A highly efficient scheme for library preparation from single-stranded DNA

Abstract

Although methods for sequencing library preparation from double-stranded DNA are well established, those from single-stranded DNA (ssDNA) have not been well studied. Further, the existing methods have limitations in efficiency and yield. Therefore, we developed a highly efficient procedure for sequencing library preparation from ssDNA. In this method, the first adaptor tagging of ssDNA is performed using terminal deoxyribonucleotidyl transferase (TdT)-assisted adenylate connector-mediated ssDNA (TACS) ligation, which we reported recently. After complementary strand synthesis using the adaptor-tagged ssDNA, second adaptor tagging via Vaccinia virus topoisomerase I (VTopoI or TOPO)-based adaptor ligation is performed. With additional steps for degradation, repression, and removal of the adaptor dimer, the proposed TACS-TOPO scheme realizes adaptor dimer-free sequencing library preparation from ssDNA samples of 24 pg. The TACS-TOPO scheme was successfully applied to cell-free DNA analysis with amplification-free library preparation from 50 µL of human serum. A modified TACS-TOPO scheme was also applied to DNA extracted from ancient human bones, bringing two to eight times more library yields than those using a conventional library preparation protocol. The procedures for preparing VTopoI and its complex with a double-stranded oligonucleotide adaptor are also described. Overall, the proposed TACS-TOPO scheme can facilitate practical and sensitive sequencing analysis of ssDNA.

Figure 1

Formation and purification of the VTopoI-oligonucleotide complex (VOC). (A) Scheme of complex formation. The formation of VOC is an equilibrium reaction. Phosphorylation of released DNA effectively increases the amount of complex formed. (B,C) An example of VOC formation. The effects of T4 PNK and buffer compositions on complex formation are shown. Gel images of SDS-PAGE (B) and agarose gel electrophoresis with E-Gel Ex (C) are shown. VTopoI expressed with pColdTF was used for VOC formation. The expected VTopoI protein expressed with the pColdTF vector is 91.0 kDa. (D) Purification of the VOC with Heparin column chromatography. SDS-PAGE (top) and chromatogram (bottom) analyses of the fractions are shown. The blue and red lines in the chromatogram indicate absorbance at 280 nm and the solution B content, respectively. The numbers on the gel image and the chromatogram indicate the fraction numbers. The VOC is eluted earlier than VTopoI. The original gel images for B, C, and D are provided in Supplementary Figure S5.

Figure 2

Substrate specificities of the VTopoI-oligonucleotide complex (VOC). (A) Structures of VOCs (left) and TOPO substrates (right). (B) Images of gel electrophoresis after mixing the VOC and TOPO substrates. Blunt (right) and TA (left) types of VOC were investigated for their reactivity with substrates with or without the 5′-phosphate and one base protrusions. If ligation occurs, the band shifts upward. The symbols indicate the following: o, hydroxyl end; p, phosphorylated end; A, C, G, or T, terminal nucleotide attached at the 5′-end of the substrate. For experimental details, see Supplementary Methods. The original gel images for B are provided in Supplementary Figure S6.

Figure 3

Comparison of library preparation schemes from single-stranded DNA. (A,B) TACS-T4 (A) and TACS-TOPO (B). The scheme (left) and analysis of DNA at each reaction step with denaturing gel electrophoresis (right) are shown. The numbers in italic fonts indicate the length of DNA used. The colored stars in the left scheme correspond to the bands in the right gel images. The lanes are labeled as follows: (1) before TACS ligation; (2) after TACS ligation; (3) after complementary strand synthesis with Taq DNA polymerase; (4) after second adaptor tagging; and (5) after SPRI purification. (C) A gel image showing the analysis of PCR-amplified libraries. DNA fragments amplified from libraries prepared using TACS-T4 (lanes 1 and 2: technical duplicates) and TACS-TOPO (lanes 3 and 4: technical duplicates) were loaded onto a denaturing gel. For both libraries, 20 cycles of PCR amplifications were conducted. Adaptor dimers were only evident in the libraries prepared with TACS-T4. The original gel images are provided in Supplementary Figure S7.

Figure 4

Improvements to the TACS-TOPO scheme. (A) The TACS-TOPO scheme. (B) Improvements applied to different versions of the protocol. OPC: oligonucleotide purification cartridge grade, HPLC: high-performance liquid chromatography purification grade. (C) The first version (ver. 1) produced several adaptor dimers. (D) The adaptor dimer degradation scheme introduced in the ver. 4 protocol. The uracil residue in the adaptor is degraded using UDG and APE 1 before PCR amplification of the library. This treatment causes the degradation of the adaptor dimer. (E) Library preparation using the TACS-TOPO ver. 4 protocol. The sensitivity of dimer-less library preparation was improved to realize highly sensitive library preparations from ssDNA. The original gel images for C and E are provided in Supplementary Figure S8.

Figure 5

Sequencing library preparation from cfDNA. (A) Comparison of libraries prepared using TACS-T4 and TACS-TOPO. (B) Library yields of TACS-TOPO ver. 4 from various volumes of serum are shown. The zones separated with horizontal lines indicate the magnitude of amplification needed for sequencing to obtain 90 Gb reads. (C) Gel image showing the analysis of amplified cfDNA libraries prepared using TACS-TOPO ver. 4. After denaturing gel electrophoresis on a 6% gel, the SYBR gold-stained gel was photographed. (D) Size distribution of the reads mapped to the human reference genome. A library prepared from 50 µL serum using TACS-TOPO ver. 4 was sequenced and mapped to the human reference genome. (E) Colocalization of mapped reads on G4-seq peaks and G4 motifs. From the mapped reads, fragments of 35–75 nt were selected and analyzed for their colocalization with the G4-seq peaks and G4 motifs. The same serum (No. 1 in Supplementary Table S2) was used for (A–E). (F–H) Sequencing libraries prepared from 50 µL serum or plasma obtained from several sources are shown (Supplementary Table S2). The library yields (F) and gel electrophoresis images of the amplified libraries are shown for the ThruPlex DNA seq kit (G) and TACS-TOPO ver. 4 (H). The 6% TBE-Urea at the bottom of the gel image indicates analysis on denaturing gel electrophoresis using 6% Novex TBE-Urea gel (Thermofisher Scientific), and 2% E-Gel Ex is indicative of analysis on E-Gel Ex (2%) (Thermofisher Scientific). The original gel images for A, C, G, and H are provided in Supplementary Figure S9.

Figure 6

Comparison of library preparation from ancient DNA using TACS-TOPO and a conventional protocol adapted only for dsDNA (KAPA Hyper Prep kit). (A–D) Comparisons of library yields after 12 cycles of PCR amplification (A), gel electrophoresis of amplified sequence libraries (B), mean alignment length of sequenced reads (C), and mean mapping rates of the reads on the human reference genome (D) are shown. The KAPA Hyper Prep kit was chosen as a representative protocol for its superior sensitivity among commercial library preparation kits and its tolerance for uracil-containing DNA. E and F. Mutation patterns were drawn using mapDamage^, for libraries prepared using the KAPA Hyper Prep kit (E) and TACS-TOPO ver. 5 (F). The four upper plots for each library show the base frequency outside and inside the read. The open grey box in the plots corresponds to the read. The bottom plots are the base substitution frequencies at relative positions from the 5′-(left) and 3′-(right) ends of the reads. The frequencies of C to T (red), G to A (blue), all other substitutions (gray), and soft-clipped bases (orange) are shown. Kapa: KAPA Hyper Prep kit; TT ver. 5: TACS-TOPO ver. 5. The original gel images for B are provided in Supplementary Figure S10.