合理优化 siRNA,确保与 RISC 复合物相互作用时存在链偏向

文摘   科学   2023-07-07 19:00   浙江  

药悟

读完需要

40
分钟

速读仅需 14 分钟


Dhrubajyoti Datta a, Christopher S. Theile a, Kelly Wassarman a, June Qin a, Tim Racie a, Karyn Schmidt a, Yongfeng Jiang a, Rachel Sigel a, Maja M. Janas a, Martin Egli b and Muthiah Manoharan *a

*a*Alnylam Pharmaceuticals, 675 West Kendall Street, Cambridge, Massachusetts 02142, USA. E-mail: mmanoharan@alnylam.com

*b*Department of Biochemistry, School of Medicine, Vanderbilt University, Nashville, Tennessee 37232, USA

👀

   

Abstract 摘要

To ensure specificity of small interfering RNAs (siRNAs), the antisense strand must be selected by the RNA-induced silencing complex (RISC). We have previously demonstrated that a 5′-morpholino-modified nucleotide at the 5′-end of the sense strand inhibits its interaction with RISC ensuring selection of the desired antisense strand. To improve this antagonizing binding property even further, a new set of morpholino-based analogues, Mo2 and Mo3, and a piperidine analogue, Pip, were designed based on the known structure of Argonaute2, the slicer enzyme component of RISC. Sense strands of siRNAs were modified with these new analogues, and the siRNAs were evaluated in vitro and in mice for RNAi activity. Our data demonstrated that Mo2 is the best RISC inhibitor among the modifications tested and that it effectively mitigates sense strand-based off-target activity of siRNA.

为了确保小干扰 RNA(siRNA)的特异性,RNA 诱导的沉默复合物(RISC)必须选择反义链。我们先前已经证明,在正义链的 5'-末端引入 5'-morpholino 修饰的核苷酸可以抑制其与 RISC 的相互作用,确保选择所需的反义链。为了进一步改进这种拮抗结合特性,我们设计了一组基于 morpholino 的新类似物,Mo2Mo3,以及一种基于哌啶的类似物 Pip,这些设计是基于已知的 RISC 中切割酶组分 Argonaute2 的结构。用这些新的类似物对 siRNA 的正义链链进行修饰,并对其在体外和小鼠体内的 RNAi 活性进行评估。我们的数据表明,在所测试的修饰中,Mo2是最好的 RISC 抑制剂,并且有效减轻了 siRNA 基于正义链的非特异性靶向活性。


Strand selection of small interfering RNAs (siRNAs) is a critical step in RNA interference (RNAi)-mediated gene silencing, as loading of the sense strand into the RNA-induced silencing complex (RISC) can lead to off-target effects through silencing of mRNAs complementary to this strand. One driver of strand selection is thermodynamics: the strand with its 5′-terminus at the thermodynamically less stable end of the siRNA duplex is selected as the antisense strand. Moreover, 5′-end phosphorylation is a requirement for efficient loading into the RISC. Therefore, the presence of a monophosphate group or phosphate analog at the 5′-end can ensure selection of the desired strand.

小干扰 RNA(siRNA)的链选择是 RNA 干扰(RNAi)介导的基因沉默中的关键步骤,因为将正义链加载到 RNA 诱导的沉默复合物(RISC)中可能会通过沉默与该链互补的 mRNA 而产生非特异性效应。链选择的一个驱动因素是热力学:siRNA 双链结构中具有 5'-热力学不稳定末端的链被选择为反义链。此外,5'-末端的磷酸化是有效加载到 RISC 中的要求。因此,通过在 5'-末端引入磷酸单酯基团或磷酸类似物,可以确保选择所需的链。

The presence of a group that blocks 5′-end phosphorylation of the sense strand also reduces off-target effects. We previously reported synthesis of a 5′-morpholino modified nucleoside (Mo1, Fig. 1) and demonstrated that its presence at the 5′-end of the sense strand improves antisense strand selection more effectively than 5′-O-methyl or unlocked nucleic acid. When the interaction of an siRNA with a Mo1-modified strand with the MID domain of Ago2 was modeled, there was not a snug fit. We reasoned that extension of the morpholino group at the point of attachment to the nucleoside 5′-end might result in a better RISC antagonist. Thus, an extended morpholino, Mo2, was designed and synthesized (Fig. 1). We synthesized two additional novel potential antagonists, piperidine (Pip) and morpholino-N-oxy (Mo3) (Fig. 1) using aminooxy click chemistry.

阻断正义链 5'-末端磷酸化的修饰基团的存在也可以减少非特异性作用。我们先前报道了一种 5'-吗啡啉修饰核苷(Mo1,见图 1)的合成,并证明它在正义链 5'-末端比 5'-O-甲基或未锁核酸更有效地改善了反义链选择。当建模分析 siRNA 与Mo1修饰的链与 Ago2 蛋白 MID 结构域的相互作用时,发现其结合不紧密。因此,我们推测在核苷 5'-末端连接点扩展吗啡啉基团可能会产生更好的 RISC 拮抗剂。因此,设计并合成了扩展的吗啡啉(Mo2,见图 1)。我们还使用氨氧肟点击化学合成了另外两种新的潜在拮抗剂,哌啶(Pip)和吗啡啉-N-氧(Mo3,见图 1)。

[
图. 1 所研究的修饰.

Mo2 was synthesized as shown in. Commercially available nucleoside 1 was oxidized to the aldehyde 2 following the literature procedure. A Wittig reaction on compound 2 produced compound 3 in good yield. Hydroboration of 3 with 9-borabicyclo[3.3.1]nonane followed by oxidation afforded compound 4 (Table S1, ESI). The primary hydroxyl group of 4 was then tosylated to compound 5 followed by neat morpholine treatment to produce 6. Deprotection of the tert-butyldimethylsilyl (TBS) group using tetrabutylammonium fluoride (TBAF) afforded 7, which was then converted to phosphoramidite 8 by standard phosphitylation.

根据文献方法,通过将商业上可获得的核苷1氧化为醛2,合成了Mo2(详见图示 1 合成路线)。在2化合物上进行的维蒂希缩合反应以良好的产率合成了化合物3。将3与 9-硼杂环[3.3.1]壬烷进行氢硼化反应,然后氧化得到化合物4(见表 S1,ESI)。将4的主要羟基进行甲磺酰化反应,得到化合物5,然后经过无溶剂吗啡啉处理得到6。使用四丁基氟化铵(TBAF)去保护6中的tert-丁基二甲基硅基(TBS)基团,得到7,然后通过标准磷酸化反应将其转化为磷酰胺8

图示 1 Mo2磷酰胺酯8的合成方案。

To synthesize the Pip and Mo3 building blocks (Scheme 2), 9 was synthesized with an aminooxy (-ONH2) group at the 5′-end of nucleoside. Reductive amination of 9 with glutaraldehyde resulted in compound 10. Removal of the TBS protecting group afforded compound 12. Phosphitylation of 12 yielded phosphoramidite 14. Similarly, reaction of 2-(2-oxoethoxy)acetaldehyde with 9 under reductive amination conditions produced 11, which, upon deprotection of the silyl group with TBAF, resulted in compound 13. Compound 13 was phosphitylated to afford the phosphoramidite 15 in moderate yield.

为合成PipMo3的构建块(图示 2),我们使用带有氨氧基(-ONH2)的核苷酸合成了9。将9与戊二醛进行还原胺化反应得到化合物10。去除 TBS 保护基后得到化合物12。对12进行磷酸化反应得到磷酰胺酯14。类似地,将 2-(2-氧乙基)乙醛与9在还原胺化条件下反应产生11,通过 TBAF 去保护矽基后得到化合物13。对13进行磷酸化反应得到磷酰胺酯15,收率适中

图示 2 展示了Pip 磷酸酯14Mo3 磷酸酯15 的合成过程。

The modified building blocks 8, 14, and 15 were incorporated at the 5′-ends of oligonucleotides using standard oligonucleotide synthesis (Fig. S1, ESI). These building blocks were used to synthesize both sense and antisense strands of siRNAs (Table S2, ESI). In the parent siRNA, the 5′-terminal nucleotide is 2′-O-methyl (2′-OMe) U. We first evaluated silencing of ApoB expression in mice by siRNA with sense strands modified with Mo1, Mo2, Pip, and Mo3 (siRNAs I, III, IV, and V, respectively, Table 1). Mice were treated subcutaneously with 3 mg kg−1 of siRNA. The sense strands of these siRNAs were conjugated to trivalent N-acetylgalactosamine (Fig. S2, ESI), to deliver siRNAs into hepatocytes after subcutaneous injection. Circulating ApoB protein was quantified using an ELISA assay. As previously observed, we found that siRNA activity was improved compared to the parent compound when the sense strand was conjugated with Mo1. The duplexes with sense strands modified with Mo2, Pip, and Mo3 had better activity than the siRNA with the Mo1 modification (Fig. 2).

经标准寡核苷酸合成方法,在寡核苷酸的 5'-端引入了改良的磷酸酯81415(见图 S1,ESI)。这些磷酸酯用于合成 siRNA 的正义链和反义链(见表 S2,ESI)。在原始 siRNA 中,5'-端核苷酸为 2'-O-甲基尿苷(2'-OMe U)。我们首先评估了正义链使用Mo1Mo2PipMo3修饰的 siRNA 对小鼠ApoB基因表达的抑制作用(分别为 siRNA IIIIIVV,表 1)。小鼠皮下注射 3 mg kg-1 的 siRNA。这些 siRNA 的正链与三价的N-乙酰半乳糖胺共轭(见图 S2,ESI),在皮下注射后将 siRNA 输送至肝细胞。采用 ELISA 测定循环中的ApoB蛋白。与原始化合物相比,如先前观察到的,当正链与Mo1共轭时,siRNA 的活性得到改善。正链修饰有Mo2PipMo3的双链比具有Mo1修饰的 siRNA 具有更好的活性(图 2)。

表 1 用于in vivo ApoB试验的 iRNA 双链

DuplexSense strand (upper) and antisense strand (lower)a(5′–3′)
Parentu•g•UgAcAaAUAuGgGcAuCa*A*L

u•U•gAuGcCcAuauUuGuCaCa•a•a
IMo1•g•UgAcAaAUAuGgGcAuCa*A*L

u•U•gAuGcCcAuauUuGuCaCa•a•a
IIu•g•UgAcAaAUAuGgGcAuCa*A*L

Mo1U•gAuGcCcAuauUuGuCaCa•a•a
IIIMo2•g•UgAcAaAUAuGgGcAuCa*A*L

u•U•gAuGcCcAuauUuGuCaCa•a•a
IVPip•g•UgAcAaAUAuGgGcAuCa*A*L

u•U•gAuGcCcAuauUuGuCaCa•a•a
VMo3•g•UgAcAaAUAuGgGcAuCa*A*L

u•U•gAuGcCcAuauUuGuCaCa•a•a
VIu•g•UgAcAaAUAuGgGcAuCa*A*L

Mo2UsgAuGcCcAuauUuGuCaCa•a•a
VIIu•g•UgAcAaAUAuGgGcAuCa*A*L

PipU•gAuGcCcAuauUuGuCaCa•a•a
VIIIu•g•UgAcAaAUAuGgGcAuCa*A*L

Mo3U•gAuGcCcAuauUuGuCaCa•a•a

a Chemical modifications are indicated as follows: •, PS linkage; lower case, 2′-OMe; italicized upper case, 2′-fluoro; L, trivalent-GalNAc respectively.

图 2 在小鼠中,对感受链进行 morpholino 或哌啶衍生物修饰增强了沉默效果。小鼠(每组n=3)接受针对ApoB的 siRNA IIIIIVV的单剂量治疗(3 mg/kg)。在 3、7、14 和 21 天时,测定循环中的ApoB蛋白水平。数据以 PBS 处理的对照动物中ApoB的比例表示。

In contrast to the effects of morpholino and piperidine modifications to the sense strand, when placed at the 5′ end of the antisense strand (siRNAs II, VI, VII, and VIII, Table 1), all modifications resulted in loss of activity compared to the parent siRNA (Fig. 3). The presence of a Mo or *Pip* derivative likely interferes with RISC-mediated gene silencing through two mechanisms. These modifications block 5′ phosphorylation and sterically interfere with effectively loading of the strand on Ago2.

相比于正义链的 morpholino 和哌啶修饰的效果,在反义链的 5'末端引入这些修饰(siRNA IIVIVIIVIII,表 1),所有修饰都导致了活性的丧失(图 3)。MoPip衍生物的存在可能通过两种机制干扰 RISC 介导的基因沉默。这些修饰阻碍了 5'磷酸化并在立体上干扰了链在 Ago2 上的有效加载。

图 3 Mo1Mo2对反义链的修饰抑制了沉默效果。小鼠(每组n=3 只)接受单剂量(3 mg/kg)的针对ApoB的 siRNA IIVIVIIVIII的治疗。在 3 天、7 天、14 天和 21 天时,测定了循环中的ApoB蛋白水平。数据以 PBS 处理的对照动物中ApoB的百分比表示。

To distinguish which of these modifications more effectively inhibit strand use, antisense strands targeting TTR were modified with the morpholino and piperidine analogues, and siRNAs (siRNAs IXXII, Table S3, ESI) were evaluated in a previously described in vitro luciferase reporter assay. In this assay, the 3′ UTR of the reporter gene that encodes luciferase contains a single binding site for the antisense strand. The siRNA X with the Mo2 modification was 30-fold less potent than parent (Fig. 4 and Table S3, ESI). This resulting antagonising effect was superior to that of Mo1, which was 13-fold less potent than the parent siRNA. Activities of siRNAs modified with Pip and Mo3 (XI and XII, respectively) were similar to that of the parent siRNA.

为了区分哪种修饰更有效地抑制链的使用,以靶向TTR的反义链为目标,使用 morpholino 和哌啶类似物对 siRNA 进行修饰,并在先前描述的体外荧光素酶报告基因分析中对 siRNA(siRNA IXXII,表 S3,附录)进行评估。在这个分析中,编码荧光素酶的报告基因的 3' UTR 包含反义链的单个结合位点。经修饰的 siRNA X(具有Mo2修饰)的效力比原始 siRNA 低 30 倍(图 4 和表 S3,附录)。这种结果的拮抗效应优于Mo1,其效力比原始 siRNA 低 13 倍。带有PipMo3修饰的 siRNA(分别为XIXII)的活性与原始 siRNA 相似。

图 4 反义链的Mo2修饰抑制了基因沉默活性。根据 siRNA 浓度,TTR报告分析中的荧光素酶表达百分比。针对TTR的 siRNA 的反义链经过指定的 morpholino 类似物修饰。原始链没有 5'修饰。

The residue at position 2 of the antisense strand strongly influences the stability of the complex with Ago2. Only the natural RNA, deoxy, or 2′-fluoro are tolerated at this position. For example, the methyl group of the commonly used 2′-OMe sugar modification results in a steric conflict with an α-helical curl of the Ago2 MID domain. Therefore, we reasoned that the RISC inhibiting ability of Mo2 would be enhanced if we replaced the 2′-fluoro at position 2 with 2′-OMe. In the reporter assay, the siRNA with an antisense strand containing these two was inactive even at the highest dose tested (siRNA XIII, Table S3 and Fig. S3, ESI). In the in vitro silencing assay, the modification of the sense strand with Mo2 and a 2′-OMe at position 2 resulted in an siRNA that was active even at the lowest dose (Table S4 and Fig. S4, ESI).

反义链第 2 位的残基强烈影响与 Ago2 结合复合物的稳定性。只有天然 RNA、去氧核苷酸或 2'-氟核苷酸在这个位置上被接受。例如,常用的 2'-OMe 糖修饰的甲基基团与 Ago2 MID 结构域的α螺旋卷曲存在立体冲突。因此,我们推测如果在第 2 位上用 2'-OMe 替换 2'-氟核苷酸,Mo2的 RISC 抑制能力将会增强。在报告基因检测中,含有这两个修饰的反义链的 siRNA 甚至在最高剂量下也无活性(siRNA XIII,表 S3 和图 S3,ESI)。在体外沉默检测中,将感应链修饰为Mo2和第 2 位的 2'-OMe,得到的 siRNA 即使在最低剂量下也具有活性(表 S4 和图 S4,ESI)。

To assess the impact of the modifications on the relative binding affinities to Ago2, the parent or morpholino-modified single strands were incubated with a commercially available recombinant human Ago2, and total RNA bound was quantified by stem-loop RT-qPCR. Significantly less oligonucleotide was loaded onto Ago2 when the 5′-position of the antisense strand was modified with Mo2 or, to a lesser extent Pip, than when the antisense strand was not modified with a morpholino or when the Mo1 modification was present (Fig. 5). The ON12 VP-modified oligonucleotide has the phosphate mimic (E)-vinyl phosphonate which is known to favor MID domain binding and was used as a positive control in this experiment. (See Table S5, ESI).

为评估这些修饰对与 Ago2 的相对结合亲和力的影响,我们将原始或经过 morpholino 修饰的单链与商业化的重组人类 Ago2 一起孵育,通过茎环 RT-qPCR 来定量结合的总 RNA。当反义链的 5'位置被Mo2或在较小程度上被Pip修饰时,加载到 Ago2 上的寡核苷酸显著减少,而当反义链未被 morpholino 修饰或存在Mo1修饰时,加载量较高(图 5)。ON12 VP 修饰的寡核苷酸含有磷酸酯类似物(E)-乙烯基膦酸酯,已知有利于与 MID 结构域的结合,在这个实验中被用作阳性对照(见表 S5,ESI)。

图 5 Mo2修饰在 5'位置抑制了 RISC 的加载。通过茎环 RT-PCR 定量与重组人类 Ago2 结合的反义链 RNA。图中显示的是三次重复实验的均值±标准偏差。

To rationalize the observation that Mo2 more effectively inhibited use of an siRNA strand by the RNAi silencing machinery than other modifications tested, we modeled complexes between Ago2 and antisense strands containing Mo1, Mo2, Pip, or Mo3 at their 5′-termini using the crystal structure of Ago2 bound to miR-20a (PDB ID 4f3t) as the starting structure. All models were built using UCSF Chimera and energy-minimized with Amber 14 as we did previously for modeling of the Mo1-modified strand bound to MID. Multiple basic side chains are gathered around the 5′ phosphate of the antisense strand inside the Ago2 MID pocket, and interactions of Mo2, Pip, and Mo3 docked to MID are fairly similar to those seen with Mo1 (Fig. 6 and Fig S5 in ESI). The Mo2 and Mo3 analogs, which are longer than Mo1, do not adopt a stretched orientation and thus are not inserted deeply into the binding pocket. These modifications are displaced by about 1 Å compared to the parent Mo1 (Fig. S5E in ESI). In the energy minimizations, we assigned a +1 positive charge to Mo1 and Mo2, but the corresponding nitrogen was neutral in Pip and Mo3. The only H-bond acceptors or donors on the morpholinos are a nitrogen lone pair or an N–H, respectively. The latter interaction is observed in the case of Mo2 with Tyr-529, which acts as an acceptor (Fig. S5F in ESI). The Mo ring oxygen Pip that is a rather weak acceptor is within H-bonding distance of Lys-570 (Fig. S5 in ESI). Interactions of the 5′ phosphate with all positively charged lysine and arginine residues as well Gln-545 are disrupted by the insertion of the 5′-Mo and -Pip modifications. Compared to Mo1, the slightly longer Mo2, Pip, and Mo3 modifications have even more unfavorable steric interactions with these basic side chains (Fig. S5E in ES). For example, Lys-533 has its NH3+ headgroup turned away and presents a methylene in the direction of Mo and Pip moieties in complexes with strands with the longer modifications (Fig. 6 and Fig. S5E in ESI). All morpholino groups are electrostatically incompatible with the MID domain binding site (Fig. S5 in the ESI). Thus, Mo2 incorporated at the 5′-end of the strand appears to be more disruptive than Mo1 as observed experimentally. Further, modeling indicates that the combination of Mo2 with 2′-OMe results in multiple steric clashes within the binding site for the antisense strand within Ago2 (Fig. 6C).

为了解释Mo2比其他修饰更有效地抑制 siRNA 链被 RNAi 沉默机制利用的观察结果,我们使用 Ago2 与 miR-20a 结合的晶体结构(PDB ID 4f3t)作为起始结构,建立了包含Mo1Mo2PipMo3在其 5'端的反义链与 Ago2 之间的复合物模型。所有模型都使用 UCSF Chimera 进行构建,并在 Amber 14 中进行了能量最小化,就像我们之前对与 MID 结合的Mo1修饰链进行建模一样。Ago2 MID 口袋内聚集了多个碱性侧链,与反义链的 5'磷酸发生相互作用,Mo2PipMo3与 MID 结合的相互作用与Mo1的相似(图 6 和 ESI 中的图 S5)。Mo2Mo3衍生物比Mo1更长,因此它们不会采用拉伸的方向,也不会深入结合口袋。这些修饰相对于父代Mo1而言,被推离约 1 埃(图 ESI 中的图 S5E)。在能量最小化中,我们为Mo1Mo2分配了+1 的正电荷,但相应的氮在PipMo3中是中性的。morpholinos 上的唯一 H 键受体或给体是氮孤对电子或 N–H。后者的相互作用在Mo2与 Tyr-529 之间被观察到(图 ESI 中的图 S5F)。Pip中的Mo环氧在较大程度上是一个较弱的受体,与 Lys-570 的 H 键距离处于 H 键的范围之内(ESI 中的图 S5)。5'磷酸与所有带正电的赖氨酸、精氨酸残基以及 Gln-545 的相互作用受到 5'-Mo和-Pip修饰的插入的破坏。与Mo1相比,Mo2PipMo3修饰稍长的修饰与这些碱性侧链之间的位阻作用更加不利(图 ESI 中的图 S5E)。例如,Lys-533 的 NH3+头基向外旋转,并向MoPip官能团的方向呈现出一个亚甲基,在较长修饰的链与这些碱性侧链的复合物中(图 6 和图 ESI 中的图 S5E)。所有 morpholino 基团在 MID 结合位点上与静止蛋白结构发生静电不相容性相互作用(ESI 中的图 S5)。因此,在链的 5'端引入Mo2似乎比实验观察到的Mo1更具破坏性。此外,建模表明,Mo2与 2'-OMe 的组合在 Ago2 内部反义链结合位点中产生多个位阻冲突(图 6C)。

图 6 Morpholino 类似物破坏 5'磷酸与 Ago2 的 MID 结构域的相互作用。Ago2 与含有(A)Mo1、(B)Mo2的链的模型,(C)与 Mo2 和 2'-OMe 在第一和第二位置(AS1 和 AS2)的复合物的模型。AS2 上的 2'-OMe 稍微穿透 Ago2 表面,并与 N562 发生碰撞(实心圆)。2'-OMe 以球棒模式显示,碳和氢原子分别以黄色和白色表示。AS3 上的 2'-F 显示为绿色球体。

In conclusion, three phosphoramidite building blocks were synthesized to allow incorporation of extended morpholino or piperidine functional groups at the 5′-position of oligonucleotides. In mice, in a reporter gene assay, and in an assay to monitor loading onto RISC, the Mo2 modification most effectively inhibited loading of an siRNA strand of the modifications tested. This extended morpholino derivative should mitigate previously described sense strand-mediated off-target effects and will be useful for studies of ascertaining the role of the antisense strand in downstream RNAi mediated side-effects if any, by blocking the antisense strand using this chemistry. These modifications are also expected to improve resistance to degradation by 5′-exonucleases. In this context, the combination of Mo2 with 2′-OMe at AS2 functions as the effective antisense blocker (Fig. 6C). When used in sense strands, the Mo2 modification should enhance potency and specificity by inhibiting use of this strand by blocking 5′-phosphorylation and by sterically hindering interaction with Ago2. In summary, the modifications evaluated here have the potential of improving the efficacy and safety of RNAi therapeutics and deserve further evaluation.

综上所述,我们合成了三种磷酸酰胺酯建构单元,用于在寡核苷酸的 5'位置引入扩展的 morpholino 或 piperidine 功能基团。在小鼠中,通过报告基因测定、RISC 载荷监测测定以及其他实验中,Mo2修饰最有效地抑制了 siRNA 链的载荷。这种扩展的 morpholino 衍生物应该能够减轻先前报道的由于 sense 链引起的非特异性效应,并可用于通过阻断 antisense 链的方式研究确定 antisense 链在下游 RNAi 介导的副作用中的作用。这些修饰还有望提高对 5'外切核酸酶的降解抗性。在这种情况下,Mo2与 AS2 上的 2'-OMe 的组合可作为有效的 antisense 阻断剂(图 6C)。当用于 sense 链时,Mo2修饰应通过阻断 5'-磷酸化和立体阻碍与 Ago2 的相互作用,增强效力和特异性。总之,我们评估的这些修饰具有提高 RNAi 治疗的疗效和安全性的潜力,并值得进一步评估。

Author contributions: The manuscript was written through contributions of all authors. All authors contributed to the design, execution of experiments, interpretation and writing of the manuscript and ESI.

作者贡献:本文的撰写是所有作者共同贡献的结果。所有作者对实验的设计、执行、结果解释和文稿及 ESI 的撰写都做出了贡献。

👀

   

Conflicts of interest 冲突利益声明

All the authors except ME are employees of Alnylam Pharmaceuticals.

除了 ME 以外的所有作者都是 Alnylam Pharmaceuticals 的雇员。


药悟
药相品悟。
 最新文章