High‐molecular‐weight DNA extraction for long‐read sequencing of plant genomes: An optimization of standard methods

Abstract Premise Developing an effective and easy‐to‐use high‐molecular‐weight (HMW) DNA extraction method is essential for genomic research, especially in the era of third‐generation sequencing. To efficiently use technologies capable of generating long‐read sequences, it is important to maximize both the length and purity of the extracted DNA; however, this is frequently difficult to achieve with plant samples. Methods and Results We present a HMW DNA extraction method that combines (1) a nuclei extraction method followed by (2) a traditional cetyltrimethylammonium bromide (CTAB) DNA extraction method for plants with optimized extraction conditions that influence HMW DNA recovery. Our protocol produced DNA fragments (percentage of fragments >20 kbp) that were, on average, ca. five times longer than those obtained using a commercial kit, and contaminants were removed more effectively. Conclusions This effective HMW DNA extraction protocol can be used as a standard protocol for a diverse array of taxa, which will enhance plant genomic research.

The successful application of third-generation sequencing technologies for sequencing nuclear genomes requires highmolecular-weight (HMW) DNA in sufficient quantity and quality for library preparation and sequencing (Healey et al., 2014). These DNA requirements are often challenging for non-model plant species and represent an important bottleneck for plant genome research; therefore, the development of an efficient HMW DNA extraction method is essential for the plant genomics community. Although several approaches have recently been provided for HMW DNA extraction from plants, they were only applied to a few taxa, required additional purification steps, or the essential factors influencing the process were not adequately discussed (Healey et al., 2014;Mayjonade et al., 2016;Li et al., 2020;Cai et al., 2021;Jones et al., 2021;Mavrodiev et al., 2021;Zerpa-Catanho et al., 2021). Therefore, there is a need for an easy-to-use protocol that can produce HMW DNA from a wide range of plant taxa at a low cost.
In this study, we propose a HMW plant DNA extraction method that combines two classic protocols: (1) a nuclei extraction method (Green et al., 1987) and (2) a cetyltrimethylammonium bromide (CTAB) plant DNA extraction method (Doyle and Doyle, 1987), with modifications. The nuclear extraction step reduces the ratio of organelle genomes in the extracted DNA (Hanania et al., 2004). The CTAB method has been modified in our protocol to solve the problems associated with phenolics and polysaccharides: polyvinylpyrrolidone (PVP) was added to isolate genomic DNA, as suggested by Healey et al. (2014). To more efficiently meet the needs of genome sequencing, our combined protocol includes (1) improvements to optimize time and reagent requirements and (2) suggestions of favorable conditions for factors influencing the results (number of pipetting steps, grinding time in liquid nitrogen, and centrifugation force in g). A combination of these two classic protocols has already been proposed for high-quality DNA extraction from Vitis vinifera L. (Hanania et al., 2004), but not with regard to HMW DNA and applicability in other taxa. Similarly, a method combining the nuclear isolation process and sodium dodecyl sulfate (SDS)-based DNA extraction protocol has recently been proposed for HMW DNA extraction (Zerpa-Catanho et al., 2021); however, its effectiveness has only been confirmed in a few plant taxa (six genera in three families), and it requires an extra purification step (QIAGEN Genomic Tip 20/G columns; QIAGEN, Hilden, Germany). By contrast, we have assessed the broad applicability of our protocol in species representing 18 orders of flowering plants from all major angiosperm lineages (Angiosperm Phylogeny Group, 2016), as well as a gymnosperm, Pinus L.
To confirm the effectiveness of our HMW DNA extraction method, we compared the results with those obtained using a commercial plant DNA extraction kit. The DNA length distributions and purity were evaluated as validation criteria for comparing the two methods. We also discuss factors influencing the results, such as the number of pipetting steps, grinding time in liquid nitrogen, and centrifugation force in g.

HMW DNA extraction method
We sampled leaves of species from each of 18 major angiosperm orders and one gymnosperm to test the taxonspecific efficiency of our protocol. For details of all samples used in this study, see Appendix 1. Reagents, recipes, and a stepwise protocol are provided in Appendix 2. Our HMW DNA extraction protocol consists of three major steps: (1) grinding and nuclei isolation, (2) nuclear DNA extraction using CTAB buffer, and (3) RNase A and proteinase K treatment. We started with 2 g of tissue (preferably fresh, young leaves) and used a vacuum-aided cell strainer (40 μm and 100 μm; pluriSelect Life Science, Leipzig, Germany) to collect the nuclei suspension. We also conducted additional DNA extractions using the same samples from our HMW DNA extraction protocol. For this, we employed the Exgene Plant SV kit (GeneAll Biotechnology, Seoul, Republic of Korea), a commercial plant DNA extraction kit based on the DNA-binding filter method. Following the instructions in the manufacturer's manual, we used 0.1 g of leaf tissue, which is the recommended amount for fresh leaves.

Grinding and nuclei isolation
The protocol starts with 2 g of fresh, young leaves. We ground the leaves into a powder in liquid nitrogen (−80°C) and placed the powder in 20 mL of nuclei isolation buffer (IB). After 30 s of vortexing, we added Triton X-100 (20 μL) and β-mercaptoethanol (1.5 mL). This step should be conducted inside a fume hood as β-mercaptoethanol is toxic. The samples were placed on ice for 10 min, and then the mixture was filtered through a 100-µm cell strainer (pluriStrainer 100 µm; pluriSelect Life Science) seated in a 50-mL conical tube to collect the nuclear suspension. During filtration, gently scraping plant material accumulated on the filter with the side of a 1000-µL pipette tip may facilitate a smoother filtration. The filtering step was repeated with a 40-µm cell strainer (pluriStrainer 40 µm; pluriSelect Life Science), and Triton X-100 (200 µL) was added to the obtained nuclear suspension. This process lyses the cell and organellar membranes but not the nuclear membrane (Peterson et al., 1997). As a non-ionic detergent, Triton X-100 facilitates the release of nuclei from cells and prevents nuclei from clumping (Loureiro et al., 2007). To pellet the nuclei, the samples were centrifuged, and the supernatant was discarded. Centrifugation for 10 min at 3000 × g (4°C) is recommended to prevent fragmenting long DNA molecules (see Results).

Nuclear DNA extraction using CTAB buffer
The nuclei pellet was resuspended in 5 mL of Carlson Lysis Buffer (Carlson et al., 1991). Adding β-mercaptoethanol (12.5 µL) denatures globular proteins to make them insoluble in water (Jadhav et al., 2015). An incomplete resuspension can reduce yield; thus, we incubated the samples at 65°C for a minimum of 15 min for efficient resuspension. If the pellet still does not suspend, crushing the pellet with a pipette tip might be helpful. For easy handling, we transferred the suspended nuclei pellet to a 15-mL tube instead of proceeding with the 50-mL tube. We added 5 mL (equal volume) of chloroform:isoamyl alcohol (24:1 [v/v]) to remove impurities. During this step, chloroform (CHCl 3 ; a non-polar 3-hydrophobic solvent) dissolves non-polar proteins and lipids to promote the partitioning of lipids and cellular debris into the organic phase. Isoamyl alcohol (C 5 H 12 O) prevents the emulsification of the solution (Jadhav et al., 2015). After centrifugation (3000 × g for 10 min at 4°C), the aqueous upper phase containing DNA was collected and transferred into a new tube, while the organic phase containing lipids, proteins, and other impurities was discarded. The separation of a pure aqueous phase is critical for the purity of the end product, and we recommend collecting just four-fifths of the upper liquid volume to avoid including any cellular debris. Adding the proper ratio of sodium acetate (NaOAc) and isopropanol to the acquired supernatant is essential for precipitating the DNA: for every 10 mL of supernatant, a 1/10 volume of 3 M NaOAc (1 mL) and the same volume (including NaOAc) of room-temperature isopropanol (11 mL) are needed. It is essential to use roomtemperature isopropanol for this step; otherwise, both polysaccharides and DNA will precipitate (Shepherd and McLay, 2011). The precipitated DNA was separated from other solvents through centrifugation (3000 × g for 10 min at 4°C), and the resulting DNA pellet was washed with 70% cold ethanol, recentrifuged (3000 × g for 10 min at 4°C), and thoroughly dried. We recommend rapidly drying samples using room-temperature air blown by a hair dryer.

RNase A and proteinase K treatment
The DNA pellet was dissolved in 2 mL Tris-EDTA (TE) buffer. To remove RNA and protein efficiently, which account for most of the impurities in extracted DNA, we treated the samples with RNase A (10 mg/mL) and proteinase K (>600 units/mL), respectively. For each treatment, the proper incubation time and enzyme activation temperature are important: 5 min at 37°C for RNase A and 15 min at 50°C for proteinase K. The enzymes used in each step are removed by a treatment with 2 mL of chloroform: isoamyl alcohol (24:1 [v/v]). After treatment with RNase A and proteinase K, the same precipitation procedure as for the CTAB extraction is followed. The resulting pellet is dissolved using an appropriate amount of deionized water (50-500 µL) according to the size of the pellet (recommended final concentration is ca. 200 ng/µL). If it is difficult to dissolve the pellet, we recommend incubating the tube at 50°C. If the pellet remains after incubation at 50°C, it is recommended to take only the dissolved aqueous layer after brief centrifugation.

Quality evaluation of extracted DNA
The quantity and purity (A 260 /A 280 and A 260 /A 230 ratios) of the extracted DNA were measured using a Qubit 4 Fluorometer (Thermo Fisher Scientific, Waltham, Massachusetts, USA) and a NanoDrop 2000 spectrophotometer (Thermo Fisher Scientific), respectively. The length distribution of the extracted DNA was evaluated using a Femto Pulse system (Agilent Technologies, Santa Clara, California, USA).

Optimization of conditions for HMW DNA recovery
We tested three factors influencing the results: (1) the number of pipetting steps, (2) the grinding time in liquid nitrogen, and (3) the centrifugation force in g. Three independent experiments were performed on different taxa in each case to evaluate each factor. First, we tested the impact of high g forces during centrifugation on DNA damage by comparing the setting in our protocol (3000 × g; control group) and a higher setting (5000 × g; experimental group). Second, the amount of grinding was compared. The control group was subjected to one minute of grinding (ensuring the sample was fully chilled before grinding began). The experimental group was subjected to an additional two minutes of grinding after adding extra liquid nitrogen. Third, we assessed whether high-speed and multiple pipetting steps could potentially damage DNA. We conducted pumping at the maximum-achievable speed 200 times in a tube using a P200 tip (experimental group) and compared the resulting DNA size distribution with the original DNA (control group).

DNA quantity, size, and purity measurements
Usually, the quantity of the end DNA product per extraction is enough to generate 4-5 libraries (8-15 µg) for long DNA sequencing with MinION or GridION (Oxford Nanopore Technologies, Oxford, United Kingdom), based on the library construction protocol (Ligation Sequencing Kit). The measurements obtained through the Femto Pulse system (peak height and percentages of fragments >20 kbp in the fragment-length distribution graph) confirm that our protocol successfully produced DNA fragments an average of five times longer than those generated using the commercial kit (Table 1, Figure 1), although the results of our standard HMW DNA extraction protocol showed different patterns depending on the taxon (Figure 2A, B). With our protocol, the taxon with the highest portion of >20-kbp fragments was Chloranthus fortunei Solms (Chloranthales; 83.6%), and the longest peak of DNA fragment distribution was obtained from Alisma plantago-aquatica subsp. orientale (Sam.) Sam. (Alismatales; 183.0 kbp) ( Table 1). In the most efficient instance, our protocol yielded 35 times more DNA fragments over 20 kbp (77.1%) in Lysimachia davurica Ledeb. (Ericales) than the commercial kit, for which only 2.2% of fragments were greater than 20 kbp.
The quality of DNA extracted using the HMW method was superior to that obtained using the kit method in most samples. In the context of next-generation sequencing, highquality DNA is characterized as predominantly HMW with an A 260 /A 280 ratio over 1.8 and without contaminating substances, such as polysaccharides or phenolics (Kasem et al., 2008;Desjardins and Conklin, 2010). With both methods, the A 260 /A 280 absorbance ratio, which measures protein contamination, showed similar results with low contamination (both averaged 1.83); however, our standard protocol more effectively removed carbohydrates and organic solvents (average A 260 /A 230 ratio = 1.88) than the commercial kit (average A 260 /A 230 ratio = 1.49) (Table 2; Figure 2C, D). Generally, A 260 /A 230 values between 1.8-2.2 indicate DNA is free of carbohydrates and organic solvents (Kasem et al., 2008;Desjardins and Conklin, 2010).
To address the statistical difference between the results from our protocol and a commercial kit, we performed paired t-tests on all pairs of DNA length and quality, with P < 0.05 considered significant. In the DNA length criteria, the peak height and percentages of fragments >20 kbp show significant differences (P = 0.002 and P = 1.66e-05, respectively). Because the DNA extracts from our method and the commercial kit both showed excellent A 260 /A 280 ratios (both averaged 1.83), their quality was not significantly different (P = 0.7769); however, the A 260 /A 230 ratio was significantly different (P = 0.001), indicating that our protocol provided an advantage.

Factors influencing the results
Pipetting: avoid fast and frequent pipetting Some long DNA extraction protocols suggest pipetting as little as possible or using a wide-bore tip to avoid shearing (Zerpa-Catanho et al., 2021). We confirmed that highspeed repeated pipetting damages DNA. For samples extracted based on our protocol (control group), 17.0% of the DNA fragments were >50 kbp in length, with the peak being 76.46 kbp. In contrast with the control, the sample subjected to repeated pipetting (200 times) at high speed (experimental group) yielded just 13.4% of the fragments >50 kbp, with the peak being 49.93 kbp (Appendix 3A). High-speed over-pipetting, therefore, does affect HMW DNA extraction. The number of pipetting steps in our extraction protocol is fewer than 20, which is recommended to be performed gently with wide-bore tips to reduce the likelihood of DNA shearing.
Grinding: avoid excessive grinding Generally, it is important to grind samples as long as possible (at least 25 min or more [Circulomics, 2021], although in practice the grinding time is much shorter) in DNA extraction to transform the plant tissue into a powder. Excessive grinding can provide a yield advantage, but it can also shear the DNA. The DNA sizes of the samples ground to different degrees were compared with the Femto Pulse system, and we concluded that additional grinding for 2 min (experimental group) has a negative effect on DNA fragment length (Appendix 3B). One minute of grinding is optimal, and additional liquid nitrogen is not needed.
T A B L E 1 A comparison between our HMW DNA extraction method and a commercial kit. Fragment lengths were estimated using the Femto Pulse system.  Centrifugation: avoid high speeds Centrifugation causes DNA molecules to collide, resulting in their molecular structure being subjected to high shearing forces (Peterson et al., 2012). Our results showed that the DNAs of the experimental group (centrifuged at 5000 × g) are more fragmented than the control group (centrifuged at 3000 × g), with measured DNA peaks (indicating HMW DNA) of 88.73 kbp and 146.94 kbp, respectively, yielding 30.0% and 39.7% of DNA fragments >50 kbp, respectively (Appendix 3C). Given this difference in producing very long fragments (e.g., >50 kbp or more), high-speed centrifugation over 3000 × g is not recommended.

DISCUSSION
Here, we focused on optimizing and standardizing a HMW DNA extraction protocol for various plant taxa using economical techniques. We confirmed that our protocol successfully produced HMW DNA from various taxa in most cases; however, we expect that the experimental results will differ depending on the taxa investigated because each species has a different polysaccharide or phenolic content. Although not all species yielded good results using our protocol, we nevertheless confirmed that our protocol yielded DNA superior to the commercial kit in terms of length and purity, with statistically significant results.
To evaluate the results of various HMW DNA extraction methods, it is important to select an appropriate method and instrument with which the results can be compared. A common method to evaluate the length of the extracted DNA is a visualization of the position and brightness of DNA bands using electrophoresis through a low-concentration agarose medium (typically 0.7%) containing ethidium bromide. Alternatively, more efficient electrophoresis can be performed using a pulse-field power supply (e.g., Pippin Pulse system; Sage Science, Beverly, Massachusetts, USA). We tried pulse-field electrophoresis to check the quality of HMW DNA at the initial stage of our study; however, we confirmed that the result (the brightest position of a smeared DNA band) varied depending on the amount of loaded DNA (Appendix 4). As the quantity of DNA loaded in the agarose gel for electrophoresis is increased, the brightest position of the DNA band is shifted to a higher position (a position of higher molecular weight); that is, the quantity of DNA and the brightest position of the DNA band are positively correlated. Special attention is therefore needed to ensure that the same quantity of DNA is used for each sample when evaluating DNA length using pulsefield electrophoresis. Several automated electrophoresis techniques with fluorescence dye have been proposed for DNA length analysis to improve the unstable ethidium bromide visualization in normal electrophoresis (including pulse-field). Although the TapeStation (Agilent Technologies) and the Fragment Analyzer (Agilent Technologies) are frequently used for size evaluations of extracted DNA fragments, they are not sensitive enough to separate HMW DNA (>60 kbp is not recommended in either instrument; Agilent Technologies, 2020a). Remarkably, the latter was used in a study of the development of a HMW DNA extraction protocol (Zerpa-Catanho et al., 2021). By contrast, the Femto Pulse system is the automated pulsed-field instrument designed for the purpose of analyzing HMW DNA. An automated pulsed-field power supply in the Femto Pulse system allows the separation of DNA up to 165 kbp (Agilent Technologies, 2020b).

CONCLUSIONS
The protocol introduced here can be used to efficiently extract HMW DNA using standard laboratory equipment (an average peak of 77.88 kbp and an average of 67.53% of fragments >20 kbp). Given its success with diverse flowering plant species and one gymnosperm, we hope our method will contribute to plant genome studies as a broadly applicable protocol for poorly studied taxa. Additional investigations comparing DNA length, purity, and extraction cost between our protocol and commercial HMW DNA extraction kits will provide a more comprehensive understanding of the benefits of our approach.
AUTHOR CONTRIBUTIONS S.K. and A.C. developed the experimental protocol. M.K. and S.K. performed all experiments and analyses. S.K. and M.K. wrote the preliminary manuscript draft, and all authors revised and approved the manuscript before submission. All authors approved the final version of the manuscript.

ACKNOWLEDGMENTS
This work was supported by the Sungshin Women's University Research Grant of 2020 to S.K.

DATA ACCESSIBILITY STATEMENT
The taxonomic locations, vouchers, herbarium, and collection sites of all species used in this study are provided in Appendix 1.
T A B L E 2 A comparison between our HMW DNA extraction method and a commercial kit. DNA purity was evaluated using a NanoDrop.