=0.002). The change in RIN of the PBMC separation method was equivalent to that of the RBC lysis method. After 12months, IL6 mRNA expression from stored clinical samples in cell separation/TRIzol was stable.
CONCLUSIONS: The blood samples frozen in TRIzol after RBC removal preserved RNA quality well. PBMC/TRIzol preservation for storage of blood samples could be a simple protocol for rapid, low-cost biobanking.

Effects of delay in the snap freezing of colorectal cancer tissues on the quality of DNA and RNA.
Hong SH, Baek HA, Jang KY, Chung MJ, Moon WS, Kang MJ, Lee DG, Park HS.
J Korean Soc Coloproctol. 2010 Oct26(5): 316-323
Chonbuk National University Hospital National Biobank of Korea, Jeonju, Korea.

PURPOSE: The success of basic molecular research using biospecimens strongly depends on the quality of the specimen. In this study, we evaluated the effects of delayed freezing time on the stability of DNA and RNA in fresh frozen tissue from patients with colorectal cancer.
METHODS: Tissues were frozen at 10, 30, 60, and 90 minutes after extirpation of colorectal cancer in 20 cases. Absorbance ratio of 260 to 280 nm (A(260)/A(280)) and agarose gel electrophoresis were evaluated. In addition, the RNA integrity number (RIN) was assayed for the analysis of the RNA integrity.
RESULTS: Regardless of delayed freezing time, all DNA and RNA samples revealed A(260)/A(280) ratios of more than 1.9, and all DNA samples showed a discrete, high-molecular-weight band on agarose gel electrophoresis. The RINs were 7.53 ± 2.04, 6.70 ± 1.88, 6.47 ± 2.58, and 4.22 ± 2.34 at 10, 30, 60, and 90 minutes, respectively. Though the concentration of RNA was not affected by delayed freezing, the RNA integrity was decreased with increasing delayed freezing time.
CONCLUSION: According to the RIN results, we recommend that the collection of colorectal cancer tissue should be done within 10 minutes for studies requiring RNA of high quality and within 30 minutes for usual RNA studies.

I mproved protocol for high-quality co-extraction of DNA and RNA from rumen digesta.
Popova M, Martin C, Morgavi DP.
Folia Microbiol (Praha). 2010 Jul55(4): 368-372
INRA Clermont-Ferrand-Theix, 63122, Saint Gènes Champanelle, France.

We report an improved method for total nucleic acids extraction from rumen content samples. The method employs bead beating, and phenol-chloroform extraction followed by saline-alcohol precipitation. Total nucleic acids and RNA yield and purity were assessed by spectrophotometric measurements RNA integrity was estimated using Agilent RNA 6000 Nano Kit on an Agilent 2100 Bioanalyzer. The method provided total nucleic acids and RNA extracts of good quantity and quality. The extraction is not time consuming and it is valuable for ecological studies of rumen microbial community structure and gene expression.

mRNA profiling in forensic genetics I - Possibilities and limitations.
Vennemann M & Koppelkamm A.
Institute of Legal Medicine, University of Freiburg, Albertstr. 9, 79104 Freiburg, Germany
Forensic Sci Int. 2010 Dec 15203(1-3): 71-75

Postmortem mRNA profiling II - Practical considerations.
Vennemann M & Koppelkamm A.
Institute of Legal Medicine, University of Freiburg, Albertstr. 9, 79104 Freiburg, Germany
Forensic Sci Int. 2010 Dec 15203(1-3): 76-82

Maintaining RNA integrity in a homogeneous population of mammary epithelial cells isolated by Laser Capture Microdissection
Claudia Bevilacqua email, Samira Makhzami email, Jean-Christophe Helbling email, Pierre Defrenaix email and Patrice Martin email
BMC Cell Biology 2010, Published: December 2010

Robust microRNA stability in degraded RNA preparations from human tissue and cell samples.
Jung M, Schaefer A, Steiner I, Kempkensteffen C, Stephan C, Erbersdobler A, Jung K.
Department of Urology, University Hospital Charité, Berlin, Germany.
Clin Chem. 2010 Jun56(6): 998-1006.

Assessment of mRNA and microRNA stabilization in peripheral Human Blood for Multicenter studies and Biobanks
Daniel Gilbert Weber1, Swaantje Casjens1, Peter Rozynek1, Martin Lehnert1, Sandra Zilch-Schöneweis1, Oleksandr Bryk1, Dirk Taeger1, Maria Gomolka2, Michaela Kreuzer2, Heinz Otten3, Beate Pesch1, Georg Johnen1 and Thomas Brüning11
Institute for Prevention and Occupational Medicine of the German Social Accident Insurance—Institute of the Ruhr-Universität Bochum (IPA), Bochum, Germany. 2Department of Radiation Protection and Health, Federal Office for Radiation Protection, Oberschleissheim, Germany. 3German Social Accident Insurance (DGUV), Sankt Augustin, Germany.
Biomarker Insights 2010(5): 95�

Pre-PCR Processing - Strategies to Generate PCR-Compatible Samples
Peter Rådström,* Rickard Knutsson, Petra Wolffs,Maria Lövenklev, and Charlotta Löfström
MOLECULAR BIOTECHNOLOGY Volume 26, 2004: 133-146

I. Riedmaier, M. Bergmaier and M.W. Pfaffl
Technische Universitat Munchen, Physiology Weihenstephan, Freising, Germany
Biotechnol. & Biotechnol. Equipment 2010, 24(4): 2154-2159

The integrity of RNA is a very critical aspect regarding downstream RNA based quantitative analysis like RT-qPCR. Low-quality RNA can compromise the results of such experiments. Today automated lab-on-chip capillary electrophoresis allows rapid RNA quality and quantity determination, e.g. 2100 Bioanalyzer (Agilent Technologies) and the Experion (Bio-Rad). Both platforms determine RNA quality using a numerical system which represents the integrity of RNA. The Bioanalyzer offers the RIN algorithm (RNA Integrity Number) on the Bioanalyzer 2100 and Bio-Rad developed a new Experion software version that offers an algorithm for calculating the RNA Quality Index (RQI).The aim of this study was to compare both systems regarding sensitivity, reproducibility, linearity and the influence of individual tissue extractions and different chip runs on RNA quality and quantity determination.Overall it was confirmed that both algorithms are very comparable and beneficial for the determination of RNA quality for downstream applications. The Experion showed slightly better results regarding reproducibility and absolute sensitivity, whereas the 2100 Bioanalyzer showed a higher linearity.
Quantitative assessment of the sensitivity of various commercial reverse transcriptases based on armored HIV RNA.
Okello JB, Rodriguez L, Poinar D, Bos K, Okwi AL, Bimenya GS, Sewankambo NK, Henry KR, Kuch M, Poinar HN.
Department of Anthropology, McMaster Ancient DNA Centre, McMaster University, Hamilton, Ontario, Canada. PLoS One. 2010 Nov 105(11):e13931.

CONCLUSIONS/SIGNIFICANCE: We therefore recommend the use of Accuscript or Superscript III when dealing with low copy number RNA levels, and suggest purification of the RT reactions prior to downstream applications (eg qPCR) to augment detection. Although the results presented in this study were based on a viral RNA surrogate, and applied to nucleic acid lysates derived from archival formalin-fixed paraffin embedded tissue, their relative performance on RNA obtained from other tissue types may vary, and needs future evaluation. Stabilizing RNA at room temperature in RNAstable
Sharron Ohgi1, Laurent Coulon, Rolf Muller, Judy-Muller-Cohn, and Omoshile ClementBiomatrica, Inc., 5627 Oberlin Dr, #120, San Diego, CA 92121, USA.
Biotechniques Vol. 48 (No. 6) 2010: 470

RNAstable is a novel preservation product developed to protect RNA from degradation during storage or shipment at ambient temperatures. The synthetic storage medium is based on the natural principles of anhydrobiosis (meaning “life without water”), a biological mechanism employed by some organisms that enables their survival while dry for more than 100 years. Anhyd-robiotic organisms protect their DNA, RNA, proteins, membranes and cellular systems for survival in a dry state and can be revived by simple rehydration. RNAstable was designed to mimic these unique characteristics to stabilize RNA at ambient tempera-tures for prolonged time periods. Quantitative RT-PCR analysis demonstrates successful amplification of RNA templates that were stored dry in RNAstable for 29 months at room temper-ature and under accelerated aging conditions equivalent to 12 years of room temperature storage (elevated temperatures at 45°C). Samples were sealed inside a moisture-barrier bag including a desiccant pack to ensure ideal storage conditions. Rehydrated samples were used directly in reactions without further purification and exhibited no inhibition or loss of activity. This innovative technology prevents degradation of RNA at room temperature and offers tremendous cost and energy savings as an easy-to-use alternative to conventional freezer storage.

A guide to ions and RNA structure.
Draper DE.
Department of Chemistry, Johns Hopkins University, Baltimore, Maryland 21218, USA
RNA. 2004 Mar10(3):335-43.

DNA integrity publication:

Multiplex picoliter-droplet digital PCR for quantitative assessment of DNA integrity in clinical samples.
Didelot A, Kotsopoulos SK, Lupo A, Pekin D, Li X, Atochin I, Srinivasan P, Zhong Q, Olson J, Link DR, Laurent-Puig P, Blons H, Hutchison JB, Taly V.
Université Paris Sorbonne Cité, INSERM UMR-S775, Paris, France.
Clin Chem. 2013 May59(5): 815-823

BACKGROUND: Assessment of DNA integrity and quantity remains a bottleneck for high-throughput molecular genotyping technologies, including next-generation sequencing. In particular, DNA extracted from paraffin-embedded tissues, a major potential source of tumor DNA, varies widely in quality, leading to unpredictable sequencing data. We describe a picoliter droplet-based digital PCR method that enables simultaneous detection of DNA integrity and the quantity of amplifiable DNA.
METHODS: Using a multiplex assay, we detected 4 different target lengths (78, 159, 197, and 550 bp). Assays were validated with human genomic DNA fragmented to sizes of 170 bp to 3000 bp. The technique was validated with DNA quantities as low as 1 ng. We evaluated 12 DNA samples extracted from paraffin-embedded lung adenocarcinoma tissues.
RESULTS: One sample contained no amplifiable DNA. The fractions of amplifiable DNA for the 11 other samples were between 0.05% and 10.1% for 78-bp fragments and 𕟩% for longer fragments. Four samples were chosen for enrichment and next-generation sequencing. The quality of the sequencing data was in agreement with the results of the DNA-integrity test. Specifically, DNA with low integrity yielded sequencing results with lower levels of coverage and uniformity and had higher levels of false-positive variants.
CONCLUSIONS: The development of DNA-quality assays will enable researchers to downselect samples or process more DNA to achieve reliable genome sequencing with the highest possible efficiency of cost and effort, as well as minimize the waste of precious samples.

Incorporation of measurement of DNA integrity into qPCR assays.
Brisco M, Latham S, Bartley P, Morley A.
Biotechniques. 2010 Dec49(6): 893-897.
Department of Haematology and Genetic Pathology, Flinders University and Medical Centre, Bedford Park, South Australia, Australia.

Implications of storing urinary DNA from different populations for molecular analyses.
Cannas A, Kalunga G, Green C, Calvo L, Katemangwe P, Reither K, Perkins MD, Maboko L, Hoelscher M, Talbot EA, Mwaba P, Zumla AI, Girardi E, Huggett JF TB trDNA consortium.
National Institute for Infectious Diseases L. Spallanzani, IRCCS, Roma, Italy.
PLoS One. 2009 Sep 104(9): e6985.

Method for isolation of PCR-ready genomic DNA from zebrafish tissues.
Meeker ND, Hutchinson SA, Ho L, Trede NS.
Huntsman Cancer Institute, University of Utah, Salt Lake City, UT 84112, USA.
Biotechniques. 2007 Nov43(5):610, 612, 614.

Background: To date PCR detection of Chlamydia pneumoniae DNA in atherosclerotic lesions from Danish patients has been unsuccessful. To establish whether non-detection was caused by asuboptimal DNA extraction method, we tested five different DNA extraction methods forpurification of DNA from atherosclerotic tissue.Results: The five different DNA extraction methods were tested on homogenate ofatherosclerotic tissue spiked with C. pneumoniae DNA or EB, on pure C. pneumoniae DNA samplesand on whole C. pneumoniae EB. Recovery of DNA was measured with a C. pneumoniae-specificquantitative real-time PCR. A DNA extraction method based on DNA-binding to spin columnswith a silica-gel membrane (DNeasy Tissue kit) showed the highest recovery rate for the tissuesamples and pure DNA samples. However, an automated extraction method based on magneticglass particles (MagNA Pure) performed best on intact EB and atherosclerotic tissue spiked withEB. The DNeasy Tissue kit and MagNA Pure methods and the highly sensitive real-time PCR weresubsequently used on 78 atherosclerotic tissue samples from Danish patients undergoing vascularrepair. None of the samples were positive for C. pneumoniae DNA. The atherosclerotic sampleswere tested for inhibition by spiking with two different, known amounts of C. pneumoniae DNA andno samples showed inhibition.Conclusion: As a highly sensitive PCR method and an optimised DNA extraction method wereused, non-detection in atherosclerotic tissue from the Danish population was probably not causedby use of inappropriate methods. However, more samples may need to be analysed per patient tobe completely certain on this. Possible methodological and epidemiological reasons for non-detection of C. pneumoniae DNA in atherosclerotic tissue from the Danish population are discussed. Further testing of DNA extraction methods is needed as this study has shownconsiderable intra- and inter-method variation in DNA recovery.
Comparison of methods in the recovery of nucleic acids from archival formalin-fixed paraffin-embedded autopsy tissues.
Okello JB, Zurek J, Devault AM, Kuch M, Okwi AL, Sewankambo NK, Bimenya GS, Poinar D, Poinar HN.
McMaster Ancient DNA Centre, Department of Anthropology, McMaster University, Hamilton, Ontario L8S4L9, Canada
Anal Biochem. 2010 May 1400(1): 110-117

Archival formalin-fixed paraffin-embedded (FFPE) human tissue collections are typically in poor states of storage across the developing world. With advances in biomolecular techniques, these extraordinary and virtually untapped resources have become an essential part of retrospective epidemiological studies. To successfully use such tissues in genomic studies, scientists require high nucleic acid yields and purity. In spite of the increasing number of FFPE tissue kits available, few studies have analyzed their applicability in recovering high-quality nucleic acids from archived human autopsy samples. Here we provide a study involving 10 major extraction methods used to isolate total nucleic acid from FFPE tissues ranging in age from 3 to 13years. Although all 10 methods recovered quantifiable amounts of DNA, only 6 recovered quantifiable RNA, varying considerably and generally yielding lower DNA concentrations. Overall, we show quantitatively that TrimGen's WaxFree method and our in-house phenol-chloroform extraction method recovered the highest yields of amplifiable DNA, with considerable polymerase chain reaction (PCR) inhibition, whereas Ambion's RecoverAll method recovered the most amplifiable RNA.