Assessment and Comparison of Two Different Methods to Extract Nucleic Acids From Individual Honey Bees

Abstract

The honey bee is an excellent model system to study behavioral ecology, behavioral genetics, and sociogenomics. Nucleic acid-based analyses enable a broad scope of research in functional genomics, disease diagnostics, mutant screening, and genetic breeding. Multiple levels of analysis lead to a more comprehensive understanding of the causes of phenotypic variation by integrating genomic variation, transcriptomic profiles, and epigenomic information. One limitation, however, is the sample preparation procedures to obtain high quality DNA and RNA simultaneously, particularly from small amounts of material, such as tissues of individual bees. We demonstrate that it is feasible to perform dual extractions of DNA and RNA from a single individual bee and compare the quality and quantity of the extracted nucleic acids using two different types of methods. There was a greater total yield of DNA and RNA from ethanol-based extractions with minimal differences in overall concentration in ng/uL. We describe here the first validated method for dual extraction of DNA and RNA specifically from individual honey bees (Apis mellifera).

Honey bees (Apis mellifera) are one of the primary species of social insects used for genomic biology studies. Their caste-specific social behavior, phenotypic plasticity, and genetics are different from other nonsocial model insect species such as Drosophila, and they were one of the first insects to be sequenced (Weinstock et al. 2006). Continued interest in honey bee health and the availability of its genome (Elsik et al. 2014) make it a key genomic study organism for the foreseeable future. In addition to studies using the honey bee as a model system to study social evolution, behavioral ecology, physiology, and genetics, this species is also an excellent model species for behavioral epigenetics (Yan et al. 2014, Yan et al. 2015, Li-Byarlay 2016) with the first described functional CpG methylation system in insects (Wang et al. 2006). CpG methylation plays an important role in the nutritional control of reproductive status to develop into a queen or a worker bee (Kucharski et al. 2008), as well as the phenotypic plasticity of workers (Herb et al. 2012). Simultaneous studies at the DNA and RNA level revealed that CpG methylation alters gene splicing in honey bees (Foret et al. 2012, Li-Byarlay et al. 2013, Wang and Li-Byarlay 2015).

Honey bees are also the most important managed pollinator in agriculture, required for the pollination of an annual $15 billion worth of agricultural crops (Bond et al. 2014) and responsible for pollinating a third of human food crops (Gallai et al. 2009). Thus the ongoing annual 30–40% mortality of honey bee colonies (Kulhanek et al. 2017) is a major concern. Several factors are thought to contribute to this mortality, including parasitic mites, pathogens such as bacteria and viruses, pesticide use in agriculture, climate change, poor nutrition, and low genetic diversity (Rinderer et al. 2010, Wu et al. 2011, Degrandi-Hoffman et al. 2015, Goulson et al. 2015, Wang and Li-Byarlay 2015, Simone-Finstrom et al. 2016, Gisder and Genersch 2017, Hamiduzzaman et al. 2017). Understanding the molecular biology and genomics of honey bees is crucial to countering the threats to honey bee mortality (Grozinger and Robinson 2015).

Genome sequencing and transcriptome studies in honey bees have become increasingly common since the completion of the honey bee genome in 2006 (Weinstock et al. 2006). However, the processes of obtaining experimental samples and tissues are not always consistent. Due to the limitation of technologies, DNA and RNA material are often not obtained from the same specimen but instead are extracted from different pools of specimens for next generation genomic sequencing. This creates variation that can cause misinterpretation or masking of important data due to variation among individuals (Sultan et al. 2007). Using pooled samples masks individual to individual variation and can obscure real differences in experimental results (de Jong et al. 2010). It is therefore highly beneficial to obtain both genomic and transcriptomic information from the same individual for a truly integrative study of the genome and its transcriptome product as one recent report from the honey bee brains showed (Jones et al. 2020).

The extraction of DNA and RNA molecules is the most critical method used in genomic biology (Tan and Yiap 2009) because any errors at the start will have amplified effects for downstream sequencing procedures and data analysis. Extracting high quality DNA and RNA simultaneously is also central to the studies of gene expression and gene regulation to address phenotypic plasticity and social behavior of honey bees, e.g., by eQTL mapping (Gilad et al. 2008), as well as to other interesting areas of Apis mellifera biology such as the high recombination rate (Beye et al. 2006, Wilfert et al. 2007, Wallberg et al. 2015, Rueppell et al. 2016). Dual extraction is the process of isolating the DNA and RNA from the same sample, eliminating genetic masking from pooled samples, and allowing direct comparisons between the RNA and DNA to be made at the individual level (Triant and Whitehead 2009). Studying DNA and RNA from the same individual facilitates a more accurate interpretation of the relationship between genotype and phenotype and, therefore, a potentially deeper understanding of the molecular basis of diseases and other complex traits (Reuter et al. 2016). Since dual extraction is not yet a standardized method in honey bees, researchers typically use DNA and RNA from different bee samples to carry out singular extractions. However, there are dual DNA and RNA extraction kits available for purchase on the market today. To our knowledge, these kits have not been systematically compared in honey bees, but they hold great promise and have been used on plant and animal tissues, bacteria, yeast, fungi, algae, viruses, cultured cells, and other insects.

Here, we address two important questions for implementing dual extraction in honey bees. First, we determine whether dual extraction is feasible using honey bee samples from three different life stages (larva, pupa, and adult). Second, we quantify the differences in terms of yield and purity with the different extraction procedures. We demonstrate that the dual extraction technique can be applied to individual honey bees and produce meaningful quantities of DNA and RNA, which can be used for truly integrative genomic analyses at the individual level.

Materials and Methods

Samples

All honey bee larvae, pupae, and adults were acquired from hives at the Central State University apiary in Wilberforce, Ohio. Up to eight samples were taken for each developmental stage: larval (fifth instar, L5), pupal (brown eye stage), and adult heads (one day old). Each bee sample's weight was weighed by using an analytical balance scale (Thermo Fisher Scientific accu-124, Waltham, MA). Before running the DNA or RNA isolations, the samples were flash frozen and stored in a −80 °C freezer.

Method 1

The Zymo Research dual extraction kit was purchased from Zymo Research (D7003, ZR-DuetTM DNA/RNA MiniPrep Plus Kit, CA). This extraction kit uses a column technique to isolate DNA and RNA from samples, and we followed the manufacturer's instructions. The weight of each sample is within the range of 25–50 mg and listed in the Supplemental Table 1 (online only). Frozen samples were homogenized and lysed to break open cells. Homogenization was carried out using a plastic micropestle (12-141-364, FisherScientific, USA) to grind tissues in a 1.5 ml micro-centrifuge tube using a cordless motor (Pellet Pistol #749540-0000, Kimble). After homogenization, lysis buffer was added to the sample and further homogenized with a plastic micropestle (12-141-364, FisherScientific, USA) for maximum extraction. During the initial stages of the protocol, DNA and RNA were separated by specifically binding DNA to a column while RNA was in the flow-through supernatant. Each sample was centrifuged at 10,000 × g for 30 s, and then transferred into the DNA column. Then the column was washed by prep buffer, wash buffer, and then eluted with 50 µl of DNase/RNase-free water. Each RNA sample was obtained from the mixture of 100% ethanol and DNA flow-through (1:1), then transferred to an RNA column with a collection tube and centrifuged. The prep buffer, wash buffer was added to the column and centrifuged. RNA was eluted in 50 µl of DNase/RNase-free water.

Method 2

The MasterPure dual extraction kit was purchased from Epicentre (MC85200, WI). The weight of each sample is within the range of 1–5 mg according to the manufacture's guidelines and listed in the Supplemental Table 1 (online only). The homogenization and lysation were carried out in a manner similar to Method 1 in 1.5 ml microcentrifuge tubes. Tissue samples were dipped into liquid nitrogen, homogenized in Proteinase K and cell lysis solution with the use of a cordless motor (Pellet Pistol #749540-0000, Kimble), further incubated at 65°C for 15 min, vortexed at 5-minute intervals, and stored on ice. Samples received 150 µl MPC Protein Precipitation Reagent and were centrifuged at 4°C for 10 min at 10,000 × g before receiving 500 µl of isopropanol and inverted 40 times. DNA and RNA were retained in the supernatant, then after precipitation through further centrifugation at 4℃ and 10,000 × g and washed, total nucleic acid was treated with RNase to obtain total DNA and treated with DNase to obtain total RNA. This was performed by using up to 5 mg of tissue from the same bee for each extraction. DNA samples received 1 µl of 5 µg/uL RNase A after homogenization and mixed thoroughly then incubated at 37℃ for 30 min. DNA was removed from RNA samples with the use of 200 µ l DNase 1 solution created from 5 µ l of RNase-Free DNase 1 and 195 µ l of 1X DNase Buffer and incubated at 37℃ for 10 min All samples received two washes with 70% ethanol before being suspended in 35 µ l TE buffer for analysis and storage. RNA samples also received 1 µ l of RiboGuard RNase inhibitor for protection from degradation while being stored in the TE buffer solution.

Yield and Purity

Quantity of the resulting RNA and DNA samples was assessed using a QFX fluorometer (DeNovix, USA) and Qubit RNA Assay Kit, and Qubit DSDNA BR Assay kit (Q32852, Q32850, ThermoFisher, USA). The internal Qubit quantification algorithm was used to determine the quantity. Quality of samples was determined using the RNA Integrity Number, determined by the ratio of 28S RNA–18S RNA, and the overall size of DNA samples was determined through Bioanalyzer analysis (G2939BA, CA) at The Center for Genomics Research at Wright State University. The total yield of each sample (Y in the unit of nanogram) is obtained by multiplying the DNA or RNA concentration (C in the unit of nanogram/microliter) by the final total volume (V in the unit of microliter), and then divided by the tissue mass (M in the unit of milligram) (Formula below and Supp Table 1 [online only]).

$${\displaystyle \begin{array}{l}{\displaystyle \mathrm{Y}=\mathrm{C}\ast \mathrm{V}/\mathrm{M}}\end{array}}$$

Data Analysis

A one-way analysis of variance (ANOVA) was calculated for comparing multiple treatments, used to determine the differences among means of the different populations. The online tool Interactive Dotplot (Weissgerber et al. 2017) was used to generate all the violin plots. Outliers were identified using the mean plus or minus the two folds of standard deviations.

Results

To compare the larval samples, we extracted total DNA and RNA using both methods. Our data indicated that between the two methods, there is a significant difference in the yield (F_1,12 = 16.10, P = 0.002, Fig. 1). For RNA, we harvested significantly higher yield via Method 2 (F_1,11 = 23.01, P < 0.001; Fig. 1).

Fig. 1.

Violin plots show the comparison of yield of total DNA and RNA of all larvae tested. Each circle represents a sample. The total yield of each sample is obtained by multiplying the DNA or RNA concentration by the final total volume, and then divided by the tissue mass.

In addition, we compared the pupal samples and extracted total DNA and RNA using both methods. For the concentration of DNA, there was a significant difference of yield between two methods (F_1,11 = 7.44, P = 0.02, Fig. 2). As to RNA, we noticed a different trend in total yield of DNA and RNA between the two methods, the yield of RNA between two methods was not significantly different (F_1,10 = 39.48, P < 0.001; Fig. 2).

Fig. 2.

Violin plots show the comparison of yield of total DNA and RNA of all pupae tested. Each circle represents a sample. The total yield of each sample is obtained by multiplying the DNA or RNA concentration by the final total volume, and then divided by the tissue mass.

To cover most developmental stages, we collected whole head tissues of adult worker bees to compare the yield. There was a significant difference in DNA yield (F_1,11 = 7.07, P = 0.02; Fig. 3) and RNA (F_1,11 = 22.88, P < 0.001; Fig. 3) with Method 2 yielding more than Method 1.

Fig. 3.

Violin plots show the comparison of yield of total DNA and RNA from whole head of adults tested. Each circle represents a sample. The total yield of each sample is obtained by multiplying the DNA or RNA concentration by the final total volume, and then divided by the tissue mass.

The quality of all RNA samples was assessed by the RNA integrity number obtained through bioanalyzer analysis. All data exhibited a ratio between 1 and 7.2, which is acceptable for further experiments. The differences in quality between the two techniques were not significantly different for RNA (F_1,14 = 1.15, P = 0.30, Fig. 4).

Fig. 4.

Violin plots show the comparison of RNA Integrity Number (RIN) of all RNA samples from bioanalyzer analysis. Each circle represents a sample.

The size and quality of DNA samples were then compared using data from bioanalyzer analysis, and there was no significant difference in the overall size of DNA obtained from each method (F_1,23 = 0.86, P = 0.36, Fig. 5). The quality of selected samples was shown in Fig. 6 by the gel pictures from the bioanalyzer analysis.

Fig. 5.

Dot plots show the comparison of DNA size (in base pair) of all DNA samples from bioanalyzer analysis. In the histogram. The top and bottom of the black rectangle indicate the third and first quartile. The white dot inside is the Median. The top and bottom of whiskers show the upper and lower adjacent values.

Fig. 6.

Gel images from bioanalyzer analysis to show the quality of DNA from selected samples. Methods are listed on the top of each sample ID. The DNA ladder is listed on the first lane on the left.

Discussion

Both methods enabled the dual extraction of RNA and DNA from individual honey bees at the larval, pupal, and adult life stages. Comparing the DNA yield between the techniques and each life stage, Method 2 (following a more traditional procedure) produced significantly higher yields than Method 1 (using columns). This result extends to the RNA isolations; Method 2 showed significantly higher yields of RNA for the larval, pupal, and adult life stages.

There were some notable differences between the two dual extraction kits. For instance, Method 1 had a faster turn-around time (40 min per sample, on average) and used a column-based technique. Method 2 took one hour per sample on average and uses a precipitation-based technique. Nonetheless, we demonstrate here that DNA and RNA can be dually extracted from individual honey bees. Although the yield is higher with Method 2, the faster extraction time makes Method 1 more desirable for some applications.

If we consider the time for performing the experiments, it is usually faster to use the column-based (Method 1) than the traditional alcohol precipitation (Method 2). However, Method 1 has two major drawbacks: 1) membrane spin columns cost more than traditional Method 2 of purification, and 2) membrane spin columns have specific abilities to select certain size of nucleic acid they can purify, excluding very large (>50 kb) or small nucleic acids (<100 bp) (Dowhan 2012). Based on our comparison, we see a significant lower concentration and lower yield of DNA and RNA using Method 1, which are membrane spin columns. There was neither a significant difference between the length of DNA obtained from each method nor a significant difference in RNA quality amongst the two methods.

In general, precipitation versus column methods are dependent on the desired aim of the study. When looking for the greatest yield, precipitation methods will be more effective than the column-based methods of extraction. Precipitation methods run into issues when discussing the overall purity of the extracted samples, which is when the column-based extractions are superior. The overall advantage of dual genomic extractions arises from the RNA providing an insight into the total gene expression of the individual at the moment of extraction. Further study into dual extractions amongst castes can provide greater insight into differences in gene expression.

The future technology of genomic biology may focus on novel ways to sequence DNA and RNA simultaneously to gain a complete understanding of the genomic network and gene regulation in cells (Lee et al. 2020). More social insects will be sequenced as the cost of sequencing decreases (http://antgenomics.dk/, http://i5k.github.io/) (Robinson et al. 2011). Our recent research of dual extraction of DNA and RNA reveals the importance of DNA methylation marks and how they may play a critical role in gene expression and alternative splicing (Li-Byarlay et al. 2020). In addition, previous studies showed investigations from both DNA and RNA sequencing levels are important to reveal novel molecular mechanisms for behavioral research of social insects (Foret et al. 2009, Herb et al. 2012, Galbraith et al. 2015, Standage et al. 2016).

In summary, we demonstrate that dual extraction is feasible for larval, pupal, and adult (head) life stages. The results also show that the ethanol precipitation (Method 2) produced much higher yields of both DNA and RNA from the samples than column-based extractions (Method 1). It is possible to use individual honey bees from most stages of life to extract DNA and RNA with these protocols. Using separated samples of the same individuals (as opposed to pooled sample of different individuals) can increase precision and accuracy while decreasing inter-individual noise from data by eliminating genetic variations because of other intrinsic differences among individuals. Dual extraction of individuals and even separate tissues of individuals like the head is a viable option for honey bee research.

Acknowledgments

This research was supported by funding from the National Research Council via a Senior Research Associateship to HL-B, United States Department of Agriculture (USDA) -National Institute of Food and Agriculture- Evans Allen fund (NI191445XXXXG002) to HL-B and RS, National Science Foundation Historically Black College and University-Udergraduate Program Rsearch Initiation Award grant (1900793) and USDA Sustainable Agriculture Research and Education grant (ONC19-062) to HL-B, NCSU Undergraduate Research Scholarship to BG. We thank Mike Markey at Wright State University for technical assistance on the bioanalyzer and helpful discussion about the manuscript.