Locating, tracing and sequencing multiple expanded genetic letters in complex DNA context via a bridge-base approach
通过桥基方法在复杂的 DNA 环境中定位、追踪和测序多个扩增的遗传字母

Abstract 抽象

A panel of unnatural base pairs is developed to expand genetic alphabets. One or more unnatural base pairs (UBPs) can be inserted to enlarge the capacity, diversity, and functionality of canonical DNA, so monitoring the multiple-UBPs-containing DNA by simple and convenient approaches is essential. Herein, we report a bridge-base approach to repurpose the capability of determining TPT3-NaM UBPs. The success of this approach depends on the design of isoTAT that can simultaneously pair with NaM and G as a bridge base, as well as the discovering of the transformation of NaM to A in absence of its complementary base. TPT3-NaM can be transferred to C–G or A–T by simple PCR assays with high read-through ratios and low sequence-dependent properties, permitting for the first time to dually locate the multiple sites of TPT3-NaM pairs. Then we show the unprecedented capacity of this approach to trace accurate changes and retention ratios of multiple TPT3-NaM UPBs during in vivo replications. In addition, the method can also be applied to identify multiple-site DNA lesions, transferring TPT3-NaM makers to different natural bases. Taken together, our work presents the first general and convenient approach capable of locating, tracing, and sequencing site- and number-unlimited TPT3-NaM pairs.
开发了一组非自然碱基对来扩展遗传字母表。可以插入一个或多个非天然碱基对 （UBP） 以扩大经典 DNA 的容量、多样性和功能，因此通过简单方便的方法监测含有多个 UBP 的 DNA 至关重要。在此，我们报告了一种基于桥的方法，以重新调整确定 TPT3-NaM UBP 的能力。这种方法的成功取决于 isoTAT 的设计，它可以同时与 NaM 和 G 配对作为桥接碱基，以及发现 NaM 在没有互补碱基的情况下向 A 的转化。TPT3-NaM 可以通过简单的 PCR 测定转移到 C-G 或 A-T，具有高读通率和低序列依赖性特性，首次允许对 TPT3-NaM 对的多个位点进行双重定位。然后，我们展示了这种方法在体内复制过程中追踪多个 TPT3-NaM UPB 的准确变化和保留率的前所未有的能力。此外，该方法还可用于鉴定多位点 DNA 损伤，将 TPT3-NaM 制造者转移到不同的天然碱基。综上所述，我们的工作提出了第一种能够定位、追踪和测序位点和数量不受限制的 TPT3-NaM 对的通用且方便的方法。

Graphical Abstract

We developed a bridge-base approach with high read-through ratios and low sequence-dependent properties for the first time to dually locate the multiple sites of TPT3-NaM pairs. This approach is also capable of tracing in vivo replications of TPT3-NaM pairs and sequencing multiple site DNA lesions marked with the unnatural base pairs.
我们首次开发了一种具有高读通率和低序列依赖性特性的桥基方法，以双重定位 TPT3-NaM 对的多个位点。这种方法还能够追踪 TPT3-NaM 对的体内复制，并对用非自然碱基对标记的多位点 DNA 损伤进行测序。

图形摘要

我们首次开发了一种具有高读通率和低序列依赖性特性的桥基方法，以双重定位 TPT3-NaM 对的多个位点。这种方法还能够追踪 TPT3-NaM 对的体内复制，并对用非自然碱基对标记的多位点 DNA 损伤进行测序。

Open in new tabDownload slide
在新选项卡中打开下载幻灯片

INTRODUCTION 介绍

Naturally, genetic information is stored in a four-letter A/T/G/C alphabet. Strenuous efforts have been made toward the design and synthesis of unnatural base pairs (UBPs) to expand the genetic letters (1–3), by now generating a panel of promising pairs including P-Z, Ds-Px and TPT3-NaM (4–8). Unique to the family of unnatural base pairs is their ability to act as bioorthogonal genetic letters for the replication and storage of genetic information (9,10). For example, the systematic evolution of ligands by exponential enrichment (SELEX) with DNA library bearing multiple UBPs is proven to effectively gain affinity-enhanced aptamers at pmol levels (11–13). Encoding three noncanonical amino acids by UBP-bearing codon and anticodon system (A-NaM-C as condon, G-TPT3-T as anticodon) has been succeeded in the expression of sfGFP proteins (14). More recently, an in-depth combination of UBPs and natural genetic letters has also been implemented for finding three new unnatural genetic codons, by which AGX, GXT and AXC (X = NaM) could be simultaneously and orthogonally decoded for protein expressions in semi-synthetic organisms (15). Insertion of more unnatural base pairs can enlarge the capacity, diversity, as well as functionality of canonical DNA (16–18), however, monitoring the multiple-UBPs-containing DNA by simple and convenient approaches is challenging. In addition, the in vivo efficiency of more than one UBP in DNA sequences shows more contingent during their replications in cells (14,15), which raises a new requirement to monitor the retention of multiple UBPs.
自然，遗传信息存储在四个字母的 A/T/G/C 字母表中。人们在设计和合成非天然碱基对 （UBP） 以扩展遗传字母方面付出了艰苦的努力 （1–3），现在生成了一组有前途的碱基对，包括 P-Z、Ds-Px 和 TPT3-NaM （4–8）。非天然碱基对家族的独特之处在于它们能够充当生物正交遗传字母，用于遗传信息的复制和存储 （9,10）。例如，使用带有多个 UBP 的 DNA 文库通过指数富集 （SELEX） 实现配体的系统进化已被证明可以有效地在 pmol 水平获得亲和力增强的适配子 （11–13）。通过携带 UBP 的密码子和反密码子系统（A-NaM-C 作为连接，G-TPT3-T 作为反密码子）编码三种非经典氨基酸，已在 sfGFP 蛋白的表达中取得成功 （14）。最近，还实施了 UBP 和自然遗传字母的深度组合，以寻找三种新的非自然遗传密码子，通过这些密码子，AGX、GXT 和 AXC （X = NaM） 可以同时正交解码，以在半合成生物体中表达蛋白质 （15）。插入更多非自然碱基对可以扩大经典 DNA 的容量、多样性和功能 （16–18），但是，通过简单方便的方法监测包含多个 UBP 的 DNA 具有挑战性。此外，DNA 序列中多个 UBP 的体内效率在细胞中复制过程中显示出更多的偶然性 （14,15），这提出了监测多个 UBP 保留的新要求。

TPT3-NaM is one of the advanced UPBs in the expanded genetic letters (2,19,20). Up to now, biotin shift and sequencing assays are developed to monitor DNA containing TPT3-NaM (2,21,22). However, the biotin shift assay is unable to detect multiple coding UBPs arranged tightly or UBPs with non-default loci and is also incompetent to monitor the site mutation near the UBPs (14,15), which may play an important role in precise protein expression. Sequencing assays can be divided into Sanger and nanopore sequencing. Sanger sequencing can detect a single UBP where the signal terminal can give the correct positions of unnatural bases (8,21). But the main obstacle to sequencing multiple UBPs in DNA is the disappearance of the Sanger sequencing signals after the first UBP site which makes the subsequent information hard to be acquired. Alternatively, the nanopore sequencing method has been developed to differentiate each unnatural base based on their different shapes (22,23), but the method involved tedious modifications on UBPs, specialized protein preparation procedures, and expansive apparatus, which possibly limits the general applications of this method. So the way for evaluation of complex DNA products with two or more TPT3-NaM UBPs at adjacent coding regions or with non-default positions, including calculation of their retentions, has remained undemonstrated thus far.
TPT3-NaM 是扩增遗传字母中的高级 UPB 之一 （2,19,20）。到目前为止，生物素迁移和测序检测已开发用于监测含有 TPT3-NaM 的 DNA （2,21,22）。然而，生物素迁移测定无法检测紧密排列的多个编码 UBP 或具有非默认基因座的 UBP，也无法监测 UBP 附近的位点突变 （14,15），这可能在精确蛋白表达中起重要作用。测序分析可分为 Sanger 测序和纳米孔测序。Sanger 测序可以检测到单个 UBP，其中信号末端可以给出非自然碱基的正确位置 （8,21）。但是，对 DNA 中的多个 UBP 进行测序的主要障碍是在第一个 UBP 位点之后 Sanger 测序信号的消失，这使得后续信息难以获取。或者，已经开发了纳米孔测序方法，以根据它们的不同形状来区分每个非天然碱基 （22,23），但该方法涉及对 UBP 的繁琐修改、专门的蛋白质制备程序和扩展设备，这可能限制了该方法的一般应用。因此，迄今为止，在相邻编码区或具有非默认位置具有两个或多个 TPT3-NaM UBP 的复杂 DNA 产物的评估方法（包括计算它们的保留）仍未得到证明。

Here, we design a bridge base (isoTAT) to transfer the TPT3-NaM pair to C–G natural pair exclusively (Figure 1A), which can be used to read through all the UBPs in DNA sequences. And we also demonstrate the transformation of NaM to A in the absence of its complementary base. Our results show that the efficient transformation of TPT3-NaM to C–G or A–T can be implemented by two PCR assays with high read-through ratios and low sequence-dependent properties, thus allowing for the first time to dually locate the multiple sites of TPT3-NaM pairs. Importantly, we further indicate the read-through capacity and transferring ability of the bridge base can be used to evaluate the retention rates of multiple TPT3-NaM UBPs at adjacent sites, and identify the accurate retention ratios of the adjacent unnatural bases in SSO replications. In addition, the approach is also practical to determine multiple DNA lesions marked by TPT3-NaM and then transform TPT3-NaM to different natural bases via PCR.
在这里，我们设计了一个桥基 （isoTAT） 以将 TPT3-NaM 对仅转移到 C-G 天然对（图 1A），可用于读取 DNA 序列中的所有 UBP。我们还证明了 NaM 在没有互补碱基的情况下向 A 的转化。我们的结果表明，TPT3-NaM 向 C-G 或 A-T 的高效转化可以通过两种具有高读穿率和低序列依赖性特性的 PCR 测定来实现，从而首次允许双重定位 TPT3-NaM 对的多个位点。重要的是，我们进一步表明桥基的通读能力和转移能力可用于评估相邻位点多个 TPT3-NaM UBP 的保留率，并确定 SSO 复制中相邻非自然碱基的准确保留率。此外，该方法也可用于确定 TPT3-NaM 标记的多个 DNA 损伤，然后通过 PCR 将 TPT3-NaM 转化为不同的天然碱基。

MATERIALS AND METHODS 材料和方法

Materials and analytic methods
材料和分析方法

Oligonucleotides with a length > 100 bp were purchased from GenScript, and other oligonucleotides were purchased from Sangon Biotech (see sequences used in this study). dNTPs were purchased from Solarbio. Klenow fragment DNA polymerase I was purchased from ABclonal Biotech Co., Ltd. OneTaq DNA polymerase, Deep vent DNA polymerase, glycosylase (UDG), apurinic/apyrimidinic Endonuclease 1 (APE1), as well as T4-DNA ligase, were purchased from New England Biolabs. 2 × UItraSYBR Mixture was purchased from CWBIO Biotech co., Ltd. pBLUE-T Fast Cloning Kits and BL 21 (DE3) Electrocompetent cells were purchased from Zoman Biotechnology Co., Ltd. dNaMTP, dTPT3TP, and dTPT3^PA were synthesized as reported (8,24). NMR spectra were performed on AVANCE NanoBay (400 MHz). HRMS or MS were performed on Bruker compact Ultra-high-resolution electro-spray time-of-flight mass spectrometry and Bruker Autoflex speed MALDLTOF/TOF spectrometry, respectively.
长度为 > 100 bp 的寡核苷酸购自金斯瑞，其他寡核苷酸购自桑贡生物科技（参见本研究中使用的序列）。dNTP 购自 Solarbio。Klenow 片段 DNA 聚合酶 I 购自 ABclonal Biotech Co.， Ltd.。OneTaq DNA 聚合酶、Deep vent DNA 聚合酶、糖基化酶 （UDG）、脱嘌呤/无嘧啶核酸内切酶 1 （APE1） 以及 T4-DNA 连接酶购自 New England Biolabs。2 × UItraSYBR 混合物购自 CWBIO Biotech co.， Ltd. pBLUE-T 快速克隆试剂盒，BL 21 （DE3） 电感受态细胞购自 Zoman Biotechnology Co.， Ltd. dNaMTP、dTPT3TP 和 dTPT3^PA，如报道所述 （8,24）。在 AVANCE NanoBay （400 MHz） 上进行 NMR 波谱分析。HRMS 或 MS 分别在 Bruker 紧凑型超高分辨率电喷雾飞行时间质谱仪和 Bruker Autoflex speed MALDLTOF/TOF 光谱仪上进行。

Pre-steady state kinetic assays
预稳态动力学分析

This assay was performed according to our previous report with some modifications (25). Briefly, after the annealing of primers (labeled with HEX) and templates, the incorporation reactions were initiated by adding 3 × dYTP (1.5 μM, 10 μl) to a 20 μl solution containing 113 nM primer/template and 4.5 U Klenow fragment DNA polymerase I. Reactions were set up at 37°C for 15 s, the reactions were quenched with 10 μl of 50 mM EDTA. For the extension assays, dCTP was added to the reaction system immediately to a concentration of 2 μM after the incorporation reaction, then incubated for another 15, 30 or 60 s, and finally quenched with 10 μl of 50 mM EDTA. Only isoTAT was also extended with NaMG and GG templates with different concentrations and time-coursing. All solutions were in Kf buffer (10 mM Tris–HCl, 50 mM NaCl, 10 mM MgCl₂, 1 mM DTT, pH 7.9 at 25°C). The reaction solution was concentrated, mixed with enough loading dye (90% formamide, and sufficient amounts of bromophenol blue and xylene cyanol), and analyzed by 8 M urea 15% denaturing polyacrylamide gel electrophoresis. The primer strips were imaged and quantified by AI600 software (Amersham Imager 680). The data are averages and standard deviations of three independent determinations. GraphPad Prism 8 was used to plot the data.
该测定是根据我们之前的报告进行的，并进行了一些修改 （25）。简而言之，在引物（用 HEX 标记）和模板退火后，通过将 3 × dYTP（1.5 μM，10 μl）添加到含有 113 nM 引物/模板和 4.5 U Klenow 片段 DNA 聚合酶 I 的 20 μl 溶液中来引发掺入反应。在 37°C 下构建反应 15 秒，用 10 μl 50 mM EDTA 淬灭反应。对于延伸测定，在掺入反应后立即将 dCTP 加入反应系统中至浓度为 2 μM，然后再孵育 15、30 或 60 秒，最后用 10 μl 50 mM EDTA 淬灭。只有 isoTAT 还用不同浓度和时间运行的 NaMG 和 GG 模板进行了扩展。所有溶液均溶于 Kf 缓冲液（10 mM Tris-HCl、50 mM NaCl、10 mM MgCl₂、1 mM DTT，pH 7.9,25°C 时）。浓缩反应液，与足够的负载染料（90% 甲酰胺和足量的溴酚蓝和二甲苯蓝）混合，并通过 8 M 尿素 15% 变性聚丙烯酰胺凝胶电泳进行分析。引物条带通过 AI600 软件 （Amersham Imager 680） 成像和定量。数据是三个独立测定的平均值和标准差。GraphPad Prism 8 用于绘制数据。

Steady-state kinetic assays
稳态动力学检测

The steady-state kinetic parameters were monitored according to the previous report with some changes (24,26). The reaction system and operation were the same as the incorporation assays above with changes in the total amount of Kf enzyme (0.225 U), the concentrations of dYTP or dNTP (the final concentrations ranging from 0.0039–100 μM), and the incubation time changed to 10 s. The steady-state kinetic constants for each base pair were calculated from eight concentrations vs velocities using Michaelis − Menten equation by GraphPad Prism 8. Velocities were calculated by the fractions of extended primer/min (%incor/min) with all the fractions of extended primer <20% at each concentration. The data are the averages and standard deviations of three independent experiments.
根据之前的报告监测稳态动力学参数，并有一些变化 （24,26）。反应系统和操作与上述掺入分析相同，Kf 酶总量 （0.225 U）、dYTP 或 dNTP 浓度（最终浓度范围为 0.0039–100 μM）发生变化，孵育时间变为 10 s。每个碱基对的稳态动力学常数是使用 GraphPad Prism 8 的 Michaelis − Menten 方程从八个浓度与速度计算得出的。通过延伸引物/分钟 （%incor/min） 的馏分与每个浓度下延伸引物的所有馏分 <20%） 计算速度。数据是三个独立实验的平均值和标准差。

Transformed PCR and retention assays
转化的 PCR 和保留检测

Quantitative real time PCR (qPCR) analysis with 2 × UItraSYBR was based on the Manufacturer's method with the addition of isoTAT and NaM or only NaM to a concentration of 0.1 mM following the thermal cycling conditions: 95°C, 10 min; (95°C, 15 s; 60°C, 1.5 min) ×40. OneTaq DNA polymerase was used to perform all other PCR assays. First, PCR was performed to obtain doubled strand templates under the following conditions (in a volume of 25 μl): 1× reaction buffer with OneTaq DNA Polymerases, 0.2 mM dNTP, 0.1 mM dNaMTP and dTPT3TP, 0.4 ng of the single strand 134-mer templates (nature, 1N, 2N, 3N, or UN), and 400 nM primers under the following thermal cycling conditions: 16× (96°C, 10 s; 60°C, 15 s; 68°C, 2 min for 3N template or 68°C, 1 min for other templates), and then a final extension (68°C, 5 min). The PCR products were purified with a QIAquick Gel Extraction Kit (QIAGEN) and quantified by NanoDrop™ One^C. Second, bridge base PCR with isoTAT and NaM as well as inherent base's preference PCR with only NaM were performed with 1 ng of the double strand templates (1N, 2N, 3N or UN) and amplified for 36 cycles with all other conditions same to above, besides, the inherent base's preference PCR with only TPT3 were only performed with UN template. Cultured Escherichia coil cells with replicated plasmids containing one or three NaM-TPT3 pairs (2 μl) were also used as templates for the above transferred PCR or unnatural base PCR with NaM and TPT3 with changes in the addition of pre-denaturation at 96°C for 3 min but all other conditions same to transferred PCR. The products of the unnatural base, bridge base, and inherent bases PCR were sequenced by Sangon Biotech.
使用 2 × UItraSYBR 进行定量实时 PCR （qPCR） 分析基于制造商的方法，在热循环条件下添加 isoTAT 和 NaM 或仅添加 NaM 至 0.1 mM 浓度：95°C，10 分钟;（95°C，15 秒;60°C，1.5 分钟）×40.OneTaq DNA 聚合酶用于进行所有其他 PCR 分析。首先，在以下条件下进行 PCR 以获得双链模板（体积为 25 μl）：在以下热循环条件下，使用 OneTaq DNA 聚合酶的 1× 反应缓冲液、0.2 mM dNTP、0.1 mM dNaMTP 和 dTPT3TP、0.4 ng 单链 134 聚体模板（nature、1N、2N、3N 或 UN）和 400 nM 引物： 16×（96°C，10 s;60°C，15 s;68°C，3N 模板为 2 分钟，其他模板为 68°C，1 分钟），然后是最终延伸（68°C，5 分钟）。使用 QIAquick 凝胶提取试剂盒 （QIAGEN） 纯化 PCR 产物，并使用 NanoDrop™ One^C 定量。其次，使用 1 ng 双链模板（1N、2N、3N 或 UN）进行具有 isoTAT 和 NaM 的桥式碱基 PCR 以及仅具有 NaM 的固有碱基偏好 PCR，并在所有其他条件相同的情况下扩增 36 个循环，此外，仅具有 TPT3 的固有碱基偏好 PCR 仅使用 UN 模板进行。培养的埃希菌螺旋细胞与含有 1 对或 3 对 NaM-TPT3 （2 μl） 的复制质粒也用作上述转移 PCR 或具有 NaM 和 TPT3 的非自然碱基 PCR 的模板，在 96°C 下添加预变性 3 分钟，但所有其他条件与转移 PCR 相同。非天然碱基、桥基和固有碱基 PCR 的产物由 Sangon Biotech 进行测序。

The percent retention of an unnatural base pair (F) was calculated using the raw sequencing data as reported (8,21), it was to use the lack of fluorescent channels for matching UBPs resulting in the signal terminal at the accurate position of UBPs, while the mutations to natural nucleotides resulted in readthrough. The average signal intensities of each channel (A, C, G and T) for defined points (35th–45th positions in the sequence) before (L) and after (R) the unnatural base were determined. The normalized R/l ratio was considered as the percentage of the natural sequences, and retention was calculated as 1 - normalized R/l ratio. So we used a natural template that contained nucleotides that were different from the bridge transformation in the corresponding positions and with the nucleotides behind the unnatural base were all the same, except template 1N with several nucleotides surrounding different from the natural template but did not influence, to determine the R/l′ ratio. R was the average signal intensity of C behind the unnatural base, L′ was the single transferred signal C of bridge base PCR product mixed with natural template PCR products in different fractions. PCR procedure with a natural template was the same as that of bridge base PCR without the addition of unnatural bases. The products of 1N or 3N template by transferred PCR and of natural template were purified with a QIAquick Gel Extraction Kit (QIAGEN) and quantified by NanoDrop™ One^C. The mixtures containing 100%, 90%, 80%, 70%, 60%, 50%, 40%, 30%, 20% and 0 of bridge base transferred products were prepared and sequenced. The R/l′ ratio over the percentage of bridge base transferred products was plotted and fitted by linear regression (Supplementary Figure S19). Representative sequencing data for 1N and 3N mixtures were shown in Supplementary Figures S17 and S18. The data are averages and standard deviations of three independent determinations.
使用报告的原始测序数据 （8,21） 计算非天然碱基对 （F） 的保留百分比，利用缺乏荧光通道来匹配 UBP，导致信号末端位于 UBP 的准确位置，而天然核苷酸的突变导致可读。确定了非自然碱基之前 （L） 和 （R） 之前 （R） 定义点（序列中的第 35-45 个位置）每个通道 （A、C、G 和 T） 的平均信号强度。归一化 R/l 比值被认为是自然序列的百分比，保留率计算为 1 - 归一化 R/l 比值。所以我们用了一个天然模板，它含有在相应位置与桥式转化不同的核苷酸，并且非天然碱基后面的核苷酸都是相同的，除了模板 1N 周围有几个核苷酸与天然模板不同但没有影响，来确定 R/l′ 比值。R 是非天然碱基后面 C 的平均信号强度，L′ 是桥式碱基 PCR 产物与不同组分的天然模板 PCR 产物混合的单一传递信号 C。使用天然模板的 PCR 程序与桥式 PCR 相同，未添加非天然碱基。使用 QIAquick 凝胶提取试剂盒 （QIAGEN） 纯化通过转印 PCR 制备的 1N 或 3N 模板产物和天然模板产物，并使用 NanoDrop™ One^C 定量。制备含有 100%、90%、80%、70%、60%、50%、40%、30%、20% 和 0 的桥基转移产物的混合物并排序。 绘制 R/l′ 比率与桥基转移产物百分比的关系图，并通过线性回归拟合（补充图 S19）。1N 和 3N 混合物的代表性测序数据显示在补充图 S17 和 S18 中。数据是三个独立测定的平均值和标准差。

Deep sequencing assay 深度测序检测

Deep sequencing assays were performed by Sangon Biotech., Briefly, the products obtained from bridge base PCR with isoTAT and NaM or inherent base's preference PCR with only NaM were purified with a QIAquick Gel Extraction Kit (QIAGEN) according to the manufacturer's methods. Then compatible primers for Illumina bridge PCR were ligated to our DNAs using 2× Hieff® Robust PCR Master Mix (Yeasen, 10105ES03), and the samples were enriched by Hieff NGS™ DNA Selection Beads (Yeasen, 12601ES56) and sequenced with a Miseq Reagent Kit 3. The raw data were also supplied in supporting material.
简而言之，根据制造商的方法，使用 QIAquick 凝胶提取试剂盒 （QIAGEN） 纯化使用 isoTAT 和 NaM 的桥式碱基 PCR 或仅使用 NaM 的固有碱基偏好 PCR 获得的产物。然后使用 2× Hieff® Robust PCR Master Mix （Yeasen， 10105ES03） 将 Illumina 桥 PCR 的兼容引物连接到我们的 DNA 上，并使用 Hieff NGS™ DNA 选择珠 （Yeasen， 12601ES56） 富集样品，并使用 Miseq 试剂盒 3 进行测序。原始数据也以支持材料的形式提供。

Growth and in vivo replication assay
生长和体内复制测定

This was performed according to previous reports (2,25). Briefly, the plasmid containing PtNTT2 reported in the literature was synthesized by GenScript (27). Sequences containing one or three NaM-TPT3 pairs were shown in sequences used in this study. Then they were amplified and assembled into pBLUE-T plasmids. Finally, the two plasmids were transformed into BL 21 (DE3) electrocompetent cells. And cells were cultured with 2 × YT media (200 μl, casein peptone 16 g/l, yeast extract 10 g/l, NaCl 5 g/l), 5 μg/ml chloramphenicol, 100 μg/ml ampicillin, 50 mM KPi) with unnatural bases dNaMTP and dTPT3TP at the concentration of 125 μM. The growth of bacteria (OD₆₀₀) was monitored by NanoDrop™ One^C with a cuvette of 2 mm optical path to reach an OD₆₀₀ range at 0.6–2 and then used for PCR assay. The data are averages and standard deviations of three independent determinations.
这是根据以前的报告 （2,25） 进行的。简而言之，文献中报道的含有 PtNTT2 的质粒是通过金斯瑞合成的 （27）。包含一个或三个 NaM-TPT3 对的序列显示在本研究中使用的序列中。然后将它们扩增并组装成 pBLUE-T 质粒。最后，将两个质粒转化到 BL 21 （DE3） 电感受态细胞中。用 2 × YT 培养基 （200 μl、酪蛋白蛋白胨 16 g/l、酵母提取物 10 g/l、NaCl 5 g/l）、5 μg/ml 氯霉素、100 μg/ml 氨苄青霉素、50 mM KPi）和浓度为 125 μM 的非天然碱 dNaMTP 和 dTPT3TP 培养细胞。使用 NanoDrop™ One^C 和 2 mm 光路的比色皿监测细菌 （OD₆₀₀） 的生长，以达到 0.6-2 的 OD₆₀₀ 范围，然后用于 PCR 测定。数据是三个独立测定的平均值和标准差。

Labeling dU lesion in DNA with unnatural nucleotide
用非自然核苷酸标记 DNA 中的 dU 损伤

KRAS-1U-F or 2U-F and KRAS-R or 134–2U-F and 134–2U-R (0.2 μM in 50 μl) were annealed to form double-stranded DNA in ABclonal buffer B. UDG (1 U) were added to the reaction mixture and incubated at 37°C for 30 min. Then APE1 (10 U) was added to the reaction mixture, incubated at 37°C for 1 h, and heated at 95°C for 10 min. Next, dTPT3TP^biotin (30 μM) and Kf (exo⁻) DNA polymerase (7 U) were added to the reaction mixture to react for 1 h at 37°C, and heated to 95°C to terminate the reaction. Finally, T4-DNA ligase (200 U), dimethylsulphoxide (10% (v/v)), and ATP (0.2 mM) were added to the reaction system and incubated at 25°C for 1 h. The reaction steps above were monitored by 20% denaturing PAGE gel.
将 KRAS-1U-F 或 2U-F 和 KRAS-R 或 134–2U-F 和 134–2U-R（0.2 μM 在 50 μl 中）在 ABclonal 缓冲液 B 中形成双链 DNA。将 UDG （1 U） 加入反应混合物中，并在 37°C 下孵育 30 分钟。然后将 APE1 （10 U） 加入反应混合物中，在 37°C 下孵育 1 小时，并在 95°C 下加热 10 分钟。接下来，将 dTPT3TP^生物素 （30 μM） 和 Kf （exo^-） DNA 聚合酶 （7 U） 添加到反应混合物中，在 37°C 下反应 1 小时，并加热至 95°C 以终止反应。最后，将 T4-DNA 连接酶 （200 U）、二甲基亚砜 （10% （v/v）） 和 ATP （0.2 mM） 加入反应体系中，并在 25°C 下孵育 1 小时。上述反应步骤通过 20% 变性 PAGE 凝胶监测。

Isolation of labeled DNA and bridge PCR
分离标记的 DNA 和桥式 PCR

The final reaction solution above (KRAS-1U or 134–2U) was used as the template for further PCR assays. First, the labeled DNA was further labeled and amplified via PCR according to the method reported(2,27), with some modifications: templates (the final reaction solution, 1 μl), dNTPs (400 μM), dTPT3TP^biotin (20 μM), dNaMTP (50 μM), MgSO₄ (2.2 mM), primers (400 nM each) OneTaq DNA Polymerases (0.018 U/μl), and DeepVent DNA polymerase (0.007 U/μl) in 1 × reaction buffer (a total of 25 μl), under the thermocycling conditions: 20 × (96°C, 30 s; 50°C, 10 s; 68°C, 4 min) with a final extension for 68°C, 5 min. The products (5 μl) were incubated with streptavidin (1 μg, Solarbio) for 30 min at 37°C. Then samples were mixed with loading dye and separated by a 6% (134-2U) or 10% (KRAS-1U) non-denaturant polyacrylamide gel electrophoresis. The shift strips were eluted by shaking and soaking at 37°C for 2 h, quantified by NanoDrop™ One^C, and used as templates for bridge base PCR (0.5–2 ng per sample). Bridge PCR was performed as described above, and inherent dU’s preference PCR as control was also performed with the same procedure as that of bridge PCR.
上述最终反应溶液（KRAS-1U 或 134-2U）用作进一步 PCR 分析的模板。首先，根据报道的方法进一步标记标记的 DNA 并通过 PCR 扩增 （2,27），并进行一些修改：模板（最终反应溶液，1 μl）、dNTP （400 μM）、dTPT3TP^生物素 （20 μM）、dNaMTP （50 μM）、MgSO₄ （2.2 mM）、引物（各 400 nM）、OneTaq DNA 聚合酶 （0.018 U/μl） 和 DeepVent DNA 聚合酶 （0.007 U/μl） 在 1 × 反应缓冲液（总共 25 μl）中， 在热循环条件下：20 ×（96°C，30 s;50°C，10 s;68°C，4 分钟），最后延伸 68°C，5 分钟。将产物 （5 μl） 与链霉亲和素（1 μg，Solarbio）在 37°C 下孵育 30 分钟。 然后将样品与上样染料混合，并通过 6% （134-2U） 或 10% （KRAS-1U） 非变性聚丙烯酰胺凝胶电泳分离。摇动并在 37°C 下浸泡 2 小时洗脱移位条，用 NanoDrop™ One^C 定量，并用作桥基 PCR 的模板（每个样品 0.5–2 ng）。如上所述进行桥式 PCR，固有的 dU 偏好 PCR 作为对照也使用与桥式 PCR 相同的程序进行。

Cell growth and preparation of plasmid DNA
细胞生长和质粒 DNA 的制备

Damages of Apurinic and apyrimidinic (AP) sites were induced according to the previous report with some changes (28). Briefly, pUC-19 plasmids were transformed into E. coli DH5α cells, a single colony was grown in 2 ml LB medium overnight at 37°C, and then 10 μl culture was diluted with fresh LB medium and incubated at 37°C, 230 rpm until OD₆₀₀ = 1 for H₂O₂ treatment. H₂O₂ was added to the culture at a final concentration of 1 mM to generate AP sites. After sitting at ambient temperature for 30 min, cells were harvested by centrifuge at ambient temperature and plasmid DNA was extracted by the plasmid isolation kit following the manufacturer's protocol (OMEGA). Plasmids from cells without H₂O₂ treatment were used as a control.
根据之前的报告，诱导了 Apurinic 和 apyrimidinic （AP） 位点的损伤，并发生了一些变化 （28）。简而言之，将 pUC-19 质粒转化到大肠杆菌 DH5α 细胞中，在 37°C 下在 2 ml LB 培养基中培养单个菌落过夜，然后用新鲜的 LB 培养基稀释 10 μl 培养物并在 37°C、230 rpm 下孵育直至 OD₆₀₀ = 1 用于 H₂O₂ 处理。将 H₂O₂ 以终浓度为 1 mM 的浓度添加到培养物中以生成 AP 位点。在环境温度下放置 30 分钟后，在环境温度下通过离心机收获细胞，并按照制造商的方案 （OMEGA） 通过质粒分离试剂盒提取质粒 DNA。来自未经 H₂O₂ 处理的细胞的质粒用作对照。

Data analysis 数据分析

Deep sequencing data for AP sites were analyzed by mutation levels of the nucleotides in each site. As each of the four natural bases might be damaged to AP sites, and then transformed into different natural bases after NaM-TPT3 incorporation by replacement PCR. If one nucleotide was damaged and an AP site formed, then TPT3 would be incorporated into the sites, and finally, C or A at the site would be detected by the replacement PCR with bridge base or only NaM respectively. For G and T, the mutation levels of A and C could be calculated. For A and C, only the mutation levels of C or A could be calculated, as they were one of the nucleotides being transformed. The mutation ratios of controlled PCR were used as input to eliminate the background.
通过每个位点核苷酸的突变水平分析 AP 位点的深度测序数据。由于 4 个天然碱基中的每一个都可能被 AP 位点破坏，然后在 NaM-TPT3 掺入后通过替代 PCR 转化为不同的天然碱基。如果一个核苷酸受损并形成一个 AP 位点，则 TPT3 将被掺入这些位点，最后，该位点的 C 或 A 将分别通过桥基或仅 NaM 的替换 PCR 检测到。对于 G 和 T，可以计算 A 和 C 的突变水平。对于 A 和 C，只能计算 C 或 A 的突变水平，因为它们是被转化的核苷酸之一。使用对照 PCR 的突变比率作为输入以消除背景。

Binding analyses 结合分析

The binding affinities of aptamers with TPT3 or natural bases to IFN-γ were measured by SPR (Reichert2, Reichert). Briefly, 5’-biotin labeled aptamers were diluted to 50 nM in 1× PBS, denatured at 90°C for 1 min, cooled down slowly to 25°C, placed on ice for more than 5 min, supplemented with Nonidet P-40 at a 0.005% (wt/vol) final concentration, and immobilized on a Sensor chip SA by injecting the DNA solution at a flow rate of 10 μl min⁻¹ in 1× binding buffer (immobilization level: ∼200 response units). Concentration series of IFN-γ were injected at a flow rate of 25 μl min⁻¹ for 180 s, and the dissociation was monitored for 300 s in 1 × running buffer (1 × PBS with 0.05% wt/vol Nonidet P-40 and 50 mM NaCl). After each injection, the sensor surface was regenerated with 25-μl injections of 5 mM NaOH for 60 s, and the following refolding of DNA was accomplished by 1 × running buffer for 10 min. The control sensorgrams of a reference cell lacking immobilized DNA fragments on the sensor surface with the buffer injection were subtracted from each sensorgram of the aptamers from a measuring cell, and the data were fitted with a 1:1 binding model.
通过 SPR （Reichert2， Reichert） 测量适配体与 TPT3 或天然碱基对 IFN-γ 的结合亲和力。简而言之，将 5'-生物素标记的适配体在 1× PBS 中稀释至 50 nM，在 90°C 下变性 1 分钟，缓慢冷却至 25°C，置于冰上 5 分钟以上，补充终浓度为 0.005% （wt/vol） 的 Nonidet P-40，并通过在 1× 结合缓冲液中以 10 μl ^min-1 的流速注射 DNA 溶液固定在传感器芯片 SA 上（固定水平： ∼200 个响应单位）。以 25 μl ^min-1 的流速注射 IFN-γ 浓度系列 180 秒，并在 1 × 运行缓冲液（1 × PBS，含 0.05% wt/vol Nonidet P-40 和 50 mM NaCl）中监测解离 300 秒。每次注射后，用 25 μl 5 mM NaOH 注射 60 秒再生传感器表面，然后通过 1 × 运行缓冲液 10 分钟完成 DNA 的后续重新折叠。从测量单元中适配体的每个传感图中减去缓冲液注射在传感器表面缺乏固定化 DNA 片段的参考细胞的对照传感图，并使用 1：1 结合模型拟合数据。

RESULTS 结果

Design of the bridge base isoTAT for pairing with both NaM and G.
桥基 isoTAT 的设计，用于与 NaM 和 G 配对。

To develop a bridge base for TPT3-NaM UBP, we need to design a structural skeleton that can efficiently pair with one of TPT3/NaM bases, and simultaneously effectively pair with one of the natural A/T/G/C bases. TPT3 represents an optimal base that has been proven to pair well with a panel of NaM-type unnatural bases and shows tolerance for the structural changes to keep its pairing capacity (8,20,25). Based on the properties, we reason that a moderate modification of TPT3 base may generate a new base that will keep efficiently pairing with the NaM base in the successful recognition model of TPT3-NaM pair, and simultaneously can enhance its pairing capacity with one of the natural bases. Furthermore, we expect that the incorporation of a hydrogen-bonding donor/acceptor will be a choice to alter the pairing capacity of TPT3 with natural bases, especially the addition of a hydrogen-bonding acceptor on the inner side of thienyl motif on TPT3, which is expected to form an effective hydrogen bond with N-1 of G base (G/isoTAT in Figure 1A). Finally, we suspect that the aza-modification analogue of TPT3, isoTAT, will not pair with the other three natural bases of A/T/C due to either steric hindrance or hydrogen bond repelling (Figure 1B). Thus isoTAT can be used as a bridge-base candidate.
为了开发 TPT3-NaM UBP 的桥基，我们需要设计一个结构骨架，该骨架可以与其中一个 TPT3/NaM 碱基有效配对，同时与其中一个天然 A/T/G/C 碱基有效配对。TPT3 代表一种最佳碱基，已被证明可与一组 NaM 型非天然碱基完美配对，并显示出对结构变化的耐受性，以保持其配对能力 （8,20,25）。基于这些特性，我们推断 TPT3 碱基的适度修饰可能会产生一个新的碱基，该碱基将在 TPT3-NaM 对的成功识别模型中保持与 NaM 碱基的有效配对，同时可以增强其与其中一个天然碱基的配对能力。此外，我们预计掺入氢键供体/受体将成为改变 TPT3 与天然碱基配对能力的一种选择，尤其是在 TPT3 上噻吩基序内侧添加氢键受体，预计这将与 G 碱基的 N-1 形成有效的氢键（图 1A 中的 G/isoTAT).最后，我们怀疑 TPT3 的 aza 修饰类似物 isoTAT 由于空间位阻或氢键排斥而不会与 A/T/C 的其他三种天然碱基配对（图 1B）。因此，isoTAT 可以用作 bridge-base 候选者。

To investigate the pairing capacity of isoTAT, we used presteady-state assays with 45-mer template and 23-mer primer (Figure 2A). Gratifyingly, co-incubation of the templates, primers, and disoTATs with Klenow fragment of E. coli DNA polymerase I (Kf (exo-)) for 15 s, resulted in efficient incorporations of isoTAT opposite NaM in the template with a ∼90% yield and opposite G in the template with a ∼94% yield (Figure 2B and C). Furthermore, we measured the pairing ability of isoTAT with A, T and C bases with the same assays, and less than 10% incorporation yields could be obtained (Figure 2B and C). These data indicate a high selectivity of isoTAT base to recognize both NaM and G, while much less effective to recognize A, T and C.
为了研究 isoTAT 的配对能力，我们使用了 45 聚体模板和 23 聚体引物的预稳态测定（图 2A）。令人欣慰的是，将模板、引物和 disoTAT 与大肠杆菌 DNA 聚合酶 I （Kf （exo-）） 的 Klenow 片段共孵育 15 秒，导致模板中 NaM 与 NaM 相反的 isoTAT 以 ∼90% 的产量有效掺入，在模板中以 ∼94% 的产量将 G 相反的 G（图 2B 和 C).此外，我们用相同的测定法测量了 isoTAT 与 A、T 和 C 碱基的配对能力，并且可以获得不到 10% 的掺入产量（图 2B 和 C）。这些数据表明 isoTAT 碱基识别 NaM 和 G 的选择性很高，而识别 A、T 和 C 的效率要低得多。

Steady-state kinetic assays were further performed to characterize the efficiencies ((second-order rate constant, K_m/V_max) (Figure 2A). The insertion efficiencies of disoTAT opposite dNaM or dG were 7.68 × 10⁷ and 2.14 × 10⁸ respectively, which was comparable to that of NaM-TPT3 at 3.36 × 10⁹. Moreover, no detectable reactions could be measured for A-isoTAT, T-isoTAT and C-isoTAT under the same conditions (Table 1 and Supplementary Figure S1). Usually, over two orders of magnitude for K_m/V_max values would lead to the exclusive pairing capacity for UBPs (29). The remarkable insertion efficiencies of disoTAT opposite dNaM and dG provided another evidence that isoTAT could pair both certain unnatural base NaM and natural base G. Because a high strand extension rate of UBPs was key for their replications to give full-length products, the extension abilities of the primer after isoTAT incorporated were also investigated. The incorporated products of NaM-isoTAT and G-isoTAT respectively with the next complementary dCTP were incubated, time-coursing extension reactions were detected within 60 s to give ∼80% yields of extended products, which were just slightly lower than the yield of NaM-TPT3 UBP itself (Figure 2D and E). In addition, we also found that isoTAT could extend to the next G with concentration-response and time-coursing after incorporating opposite NaM or the first G (Supplementary Figure S2). With the NaMG template, the extension yield of isoTAT: G (25 nt) arrived at 92.5% with 8 μM disoTATTP after 5 min; alternatively, the extension yield of isoTAT: G (25 nt) could also arrive at 82.6% with 2 μM disoTAT-TP after 10 min. With the GG template, the extension yield of isoTAT: G (25 nt) arrived at 95.3% with 8 μM disoTATTP after 5 min. These suggest that isoTATs can be consecutively inserted in the DNA. Taken together, these kinetic data indicate isoTAT may be used as a bridge base.
进一步进行稳态动力学测定以表征效率（（二阶速率常数，K_m/V_max）（图 2A）。与 dNaM 或 dG 相反的 disoTAT 的插入效率分别为 7.68 × 10⁷ 和 2.14 × 10⁸，与 NaM-TPT3 在 3.36 × 10 9 的插入效率^相当。此外，在相同条件下，无法测量 A-isoTAT、T-isoTAT 和 C-isoTAT 的可检测反应（表 1 和补充图 S1）。通常，K_m/V_最大值超过两个数量级将导致 UBP 的独占配对容量 （29）。与 dNaM 和 dG 相反的 disoTAT 的显着插入效率提供了另一个证据，证明 isoTAT 可以配对某些非天然碱基 NaM 和天然碱基 G。由于 UBP 的高链延伸率是其复制产生全长产物的关键，因此还研究了 isoTAT 掺入后引物的延伸能力。将 NaM-isoTAT 和 G-isoTAT 的掺入产物分别与下一个互补的 dCTP 一起孵育，在 60 秒内检测时间流动延伸反应，得到 ∼80% 的延伸产物产量，略低于 NaM-TPT3 UBP 本身的产量（图 2D 和 E).此外，我们还发现 isoTAT 在掺入相反的 NaM 或第一个 G 后可以扩展到具有浓度响应和时间流动的下一个 G（补充图 S2）。 使用 NaMG 模板，isoTAT 的延伸产量：G （25 nt） 在 5 分钟后达到 92.5%，8 μM disoTATTP;或者，isoTAT 的延伸产量：G （25 nt） 也可以在 10 分钟后用 2 μM disoTAT-TP 达到 82.6%。使用 GG 模板，isoTAT：G （25 nt） 的延伸产量在 5 分钟后达到 95.3%，8 μM disoTATTP。这些表明 isoTAT 可以连续插入 DNA 中。综上所述，这些动力学数据表明 isoTAT 可用作桥基。

Dual location of multiple NaM-TPT3 pairs.
多个 NaM-TPT3 对的双重定位。

Twice replacement PCR strategy has been used by Benner's team and Hirao's team for sequencing their P-Z and Ds-Px UBPs (7,30). This means that unnatural base pairs can be converted into different natural base pairs with two varied PCR conditions. To this end, we expected the presence of disoTATTP in the PCR buffer would be the optimal choice to induce the transformation of TPT3-NaM to natural C–G pairs during the amplification. On the other hand, the kinetic data have suggested an A-base's preference of NaM in some sequence contexts (24,26), and our kinetic data also showed that NaM could pair with A but not the other natural nucleotides (Supplementary Figure S3 and Table 1). We therefore designed PCR assays using disoTATTP/dNaMTP or only dNaMTP with natural dNTPs and expected the potential of using disoTATTP and the NaM own bases’ preferences to monitor multiple TPT3-NaM pairs (Figure 3A). Double-stranded DNA templates 1N, 2N, 3N and UN containing one, two, three, or one NaM-TPT3 (with three random nucleotides upstream and downstream), were chosen respectively. As shown in Figure 3B and C and Supplementary Figures S4 and S5, the results indicated that the NaM-TPT3 pairs surrounded by specific sequences or random nucleotides were all mainly converted to G–C with disoTATTP/dNaMTP and natural dNTPs. And it was also confirmed that the NaM-TPT3 pairs were all mainly converted to T–A with only dNaMTP and natural dNTPs. The results demonstrated that the combination of our bridge base and the inherent base's preference of NaM, through twice PCR assays and the followed Sanger-sequencing signals’ comparison, could offer an effective and concise method for determinations of multiple TPT3-NaM UBPs with known or unknown loci. Moreover, the effects of quantity of templates were also investigated via qPCR amplification using the 1N template under the two PCR conditions. After 40 cycles, we found that the amplification of 1N remained efficient and accurate even with only 0.004 pg, and PCR with only NaM showed higher amplification efficiency (Figures 3D and S6).
Benner 的团队和 Hirao 的团队已使用两次替代 PCR 策略对他们的 P-Z 和 Ds-Px UBP 进行测序 （7,30）。这意味着非天然碱基对可以在两种不同的 PCR 条件下转化为不同的天然碱基对。为此，我们预计 PCR 缓冲液中存在 disoTATTP 将是扩增过程中诱导 TPT3-NaM 转化为天然 C-G 对的最佳选择。另一方面，动力学数据表明，在某些序列环境中，A 碱基更喜欢 NaM （24,26），我们的动力学数据还表明 NaM 可以与 A 配对，但不能与其他天然核苷酸配对（补充图 S3 和表 1).因此，我们设计了使用 disoTATTP/dNaMTP 或仅使用具有天然 dNTP 的 dNaMTP 的 PCR 测定，并预期使用 disoTATTP 和 NaM 自身碱基偏好来监测多个 TPT3-NaM 对的潜力（图 3A）。选择双链 DNA 模板 1N、2N、3N 和 UN，分别包含一个、两个、三个或一个 NaM-TPT3 （上游和下游具有三个随机核苷酸）。如图 3B 和 C 以及补充图 S4 和 S5 所示，结果表明，被特定序列或随机核苷酸包围的 NaM-TPT3 对都主要通过 disoTATTP/dNaMTP 和天然 dNTP 转化为 G-C。并且还证实，NaM-TPT3 对都主要转化为 T-A，只有 dNaMTP 和天然 dNTP。 结果表明，通过两次 PCR 测定和随后的 Sanger 测序信号比较，将我们的桥式碱基与固有碱基对 NaM 的偏好相结合，可以为测定具有已知或未知基因座的多个 TPT3-NaM UBP 提供一种有效且简洁的方法。此外，在两种 PCR 条件下，使用 1N 模板通过 qPCR 扩增研究了模板数量的影响。40 个循环后，我们发现即使只有 0.004 pg，1N 的扩增仍然高效和准确，并且仅使用 NaM 的 PCR 显示出更高的扩增效率（图 3D 和 S6）。

To further evaluate the practicality of this method, PCR amplification efficiency with 1N template was analyzed using a gel-based assay by band brightness ratio (gray ratio). The result showed their considerable replication efficiency of dTPT3TP/dNaMTP, only dNaM or disoTATTP/dNaMTP with a gray ratio of 1.0:1.2:0.8 (Supplementary Figure S7). Moreover, raw sequencing signals of PCR products were also analyzed. They showed no cut-down in the presence of disoTATTP/dNaMTP or only NaM (Supplementary Figure S8), which also proved that the unnatural bases were converted to natural bases. Furthermore, we analyzed the signals of all other natural bases before and behind the UBP, aiming to find possible disadvantageous effects of the bridge base isoTAT on the replications of natural base pairs. Gratifyingly, no detectable mutations of natural bases in all sequencing signals except the UBP site (Supplementary Figure S8). These might be probably due to the higher pairing efficiency of natural base pairs, NaM-isoTAT and isoTAT-G (K_m/V_max values of both A–T and C–G are over 10⁹) as well as the undetectable pairing efficiency between isoTAT and other natural bases. Finally, dTPT3TP and natural dNTPs were used for PCR assays to explore whether TPT3 had its inherent preference in the absence of NaM, only messy signals were obtained at the predetermined position of TPT3-NaM UBP in template UN (Supplementary Figure S9), suggesting that TPT3 could not be used to improve the dual location of TPT3-NaM pairs. These experiments identified two transferred PCR that can efficiently and specifically transfer NaM-TPT3 to different natural pairs, thus offering a promising TPT3-NaM location method.
为了进一步评价该方法的实用性，使用基于凝胶的分析通过条带亮度比 （灰度比） 分析 1N 模板的 PCR 扩增效率。结果表明，它们对 dTPT3TP/dNaMTP 的复制效率相当高，仅对 dNaM 或 disoTATTP/dNaMTP 具有 1.0：1.2：0.8 的灰度比（补充图 S7）。此外，还分析了 PCR 产物的原始测序信号。它们在 disoTATTP/dNaMTP 或仅 NaM 存在下没有减少（补充图 S8），这也证明了非天然碱基转化为天然碱基。此外，我们分析了 UBP 前后所有其他天然碱基的信号，旨在发现桥基 isoTAT 对天然碱基对复制的可能不利影响。令人欣慰的是，除 UBP 位点外，所有测序信号中均未检测到天然碱基突变（补充图 S8）。这可能是由于天然碱基对 NaM-isoTAT 和 isoTAT-G 的配对效率较高（A-T 和 C-G 的 K_m/V_最大值均超过 10⁹）以及 isoTAT 和其他天然碱基之间检测不到的配对效率。最后，使用 dTPT3TP 和天然 dNTPs 进行 PCR 测定，探讨 TPT3 在没有 NaM 的情况下是否具有其固有的偏好，在模板 UN 中 TPT3-NaM UBP 的预定位置仅获得杂乱信号（补充图 S9），表明 TPT3 不能用于改善 TPT3-NaM 对的双重定位。 这些实验确定了两种转移的 PCR，它们可以有效、特异性地将 NaM-TPT3 转移到不同的天然对，从而提供了一种有前途的 TPT3-NaM 定位方法。

Sequence-dependence analysis.
序列依赖性分析。

Because the sequence dependence has shown a detrimental effect on both in vitro and in vivo replication of hydrophobic UBPs (27,31), we considered that a detailed investigation of the sequence dependence of the current transferred PCRs should be necessary for promoting its general applications. So we ran the deep sequencing on the PCR products of UN template with the nucleotide conditions of Group-I (disoTATTP, dNaM, and natural dNTPs) and Group-II (dNaM and natural dNTPs) respectively. Samples with triplicate were normalized to 1 × 10⁶ read counts and further analyzed. For the bridge-base transformation samples, a standard bell-shape frequency distribution around the median number (244) indicated that there were no significant deviation differences for the detected sequence combinations (Figure 4A). In contrast, the frequency distribution of NaM-mediated transformation samples demonstrated a less regular bell shape with some sequence combinations bearing less than 100 reads or more than 750 reads, suggesting a few special sequence contexts bearing abnormalized preferences (Figure 4B). However, for both two transformations, we could obtain 4096 kinds of readthrough sequences, which covered all the possible sequences of the random arrangements. The total 100% read-through ratios guaranteed the generality of the methods for applications in various DNA technologies.
由于序列依赖性已显示出对疏水性 UBP 的体外和体内复制的不利影响 （27,31），因此我们认为有必要详细研究当前转移的 PCR 的序列依赖性，以促进其一般应用。因此，我们分别以 I 组 （disoTATTP、dNaM 和天然 dNTPs） 和 II 组 （dNaM 和天然 dNTPs） 的核苷酸条件对 UN 模板的 PCR 产物进行深度测序。将一式三份的样品归一化为 1 × 10⁶ 个读取计数并进一步分析。对于桥基变换样本，围绕中位数 （244） 的标准钟形频率分布表明检测到的序列组合没有显着的偏差差异（图 4A）。相比之下，NaM 介导的转化样本的频率分布表现出不太规则的钟形，一些序列组合的读数少于 100 个或超过 750 个读数，这表明一些特殊的序列上下文具有异常偏好（图 4B）。然而，对于这两种变换，我们都可以获得 4096 种透读序列，涵盖了随机排列的所有可能序列。总的 100% 读通率保证了方法在各种 DNA 技术中的应用的通用性。

The detailed effects of nearby bases on the UBP’s transformations could also be analyzed by the deep-sequencing data. The single base distribution from the before and behind three positions for the bridge-base transformation was calculated first. As shown in Figure 4C, no base preferences could be found for the detected six natural bases’ positions (each of them was about 25%), and all over 66% G signals indicated the effective transformation of NaM to G. With the same analysis method, all over 60% T signals indicated the effective transformation to T under the NaM-mediated transformation and base's preferences could be found for the three position downstream (Supplementary Figure S10). Secondly, we evaluated the effects of dinucleotides in the bridge-base transformation. As shown in Figure 4D, for all the 32 combinations upstream or downstream, no special dinucleotide combinations showed dramatic preferences, with the number of reads ranging from about 40 000 to 80 000. For the NaM-mediated transformation, the read counts of combinations upstream showed no dramatic preferences and were similar to those of bridge-base transformation, while the read counts of combinations downstream showed dramatic preferences, TT and TA combinations showed high frequency with a number over 100000, and AA, AG, GA, and GT showed low frequency with the number <40 000 (Supplementary Figure S11). Such preferences could give a reasonable explanation for the less regular bell-shape distribution shown in Figure 4B and S10.
附近碱基对 UBP 转化的详细影响也可以通过深度测序数据进行分析。首先计算桥基变换前后三个位置的单个碱基分布。如图 4C 所示，检测到的 6 个天然碱基位置没有找到碱基偏好（每个碱基约为 25%），所有超过 66% 的 G 信号都表明 NaM 向 G 的有效转化。使用相同的分析方法，所有超过 60% 的 T 信号都表明在 NaM 介导的转化下有效转化为 T，并且可以在下游的三个位置找到碱基的偏好（补充图 S10）。其次，我们评估了二核苷酸在桥基转化中的作用。如图 4D 所示，对于上游或下游的所有 32 种组合，没有特殊的二核苷酸组合显示出显著的偏好，读取数范围约为 40 000 至 80 000 次。对于 NaM 介导的转化，上游组合的读取计数没有表现出显着的偏好，与桥基转化相似，而下游组合的读取计数显示出显着的偏好，TT 和 TA 组合显示高频，数字超过 100000，AA、AG、GA 和 GT 显示低频率，数字<40 000 (补充图 S11).这种偏好可以为图 4B 和 S10 中所示的不太规则的钟形分布提供合理的解释。

Finally, we analyzed the effects of triple bases on the UBP’s transformations, the results were shown in Figures 5A, B, and S12, 13. For all the 4096 combinations of bridge-base transformation, TPT3-NaM were mainly transferred to C–G in about 95% of the combinations, except sequence contents with CGG and GGG downstream and a few specific sequence combinations were transformed to G–C, A–T and T–A (Figure 5A and S12). For the 4096 combinations of NaM-mediated, TPT3-NaM were mainly transferred to A–T in about 91% of the combinations, except part of sequence contexts with GGA, all sequence contexts with GGG downstream, and a few specific sequence combinations were transformed to G–C, A–T and T–A (Figures 5B and S13). Although variable transformation in some sequence contexts mediated by bridge-base or NaM were also different, about 96% of the total sequence combinations could complete the dual location of NaM-TPT3 pair. There were 4% triple-base sequence contexts that did not support the dual location of NaM-TPT3 pair possibly because of their unique combinations (Supplementary Table 1). Taken together, these results indicate our dual location method possesses the characteristic of high read-through ratios and low sequence-dependent properties.
最后，我们分析了三重碱基对 UBP 变换的影响，结果如图 5A 、 B 和 S12 、 13 所示。对于桥基转化的所有 4096 个组合，TPT3-NaM 在约 95% 的组合中主要转移到 C-G，除了下游具有 CGG 和 GGG 的序列内容和一些特异性序列组合被转化为 G-C、A-T 和 T-A （图 5A 和 S12）。对于 NaM 介导的 4096 个组合，TPT3-NaM 在约 91% 的组合中主要转移到 A-T，除了部分具有 GGA 的序列上下文，所有具有 GGG 下游的序列上下文，以及一些特定的序列组合被转化为 G-C、A-T 和 T-A（图 5B 和 S13）。尽管在一些由桥基或 NaM 介导的序列上下文中的可变转化也不同，但大约 96% 的总序列组合可以完成 NaM-TPT3 对的双重定位。有 4% 的三碱基序列上下文不支持 NaM-TPT3 对的双重定位，可能是因为它们的独特组合（补充表 1）。综上所述，这些结果表明我们的双定位方法具有高读通率和低序列依赖性的特点。

Tracing the in vivo replication of multiple NaM-TPT3 pairs in the plasmids.
追踪质粒中多个 NaM-TPT3 对的体内复制。

There were increasing requirements for the incorporation of multiple TPT3-NaM UBPs or UBPs with non-default loci, for example in vivo replications in SSO to insert more ncAAs and SELEX. So the tool for tracing multiple UBPs or UBPs with non-default loci is urgently needed, which is limited by existing methods such as the Sanger-terminal method and biotin-shift assays (Figure 6A). And the emerging data have shown that some sequence contexts are less acceptable for in vivo replications (more harsh sequence requirements than that of in vitro replications (27), thus needing to examine thoroughly the mutation products of UBPs-bearing plasmid after their doublings. Here we chose the plasmid that contained one or three UBPs with different neighboring bases for the proof-of-concept retention/mutation analysis (Figure 6B).
对掺入多个 TPT3-NaM UBP 或具有非默认基因座的 UBP 的要求越来越高，例如在 SSO 中进行体内复制以插入更多的 ncAA 和 SELEX。因此，迫切需要用于追踪多个 UBP 或具有非默认基因座的 UBP 的工具，而该工具受到 Sanger 末端法和生物素偏移分析等现有方法的限制（图 6A）。新出现的数据表明，一些序列环境对于体内复制不太可接受（比体外复制更严格的序列要求 （27），因此需要彻底检查携带 UBP 的质粒在倍增后的突变产物。在这里，我们选择了包含一个或三个具有不同相邻碱基的 UBP 的质粒进行概念验证保留/突变分析（图 6B）。

The plasmids were prepared, transfected based on the literature methods², and extracted until OD₆₀₀ reach a range of 0.6–2 (1N: 1.7; 3N: 0.7). The extracted plasmids were amplified via replacement PCR of 36 cycles with bridge base and NaM or only NaM. The dual location method allows us to analyze the detailed changes of the UBPs’ mutations in vivo. The results showed that NaM was almost all transformed into G or T by bridge base or only NaM PCR in the plasmid containing one UBP with the three base combination GXA (X = NaM), respectively (Figures 6C and S14). This might be due to the high retention of the UBP as seen in Supplementary Figure S15. Intriguingly, NaM bases in the plasmid containing three UBPs were converted into a mixture of A and G or A and T in the triple base combination of GXA, and into a mixture of T and G or T in the triple base combination of TXT, indicating a certain degree of retention as well as mutations of NaM to A in GXA and T in TXT combination. While in CXT base combination, a nearly complete mutation of NaM was to A (Figure 6E, Supplementary Figure S16). Our method directly gave the information of UBPs’ mutations of DNA bearing multiple UBPs replicated in SSO, we expected it would be applied to other applications to identify multiple UBPs at unknown sites.
制备质粒，根据文献方法² 转染，并提取至 OD₆₀₀ 达到 0.6–2 的范围 （1N： 1.7; 3N： 0.7）。提取的质粒通过桥基和 NaM 或仅 NaM 的 36 个循环的替代 PCR 进行扩增。双重定位方法使我们能够分析 UBP 突变在体内的详细变化。结果表明，在含有一个 UBP 的质粒中，NaM 几乎全部通过桥式碱基或仅 NaM PCR 转化为 G 或 T，分别具有三个碱基组合 GXA （X = NaM）（图 6C 和 S14）。这可能是由于 UBP 的高保留率，如补充图 S15 所示。有趣的是，含有三个 UBP 的质粒中的 NaM 碱基在 GXA 的三重碱基组合中转化为 A 和 G 或 A 和 T 的混合物，在 TXT 的三重碱基组合中转化为 T 和 G 或 T 的混合物，表明一定程度的保留以及 NaM 在 GXA 中对 A 的突变，在 TXT 组合中对 T 的突变。而在 CXT 碱基组合中，NaM 几乎完全突变为 A （图 6E，补充图 S16）。我们的方法直接给出了 UBP 突变的信息，即在 SSO 中复制的带有多个 UBP 的 DNA 突变，我们预计它会应用于其他应用，以识别未知位点的多个 UBP。

Beyond the qualitative analysis of base distribution, we attempted the possibility to employ the bridge base for quantitative analysis of contents of retained bases on the positions of UBPs after in vivo replications by the standard curve methods. 1N and 3N templates were amplified by bridge base to transform NaM-TPT3 to G–C, then the products were mixed with natural sequences in different proportions and sequenced (Supplementary Figures S17 and S18). The signal ratio of C on the antisense strand was used for generating the standard curves (Supplementary Figure S19). First, we tested the plasmid-A sample bearing one TPT3-NaM UBP with the three-base combination GXA. With the standard curve built for this sequence, the UPB retention was calculated as 88%, which was comparable to 82% that was measured by signal attenuation assays with NaM-TPT3 as some unnatural bases would be lost during PCR (Figure 6D and Supplementary Figures S15 and S20). Then the plasmid-B samples bearing three TPT3-NaM UBPs were analyzed. Interestingly, the retention or mutation ratio for the UBPs with different neighboring bases in the same strand changes obviously. Taking the antisense strand for calculation, for the GXA base combination, 72% retention could be obtained. In contrast, 76% retention of UBP was obtained for the TXT base combination, and only 7.5% retention of UBP for the CXT base combination (Figure 6F and S21, 22). Finally, sequences were completely read-through with no detectable mutations of natural bases found by our method, indicating the desirable bioorthogonality of NaM-TPT3 for applications in vivo replications (Figure 6C, E, and Supplementary Figure S23). In all, we show that our dual location method is a powerful tool for tracing multiple UBPs or UBPs with non-default loci.
除了碱基分布的定性分析之外，我们还尝试了通过标准曲线方法在体内复制后使用 UBP 位置使用桥基对保留碱基含量进行定量分析的可能性。通过桥基扩增 1N 和 3N 模板，将 NaM-TPT3 转化为 G-C，然后将产物以不同比例与天然序列混合并测序（补充图 S17 和 S18）。反义链上 C 的信号比用于生成标准曲线（补充图 S19）。首先，我们用三个碱基组合 GXA 测试了带有一个 TPT3-NaM UBP 的质粒 A 样品。通过为该序列构建的标准曲线，UPB 保留率计算为 88%，这与使用 NaM-TPT3 信号衰减测定测量的 82% 相当，因为一些非天然碱基在 PCR 过程中会丢失（图 6D 和补充图 S15 和 S20）。然后分析带有 3 个 TPT3-NaM UBP 的质粒 B 样品。有趣的是，同一链中具有不同相邻碱基的 UBP 的保留率或突变率明显变化。以反义链计算，对于 GXA 碱基组合，可以获得 72% 的保留。相比之下，TXT 碱基组合的 UBP 保留率为 76%，而 CXT 碱基组合的 UBP 保留率仅为 7.5%（图 6F 和 S21、22）。 最后，序列被完全读取，我们的方法没有发现可检测的天然碱基突变，这表明 NaM-TPT3 在体内复制应用中具有理想的生物正交性（图 6C、E 和补充图 S23）。总而言之，我们表明我们的双重定位方法是追踪多个 UBP 或具有非默认基因座的 UBP 的强大工具。

Additionally, we also compared our data with the reported retention ratios for which each triple base combination was evaluated individually (27). The triple base combinations measured by our bridge base method showed that TXT had good retention efficiency, GXA showed moderate tolerance, and CXT was the most affected (Figure 6D and F). This was in agreement with the previous report, although the report also showed that the retention efficiencies of UBPs were influenced by plasmid types, triple base combinations, and strains of E. coil (27). Apart from the copies of plasmid bearing UBPs that were different (OD₆₀₀, 1N: 1.7; 3N: 0.7), taking the retention efficiency of GXA in two plasmids into consideration (Figure 6D and F), we speculated that more sequence contexts beyond the neighboring bases and the close arrangement of multiple UBPs might also play a role in controlling the retention of UBPs during their in vivo replication. To test the hypothesis, kf-mediated kinetic assays have been applied to reflect the possible sequence dependence of TPT3-NaM UBPs that frequently occurs in in vivo replications (26). We synthesized nine flanking sequences with NaM and three flanking sequences with TPT3 in templates to explore the influence of the flanking sequence, including templates with three near-flanking nucleotides the same as that replicated in vivo. The mispairing abilities are thought to be an important factor that leads to UBP loss (27). We found that the sequences farther from the adjacent three bases could slightly affect the mispairing capacity of UBPs (Supplementary Figure S24); the adjacent three bases on the upstream and downstream of UBPs had more powerful effects on their mispairing tendency (Supplementary Figure S25). To further explore whether the adjacent three bases on the upstream and downstream of UBP affected the mispairing level, kinetic analysis using templates with three near-flanking nucleotides the same as that replicated in vivo was also performed. Mispairing levels of NaM and TPT3 in one site were added together as total mispairing rates. We found that the TXT, GXA, and CXT on the same sequence context also showed different mispairing capacities, in the order CXT > GXA > TXT (Supplementary Figure S26), which was consistent with the fidelity of in vivo replication of the three combinations (TXT > GXA > CXT). These results suggest that more flanking sequences should be taken into account for high-fidelity replication in vivo.
此外，我们还将我们的数据与报告的保留率进行了比较，其中每个三重碱基组合都经过单独评估 （27）。通过我们的桥基方法测量的三重碱基组合显示，TXT 具有良好的保留效率，GXA 表现出中等耐受性，而 CXT 受影响最大（图 6D 和 F）。这与之前的报告一致，尽管该报告还显示 UBP 的保留效率受质粒类型、三重碱基组合和 E. coil 菌株的影响 （27）。除了不同的带有质粒的 UBP 拷贝（OD_600,1N：1.7;3N：0.7）外，考虑到 GXA 在两种质粒中的保留效率（图 6D 和 F），我们推测相邻碱基之外的更多序列上下文和多个 UBP 的紧密排列也可能在控制 UBP 在体内保留中发挥作用复制。为了检验这一假设，已经应用了 kf 介导的动力学测定来反映 TPT3-NaM UBP 的可能序列依赖性，这种依赖性经常发生在体内复制中 （26）。我们在模板中合成了 9 个 NaM 侧翼序列和 3 个 TPT3 侧翼序列，以探索侧翼序列的影响，包括具有三个近侧翼核苷酸的模板，与体内复制的核苷酸相同。错配能力被认为是导致 UBP 损失的重要因素 （27）。 我们发现，远离相邻三个碱基的序列会略微影响 UBP 的错配能力（补充图 S24）;UBPs 上下游的相邻 3 个碱基对其错配倾向的影响更强（补充图 S25）。为了进一步探讨 UBP 上游和下游相邻的三个碱基是否影响错配水平，还使用具有与体内复制相同的三个近侧翼核苷酸的模板进行了动力学分析。将一个位点中 NaM 和 TPT3 的错配水平相加为总错配率。我们发现，在同一序列背景下的 TXT、GXA 和 CXT 也显示出不同的错配能力，CXT > GXA > TXT 的顺序（补充图 S26），这与三种组合的体内复制保真度一致 （TXT > GXA > CXT）。这些结果表明，对于体内的高保真复制，应考虑更多的侧翼序列。

Sequencing multiple-site DNA lesions and aptamer containing multiple TPT3.
对多位点 DNA 损伤和包含多个 TPT3 的适配体进行测序。

The exclusive pairing capacity of UBPs enables their successful applications in a variety of biotechnologies, among which Burrow and coworkers have used 5SICS-NaM as the third pair for the identification of DNA lesions such as abasic site (AP site), dU and dOG (22). Lesion sites were replaced by 5SICS-NaM pair via the base excision repair process and PCR amplification. Then Sanger sequencing was performed to determine the position of a single lesion site depending on the terminal signal. And α-hemolysin nanopore method was used to detect more than one lesion site via identifying the substituted NaM with post modified reporting group. Here we provided a proof-of-concept for improving the identification of such multiple lesions by our bridge bases (Figure 7A). As lesion excision and unnatural nucleotide insertion were similar via the base excision repair process, dU was used to represent varied lesions. More than one dU site in the strand that was marked by TPT3-NaM could be read out after running the twice PCR assays without the involvement of specialized instruments (Figure 7B).
UBP 的独有配对能力使其能够在各种生物技术中成功应用，其中 Burrow 及其同事已使用 5SICS-NaM 作为鉴定 DNA 损伤的第三对，如脱碱基位点（AP 位点）、dU 和 dOG （22）。通过基底切除修复过程和 PCR 扩增，病灶部位被 5SICS-NaM 对替换。然后进行 Sanger 测序，根据终末信号确定单个病灶部位的位置。采用 α-溶血素纳米孔法，通过鉴定取代的 NaM 和改良后报告组来检测多个病变部位。在这里，我们提供了一个概念验证，用于改进我们的桥基对此类多发性病变的识别（图 7A）。由于病灶切除和非自然核苷酸插入通过碱基切除修复过程相似，因此使用 dU 表示不同的病灶。在运行两次 PCR 检测后，可以在没有专用仪器参与的情况下读出链中由 TPT3-NaM 标记的多个 dU 位点（图 7B）。

We prepared the template DNA (KRAS-1U, 2U and 134–2U) bearing dU lesions according to the literature (22), and then validated the performance of our method. Briefly, uracil-DNA glycosylase (UDG) was employed to yield AP sites, and apurinic/apyrimidinic Endonuclease 1 (APE1) was used to yield gap sites without the sugar fragment. Then Kf (exo⁻) enzyme was used to insert the dTPT3TP^biotin (Figure 7C). In order to seal the marker nucleotide at the lesion site, T4-DNA ligase was added to the reaction mixture. The procedures for gap-formation reaction and ligation reaction were analyzed by denaturing PAGE gel (Supplementary Figure S27). Two shorter DNA strands were found after treating the template DNA with APE1, and full-length products could be detected after T4-DNA ligase was added. Then the reaction products were used for PCR amplification with dTPT3TP^biotin and NaM, and the PCR products were conveniently enriched by biotin-streptavidin-based strand shift (Supplementary Figure S25).
我们根据文献 （22） 制备了带有 dU 损伤的模板 DNA （KRAS-1U、2U 和 134-2U），然后验证了我们方法的性能。简而言之，尿嘧啶-DNA 糖基化酶 （UDG） 用于产生 AP 位点，使用脱嘌呤/脱嘧啶核酸内切酶 1 （APE1） 产生没有糖片段的间隙位点。然后使用 Kf （exo ^- ） 酶插入 dTPT3TP^生物素（图 7C）。为了将标记核苷酸密封在病变部位，将 T4-DNA 连接酶添加到反应混合物中。通过变性 PAGE 凝胶分析间隙形成反应和连接反应的过程 （补充图 S27）。用 APE1 处理模板 DNA 后发现两条较短的 DNA 链，加入 T4-DNA 连接酶后可检测到全长产物。然后，将反应产物用于 dTPT3TP^生物素和 NaM 的 PCR 扩增，并通过基于生物素-链霉亲和素的链转移方便地富集 PCR 产物（补充图 S25）。

The enriched strips, expected to contain the TPT3^biotin-NaM marked DNA lesion, were used as the templates for bridge-base PCR. The single strand DNAs containing one or two dU were also used as templates for PCR amplification to make up the dual location method through the pairing of dU with A, distinguishing signal conversion of bridge-base PCR. With the one-dU-site-containing DNA sample, the dU site replaced with TPT3^biotin was transferred to C, while the original dU site was transferred to T (Figure 7D). With the two-dU-site-containing DNA sample, dU sites replaced with TPT3^biotin were both transferred to C, while original dU sites were both transferred to T (Figure 7E). The results indicate that our TPT3^biotin can be inserted into dU lesion sites and transferred to C by bridge-base PCR, and dU can be identified in this model system.
预计含有 TPT3^生物素-NaM 标记的 DNA 损伤的富集条带被用作桥基 PCR 的模板。含有 1 个或 2 个 dU 的单链 DNA 也被用作 PCR 扩增的模板，通过 dU 与 A 配对构成双定位方法，区分桥基 PCR 的信号转换。使用含有 1 dU 位点的 DNA 样品，用 TPT3^生物素替换的 dU 位点被转移到 C，而原始 dU 位点被转移到 T（图 7D）。对于含有两个 dU 位点的 DNA 样品，用 TPT3^生物素替换的 dU 位点都转移到 C 位点，而原始 dU 位点都转移到 T 上（图 7E）。结果表明，我们的 TPT3^生物素可以插入 dU 病变部位并通过桥基 PCR 转移到 C，并且可以在该模型系统中鉴定 dU。

Furthermore, we applied our method to detect apurinic and apyrimidinic (AP) sites in simple biological samples. It has been reported that AP sites induced by H₂O₂ displayed a preference for regions undergoing replication or transcription (28), we attempted to evaluate the corresponding mutation levels of the nucleotides on DNA fragments in replication regions of pUC-19 plasmids. pUC-19 plasmids were transformed into E. coli DH5α cells, and damages of AP sites were induced by H₂O₂. Labeling AP sites in plasmid DNA with unnatural nucleotide dTPT3TP^biotin, isolation of labeled DNA, replacement PCR, and Deep sequencing were performed as described above. If one nucleotide was damaged to AP site, it will be transformed to A or C by replacement PCR. For A and C, only the mutation levels of C or A could be calculated respectively, as they were one of the nucleotides being transformed. We found that the mutation ratios of A to C and C to A were both selectively increased in some sites (Supplementary Figure S28). Moreover, we also found that mutation ratios of some G and T sites increased significantly and most sites with increased A or C signals overlapped well especially these increased most, while the mutation ratios of other sites were near the baseline (Figure 8 and Supplementary Figure S29). These indicate that AP sites may be accumulated in some preference regions, and our bridge method can be used to determine DNA lesions such as AP site.
此外，我们应用我们的方法检测简单生物样品中的无嘌呤和无嘧啶 （AP） 位点。据报道，H₂O₂ 诱导的 AP 位点表现出对正在复制或转录的区域的偏好 （28），我们试图评估 pUC-19 质粒复制区 DNA 片段上核苷酸的相应突变水平。将 pUC-19 质粒转化到大肠杆菌 DH5α 细胞中，并通过 H₂O₂ 诱导 AP 位点的损伤。如上所述，用非天然核苷酸 dTPT3TP^生物素标记质粒 DNA 中的 AP 位点，分离标记的 DNA，替换 PCR 和深度测序。如果一个核苷酸在 AP 位点受损，则通过替代 PCR 将其转化为 A 或 C。对于 A 和 C，只能分别计算 C 或 A 的突变水平，因为它们是被转化的核苷酸之一。我们发现 A 到 C 和 C 到 A 的突变比在某些位点都选择性地增加（补充图 S28）。此外，我们还发现一些 G 和 T 位点的突变率显着增加，大多数 A 或 C 信号增加的位点重叠良好，尤其是这些位点增加最多，而其他位点的突变率接近基线（图 8 和补充图 S29).这些表明 AP 位点可能积累在某些偏好区域，我们的桥接方法可用于确定 AP 位点等 DNA 损伤。

Finally, for the SELEX application, one or more UBPs will be needed to arrange in short DNA fragments (11–13). Ichiro Hirao's team has generated aptamers containing multiple Ds-Px pairs with high affinities. Aptamers generated with hydrophobic Ds-Px pair were synthesized (11). The Ds sites were replaced by TPT3 or natural bases, and affinity analysis was performed by SPR. We found that aptamer with two TPT3 showed a similar affinity with that of natural bases aptamer, and aptamer with three TPT3 showed improved affinity than that with natural bases (Supplementary Figure S30 and 9A), which indicated that NaM-TPT3 pair or at least TPT3 could be used in SELEX. The aptamer with three TPT3 was also amplified via PCR with NaMTP/TPT3TP or replacement PCR. We also showed that aptamers with three TPT3 will be terminated after the first UBP without transformation and could be sequenced by our bridge base method without sequence variations (Figure 9B and C). In a word, these results indicate that multiple NaM-TPT3 UBPs that exist in aptamers can be easily detected by our bridge base method, shedding new light on the application of NaM-TPT3 pair in SELEX screening.
最后，对于 SELEX 应用，需要一个或多个 UBP 来排列短 DNA 片段 （11–13）。Ichiro Hirao 的团队已经生成了包含多个具有高亲和力的 Ds-Px 对的适配体。合成了用疏水性 Ds-Px 对生成的适配子 （11）。Ds 位点被 TPT3 或天然碱取代，并通过 SPR 进行亲和力分析。我们发现具有两个 TPT3 的适配子显示出与天然碱基适配体相似的亲和力，具有三个 TPT3 的适配子显示出比具有天然碱基的适配子更好的亲和力（补充图 S30 和 9A），这表明 NaM-TPT3 对或至少 TPT3 可用于 SELEX。具有三个 TPT3 的适配体也通过 NaMTP/TPT3TP 的 PCR 或替代 PCR 扩增。我们还表明，具有三个 TPT3 的适配体将在第一个 UBP 后终止而不转化，并且可以通过我们的桥基方法进行测序，没有序列变化（图 9B 和 C）。总之，这些结果表明，我们的桥基方法可以很容易地检测到适配体中存在的多个 NaM-TPT3 UBP，为 NaM-TPT3 对在 SELEX 筛选中的应用提供了新的思路。

DISCUSSION 讨论

In this manuscript, we have designed and synthesized an isoTAT base to facilitate the tracing of TPT3-NaM UBPs for variant research purposes. In particular, we found that the isoTAT bearing similar shapes with TPT3 but possessing additional hydrogen bond acceptor has an intriguing pairing property to pair with NaM and G with almost equal efficiency. For example, in the single-nucleotide insertion experiments, the % incorporations of disoTATTP opposite NaM or G in the templates are 90% and 94% respectively, and the Km/Vmax constants of disoTATTP opposite NaM or G in templates are 0.77 × 10 ⁸ and 2.14 × 10 ⁸ respectively (Figure 2 and Table 1). More interestingly, the dual pairing capacity of isoTAT with NaM and G is exclusive. The % incorporations of disoTATTP opposite A/T/C in templates are all less than 10%, and the Km/Vmax constants of disoTATTP opposite A/T/C in templates are lower than the detectable level. Although nature has such kind of dual/multiple-pairing bases such as hypoxanthine (I) that can acts as a ‘wobble base’ in the third position of codons in tRNA enhancing its reading of anticodon in mRNA, which shows a promising perspective for various applications. There are seldom reports about the design of such kinds of artificial bases that can pair well with a certain expanded genetic letter and natural genetic letters. The finding of isoTAT indicates that the recognitions of UBPs themselves and natural base pairs are not insulated, which can be connected directionally by a proper structure called a bridge base. Moreover, the design of bridge bases can be dependent on the hydrophobic interaction forces, shape complementary, and hydrogen-bond rearrangements that have been shown invaluable tools for current UBP designing. In this regard, the strategy for the combination of parent skeletons and additional hydrogen-bonding acceptor/doner that are used for isoTAT will be also applicable to other unnatural base pairs for different research purposes.
在这份手稿中，我们设计并合成了一个 isoTAT 碱基，以促进 TPT3-NaM UBP 的追踪用于变体研究目的。特别是，我们发现具有与 TPT3 相似形状但具有额外氢键受体的 isoTAT 具有与 NaM 和 G 配对的有趣配对特性，效率几乎相同。例如，在单核苷酸插入实验中，模板中 NaM 或 G 对立的 disoTATTP 掺入百分比分别为 90% 和 94%，模板中 NaM 或 G 对立的 disoTATTP 的 Km/Vmax 常数分别为 0.77 × 10 ⁸ 和 2.14 × 10 ⁸（图 2 和表 1).更有趣的是，isoTAT 与 NaM 和 G 的双重配对能力是排他性的。模板中 A/T/C 对立的 disoTATTP 掺入百分比均小于 10%，模板中 A/T/C 对立的 disoTATTP 的 Km/Vmax 常数低于可检测水平。尽管自然界有这种双/多配对碱基，例如次黄嘌呤 （I），它可以在 tRNA 密码子的第三位置充当“摆动碱基”，从而增强其对 mRNA 中反密码子的读取，这为各种应用显示出广阔的前景。很少有关于这种人工碱基设计的报道，它们可以与某个扩展的遗传字母和天然遗传字母很好地配对。isoTAT 的发现表明，UBP 本身和天然碱基对的识别不是绝缘的，它们可以通过称为桥基的适当结构进行定向连接。 此外，桥梁基础的设计可能取决于疏水相互作用力、形状互补和氢键重排，这些已成为当前 UBP 设计的宝贵工具。在这方面，用于 isoTAT 的母体骨架和额外的氢键受体/供体的组合策略也将适用于用于不同研究目的的其他非天然碱基对。

TPT3-NaM is one of the advanced UPBs in the expanded genetic letters (2,8,19,20). There were increasing requirements for incorporations of multiple TPT3-NaM UBPs with or without defaulted loci. More UBPs are used as genetic letters which increases the complexity of DNA, but monitoring the DNA containing multiple UBPs with or without defaulted loci by simple and convenient approaches is challenging by existing methods. We develop a dual location method based on the use of bridge base and the inherent bases’ preference of NaM, and our method can be applied to almost all the sequence contexts, allowing to monitor almost all changes in the sequence. So it is a useful approach for detecting more than one UBP on the unknown or default positions. Besides, the applications of bridge base to analyze the in vivo replicated DNA samples containing TPT3-NaM enable researchers to evaluate the fidelity as well as changes of multiple UBPs at the adjacent sites at the same time, which is also hard to be conducted by the biotin-shift assays that are mainly used currently. Considering the unexpected complexity that may be encountered when using TPT3-NaM to recode various proteins and the emerging sequence-dependence properties that may be worse than that of in vitro applications (15,27,32), assessments and optimizations of sequence contexts around one and more TPT3-NaM UBPs may be necessary steps before protein expressions. Therefore, our bridge base method offers a simple and convenient determining approach.
TPT3-NaM 是扩增遗传字母中的高级 UPB 之一 （2,8,19,20）。对掺入多个 TPT3-NaM UBP 的要求越来越高，无论是否具有默认基因座。更多的 UBP 被用作遗传字母，这增加了 DNA 的复杂性，但通过简单方便的方法监测包含多个 UBP 的 DNA（有或没有默认位点）在现有方法中具有挑战性。我们开发了一种基于桥基的使用和 NaM 的固有基偏好的双重定位方法，我们的方法可以应用于几乎所有的序列上下文，从而可以监测序列中的几乎所有变化。因此，这是在未知或默认位置上检测多个 UBP 的有用方法。此外，桥基分析含有 TPT3-NaM 的体内复制 DNA 样品的应用使研究人员能够同时评估相邻位点多个 UBP 的保真度和变化，这也是目前主要使用的生物素转移分析难以进行的。考虑到使用 TPT3-NaM 重编码各种蛋白质时可能遇到的意想不到的复杂性，以及可能比体外应用更糟糕的新出现的序列依赖性特性 （15,27,32），围绕一个或多个 TPT3-NaM UBP 的序列上下文评估和优化可能是蛋白质表达之前的必要步骤。因此，我们的 bridge base 方法提供了一种简单方便的确定方法。

5SICS-NaM was used as a marker to identify DNA lesions (22). The single lesion site was identified using Sanger sequencing depending on the terminal signal with no strand location information, and more than one lesion site was identified using α-hemolysin nanopore depending on distinguishing the substituted NaM with post modified reporting group, which possessed a complex process and required expansive apparatus. We provided proof-of-concept and simple biological samples for improving the identification of such multiple lesions marked with UBPs by using our bridge bases. The TPT3 unnatural base can be accurately installed at one or more lesion sites of DNA, and the marked samples can be enriched by TPT3^biotin. The results indicate that our TPT3^biotin can be transferred to C by bridge-base PCR and different DNA lesions can be identified in this model system and biological samples. We also believe that other lesions can also be detected through NaM’ preferences and a bridge base regardless of their pairing abilities. While lesions in biological samples may be in one strand of double-strand DNA with low abundance, the effect of the other strand and normal sequences without lesions should be taken into account. So our enriched step seems to be necessary. It is also noteworthy that the marked samples bearing dTPT3^biotin can be directly used as the template for PCR with the bridge base disoTATTP with no requirements for removing the linker and attached streptavidin, simplifying the tedious streptavidin-trapping-then-cleavage process. The method will be more easily operated and generally applied in ordinary biological labs, owing to its total compatibility with the commercialized sequencing technology and no requirements for specialized instruments. In all, our method can provide a simple and convenient dual location of NaM-TPT3 pair for multiple lesions detection. Such a simple dual location platform with the capacity to determine/sequence multiple TPT3-NaM UBPs can therefore be used in various sequence contexts of nucleic acid for different research purposes.
5SICS-NaM 用作鉴定 DNA 损伤的标记物 （22）。根据没有链位置信息的末端信号，使用 Sanger 测序鉴定单个病灶部位，根据区分取代的 NaM 和修饰后报告组，使用 α-溶血素纳米孔鉴定多个病灶部位，具有复杂的过程和需要膨胀的装置。我们提供了概念验证和简单的生物样本，以使用我们的桥基来提高对这种用 UBP 标记的多发病灶的识别。TPT3 非天然碱基可以准确地安装在 DNA 的一个或多个损伤位点，标记的样品可以被 TPT3^生物素富集。结果表明，我们的 TPT3^生物素可以通过桥基 PCR 转移到 C 上，并且可以在该模型系统和生物样品中鉴定出不同的 DNA 损伤。我们还相信，其他病变也可以通过 NaM 偏好和桥基检测到，无论它们的配对能力如何。虽然生物样品中的损伤可能位于一条丰度较低的双链 DNA 链中，但应考虑另一条链和无损伤的正常序列的影响。因此，我们的 enriched 步骤似乎是必要的。还值得注意的是，带有 dTPT3^生物素的标记样品可以直接用作带有桥基 disoTATTP 的 PCR 模板，无需去除接头和连接的链霉亲和素，简化了繁琐的链霉亲和素捕获然后切割过程。 该方法将更易于操作，并且通常应用于普通生物实验室，因为它与商业化的测序技术完全兼容，并且不需要专用仪器。总之，我们的方法可以为多个病灶检测提供简单方便的 NaM-TPT3 对双重定位。因此，这种简单的双定位平台能够确定/测序多个 TPT3-NaM UBP，可用于核酸的各种序列环境，用于不同的研究目的。

DATA AVAILABILITY 数据可用性

Raw sequencing data for experiments from this study are available at Bioproject/SRA (BioProject ID PRJNA942583).
本研究实验的原始测序数据可在 Bioproject/SRA （BioProject ID PRJNA942583） 获得。

SUPPLEMENTARY DATA 补充数据

Supplementary Data are available at NAR Online.
补充数据可在 NAR Online 上获得。

FUNDING 资金

National Natural Science Foundation of China [22077027, 21778015 to L.L., 21877206 to G.Z., 32102612 to H.W.]; Central Plains Science and Technology Innovation Leader Project [214200510008 to L.L.]; Scientific and Technological Innovation Team of Colleges and Universities in Henan Province [21IRTSTHN001]; Doctoral Initiation Fund [31901014, qd18008]; Shenzhen Institute of Synthetic Biology Scientific Research Program [DWKF20210010]. Funding for open access charge: National Natural Science Foundation of China.
中国国家自然科学基金 [22077027， 21778015 to L.L.， 21877206 to G.Z.， 32102612 to H.W.];中原科技创新领军项目 [214200510008 to L.L.];河南省高等学校科技创新团队 [21IRTSTHN001];博士生基金 [31901014， qd18008];深圳市合成生物学研究院科研计划项目 [DWKF20210010].开放获取基金资助：国家自然科学基金。

Conflict of interest statement. None declared.
利益冲突声明。没有人宣布。

REFERENCES 引用

Author notes 作者注释