深度突变扫描技术的发展及应用

doi:10.13560/j.cnki.biotech.bull.1985.2026-0659

摘要/Abstract

摘要：

解析蛋白质序列变异与结构和功能的关联是生命科学领域的关键科学问题。深度突变扫描（deep mutational scanning, DMS）技术是全景解析蛋白质突变效应的核心手段，为解决这一科学难题提供了关键技术支撑。该技术针对目标蛋白编码基因的全序列构建大规模突变文库，依托高通量筛选结合深度测序完成海量突变体的并行功能表征，最终量化突变效应并绘制蛋白质适应性景观。本文系统综述了DMS技术的前沿进展，重点对比了基于质粒和基因组原位编辑两类突变文库构建策略的原理、优缺点及适用范围；总结了生长富集筛选结合深度测序的高通量表型评估方法；梳理了其在蛋白质工程改造、微生物耐药突变鉴定、蛋白质语言模型测评与优化等领域的应用。同时针对文库构建与筛选体系的现存瓶颈，展望了新型基因组编辑工具、多元化表型筛选等有望推动技术发展的趋势。

关键词: 深度突变扫描, 突变库构建, 高通量筛选, 深度测序, 蛋白质适应性景观

Abstract:

Understanding how protein sequence variation influences protein structure and function remains a fundamental challenge in the life sciences. Deep mutational scanning (DMS) has emerged as a powerful high-throughput approach for systematically assessing the functional consequences of protein variants, providing critical technological support for addressing this challenge. DMS involves the construction of large-scale mutant libraries covering the entire coding sequence of a target protein. By coupling high-throughput screening with deep sequencing, DMS enables the parallel characterization of the functional effects of vast numbers of variants, thereby allowing quantitative assessment of mutational effects and the mapping of protein fitness landscapes. Here, we systematically review recent advances in DMS technology. First, we compare the principles, characteristics, and applicability of two major strategies for mutant library construction: plasmid-based approaches and in situ genome-editing approaches. We further summarize high-throughput phenotypic evaluation methods based on growth-enrichment screening coupled with deep sequencing. In addition, we highlight the applications of DMS in protein engineering, identification of mutations conferring antimicrobial resistance, and benchmarking and optimization of protein language models. Finally, we discuss current bottlenecks in library construction and screening systems, and provide perspectives on future developments, including novel genome-editing tools and diversified phenotypic screening methods.

Key words: deep mutational scanning, mutant library construction, high-throughput screening, deep sequencing, protein fitness landscape

孙冠男, 魏艺芸, 何文涛, 刘娇, 郑平. 深度突变扫描技术的发展及应用[J]. 生物技术通报, doi: 10.13560/j.cnki.biotech.bull.1985.2026-0659.

SUN Guan-nan, WEI Yi-yun, HE Wen-tao, LIU Jiao, ZHENG Ping. Development and Applications of Deep Mutational Scanning[J]. Biotechnology Bulletin, doi: 10.13560/j.cnki.biotech.bull.1985.2026-0659.

图/表 6

图1 深度突变扫描流程工作流程包括大规模突变体库的构建与突变文库并行功能评估两部分，其中大规模突变体库的构建分为质粒建库以及基因组原位建库；突变文库并行功能评估包括基于生长的高通量筛选、深度测序以及适应性景观绘制

Fig. 1 Deep mutational scanning workflowThe workflow consists of two main modules: construction of a large-scale mutant library and parallel fitness evaluation of numerous mutants. Large-scale mutant library construction is further divided into plasmid-based library construction and in situ genomic library construction. Parallel fitness evaluation of numerous mutants includes growth-based high-throughput screening, deep sequencing, and fitness landscape profiling

表1 基于质粒构建突变扫描文库汇总

Table 1 Summary of plasmid-based methods for deep mutational scanning library construction

方法 Methods	蛋白 Protein	长度 Length （aa）	突变方式 Mutant types	覆盖度 Coverage（%）	文献 Reference
反向PCR技术 Inverse PCR	CcdB	101	NNK	95.0	［19］
	PTEN	122	NNK	83.6	［5］
	Src	158	NNK	87.7	［20］
	TPMT	187	NNK	85.4	［5］
	Ras	165	NNK	90.0	［21］
	BRCA1	237	NNK	96.5	［22］
盒式连接技术 Cassette ligation	Hsp90	231	NNK	98.0	［23］
	APH3Ⅱ	264	NNK	100.0	［24］
	ERK2	359	NNK	94.8	［25］
	Hp53	393	NNK	99.8	［15］
	Rubisco	462	NNK	99.1	［8］
	SHP2	593	NNK	80	［14］
切刻突变技术 Nicking mutagenesis	AmiE	71	NNK	100.0	［3］
	PYR1	86	NNK	99	［3］
	UCA9	99	NNK	97.4	［17］
	ΦX174G	172	NNK	99	［26］
	KRAS	188	NNK	99.5	［27］
	TnpB	408	NNK	95.7	［28］
	ΦX174F	421	NNK	99	［26］

图2 质粒层面DMS文库构建技术A：反向PCR技术；B：盒式连接技术；C：切刻突变技术

Fig. 2 DMS library construction technology based on plasmidA: Inverse PCR; B: cassette ligation; C: nicking mutagenesis

表2 基于基因组突变扫描文库构建汇总

Table 2 Summary of genome methods for constructing deep mutational scanning libraries

方法 Methods	蛋白 Protein	长度 Length （aa）	突变方式 Mutant types	覆盖度 Coverage （%）	文献 Reference
MAGE	InfA	71	NNN	99.0	［29］
	RpoB	62	NNK	98.0	［10］
	RpoD	234	NNN	99.0	［30］
	UgtP	391	NNK	-^#	［11］
CRISPR	FolA	157	NNK	97.0	［31］
	FabZ	151	20^*	84.0	［32］
	LpxC	305	20^*	84.0	［32］
	MurA	419	20^*	76.0	［32］

图3 基于基因组DMS文库构建技术

Fig. 3 Genome-level DMS library construction technology

图4 深度突变扫描应用研究A：蛋白工程改造；B：耐药突变位点鉴定；C：蛋白质语言模型测评与优化

Fig. 4 Application of DMSA: Protein engineering modification. B: Identification of drug resistance mutations. C: Evaluation and optimization of protein language model

参考文献 60

[1]	Fowler DM, Fields S. Deep mutational scanning: a new style of protein science [J]. Nat Methods, 2014, 11(8): 801-807.
[2]	Araya CL, Fowler DM. Deep mutational scanning: assessing protein function on a massive scale [J]. Trends Biotechnol, 2011, 29(9): 435-442.
[3]	Wrenbeck EE, Klesmith JR, Stapleton JA, et al. Plasmid-based one-pot saturation mutagenesis [J]. Nat Methods, 2016, 13(11): 928-930.
[4]	Kitzman JO, Starita LM, Lo RS, et al. Massively parallel single-amino-acid mutagenesis (vol 12, pg 203, 2015) [J]. Nat Methods, 2017, 14(5): 540-1.
[5]	Matreyek KA, Starita LM, Stephany JJ, et al. Multiplex assessment of protein variant abundance by massively parallel sequencing [J]. Nat Genet, 2018, 50(6): 874-882.
[6]	Kemble H, Nghe P, Tenaillon O. Recent insights into the genotype-phenotype relationship from massively parallel genetic assays [J]. Evol Appl, 2019, 12(9): 1721-1742.
[7]	Kuiper BP, Prins RC, Billerbeck S. Oligo pools as an affordable source of synthetic DNA for cost-effective library construction in protein- and metabolic pathway engineering [J]. ChemBioChem, 2022, 23(7): e202100507.
[8]	Prywes N, Phillips NR, Oltrogge LM, et al. A map of the rubisco biochemical landscape [J]. Nature, 2025, 638(8051): 823-828.
[9]	Küng C, Protsenko O, Vanella R, et al. Deep mutational scanning reveals a de novo disulfide bond and combinatorial mutations for engineering thermostable myoglobin [J]. Protein Sci, 2025, 34(5): e70112.
[10]	Yang KB, Cameranesi M, Gowder M, et al. High-resolution landscape of an antibiotic binding site [J]. Nature, 2023, 622(7981): 180-187.
[11]	Filsinger GT, Mychack A, et al. A diverse single-stranded DNA-annealing protein library enables efficient genome editing across bacterial Phyla [J]. Proc Natl Acad Sci U S A, 2025, 122(17): e2414342122.
[12]	Gelman S, Fahlberg SA, Heinzelman P, et al. Neural networks to learn protein sequence-function relationships from deep mutational scanning data [J]. Proc Natl Acad Sci U S A, 2021, 118(48): e2104878118.
[13]	Wu ZX, Cai XJ, Zhang X, et al. Expression level is a major modifier of the fitness landscape of a protein coding gene [J]. Nat Ecol Evol, 2022, 6(1): 103-115.
[14]	Jiang ZY, van Vlimmeren AE, Karandur D, et al. Deep mutational scanning of the multi-domain phosphatase SHP2 reveals mechanisms of regulation and pathogenicity [J]. Nat Commun, 2025, 16: 5464.
[15]	Giacomelli AO, Yang XP, Lintner RE, et al. Mutational processes shape the landscape of TP53 mutations in human cancer [J]. Nat Genet, 2018, 50(10): 1381-1387.
[16]	Cozens C, Pinheiro VB. Darwin Assembly: fast, efficient, multi-site bespoke mutagenesis [J]. Nucleic Acids Res, 2018, 46(8): e51.
[17]	Medina-Cucurella AV, Steiner PJ, Faber MS, et al. User-defined single pot mutagenesis using unamplified oligo pools [J]. Protein Eng Des Sel, 2019, 32(1): 41-45.
[18]	Kirby MB, Medina-Cucurella AV, Baumer ZT, et al. Optimization of multi-site nicking mutagenesis for generation of large, user-defined combinatorial libraries [J]. Protein Eng Des Sel, 2021, 34: gzab017.
[19]	Jain PC, Varadarajan R. A rapid, efficient, and economical inverse polymerase chain reaction-based method for generating a site saturation mutant library [J]. Anal Biochem, 2014, 449: 90-98.
[20]	Ahler E, Register AC, Chakraborty S, et al. A combined approach reveals a regulatory mechanism coupling src's kinase activity, localization, and phosphotransferase-independent functions [J]. Mol Cell, 2019, 74(2): 393-408.e20.
[21]	Bandaru P, Kuriyan J, Ranganathan R. Deconstruction of the Ras switching cycle through saturation mutagenesis reveals hot-spots of allosteric activation [J]. Protein Sci, 2017, 26: 79.
[22]	Starita LM, Islam MM, Banerjee T, et al. A multiplex homology-directed DNA repair assay reveals the impact of more than 1, 000 BRCA1 missense substitution variants on protein function [J]. Am J Hum Genet, 2018, 103(4): 498-508.
[23]	Mishra P, Flynn JM, Starr TN, et al. Systematic mutant analyses elucidate general and client-specific aspects of Hsp90 function [J]. Cell Rep, 2016, 15(3): 588-598.
[24]	Melnikov A, Rogov P, Wang L, et al. Comprehensive mutational scanning of a kinase in vivo reveals substrate-dependent fitness landscapes [J]. Nucleic Acids Res, 2014, 42(14): e112.
[25]	Brenan L, Andreev A, Cohen O, et al. Phenotypic characterization of a comprehensive set of MAPK1/ERK2 missense mutants [J]. Cell Rep, 2016, 17(4): 1171-1183.
[26]	Faber MS, Van Leuven JT, Ederer MM, et al. Saturation mutagenesis genome engineering of infective ΦX174 Bacteriophage viaUnamplified oligo pools and golden gate assembly [J]. ACS Synth Biol, 2020, 9(1): 125-131.
[27]	Weng CC, Faure AJ, Escobedo A, et al. The energetic and allosteric landscape for KRAS inhibition [J]. Nature, 2024, 626(7999): 643-652.
[28]	Thornton BW, Weissman RF, Rodriguez JE, et al. Engineered TnpB genome editors for plants and human cells identified by ribonucleoprotein mutational scanning [J]. Nat Biotechnol, 2026.
[29]	Kelsic ED, Chung H, Cohen N, et al. RNA structural determinants of optimal codons revealed by MAGE-seq [J]. Cell Syst, 2016, 3(6): 563-571.e6.
[30]	Park J, Wang HH. Systematic dissection of σ70 sequence diversity and function in bacteria [J]. Cell Rep, 2021, 36(8): 109590.
[31]	Garst AD, Bassalo MC, Pines G, et al. Genome-wide mapping of mutations at single-nucleotide resolution for protein, metabolic and genome engineering [J]. Nat Biotechnol, 2017, 35(1): 48-55.
[32]	Dewachter L, Brooks AN, Noon K, et al. Deep mutational scanning of essential bacterial proteins can guide antibiotic development [J]. Nat Commun, 2023, 14: 241.
[33]	Wang HH, Kim H, Cong L, et al. Genome-scale promoter engineering by coselection MAGE [J]. Nat Methods, 2012, 9(6): 591-593.
[34]	Wang HH, Church GM. Multiplexed genome engineering and genotyping methods [M]//Synthetic Biology, Part B - Computer Aided Design and DNA Assembly. Amsterdam: Elsevier 2011: 409-426.
[35]	Wannier TM, Ciaccia PN, Ellington AD, et al. Recombineering and mage [J]. Nat Rev Meth Primers, 2021, 1: 7.
[36]	Zwart MP, Schenk MF, Hwang S, et al. Unraveling the causes of adaptive benefits of synonymous mutations in TEM-1 β-lactamase [J]. Heredity, 2018, 121(5): 406-421.
[37]	Wei HJ, Li XH. Deep mutational scanning: a versatile tool in systematically mapping genotypes to phenotypes [J]. Front Genet, 2023, 14: 1087267.
[38]	Dietrich JA, McKee AE, Keasling JD. High-throughput metabolic engineering: advances in small-molecule screening and selection [J]. Annu Rev Biochem, 2010, 79: 563-590.
[39]	Maes S, Deploey N, Peelman F, et al. Deep mutational scanning of proteins in mammalian cells [J]. Cell Rep Meth, 2023, 3(11): 100641.
[40]	Liu Y, Wang RY, Liu JH, et al. Base editor enables rational genome-scale functional screening for enhanced industrial phenotypes in Corynebacterium glutamicum [J]. Sci Adv, 2022, 8(35): eabq2157.
[41]	Rubin AF, Gelman H, Lucas N, et al. A statistical framework for analyzing deep mutational scanning data [J]. Genome Biol, 2017, 18: 150.
[42]	Faure AJ, Schmiedel JM, Baeza-Centurion P, et al. DiMSum: an error model and pipeline for analyzing deep mutational scanning data and diagnosing common experimental pathologies [J]. Genome Biol, 2020, 21: 207.
[43]	Stein A, Fowler DM, Hartmann-Petersen R, et al. Biophysical and mechanistic models for disease-causing protein variants [J]. Trends Biochem Sci, 2019, 44(7): 575-588.
[44]	Zhang N, Chen YT, Lu HY, et al. MutaBind2: predicting the impacts of single and multiple mutations on protein-protein interactions [J]. iScience, 2020, 23(3): 100939.
[45]	Park Y, Metzger BPH, Thornton JW. Epistatic drift causes gradual decay of predictability in protein evolution [J]. Science, 2022, 376(6595): 823-830.
[46]	Cao LX, Coventry B, Goreshnik I, et al. Design of protein-binding proteins from the target structure alone [J]. Nature, 2022, 605(7910): 551-560.
[47]	Faure AJ, Domingo J, Schmiedel JM, et al. Mapping the energetic and allosteric landscapes of protein binding domains [J]. Nature, 2022, 604(7904): 175-183.
[48]	Luo YN, Jiang GD, Yu TH, et al. ECNet is an evolutionary context-integrated deep learning framework for protein engineering [J]. Nat Commun, 2021, 12: 5743.
[49]	Brandes N, Goldman G, Wang CH, et al. Genome-wide prediction of disease variant effects with a deep protein language model [J]. Nat Genet, 2023, 55(9): 1512-1522.
[50]	Wrenbeck EE, Azouz LR, Whitehead TA. Single-mutation fitness landscapes for an enzyme on multiple substrates reveal specificity is globally encoded [J]. Nat Commun, 2017, 8: 15695.
[51]	Bédard C, Gagnon-Arsenault I, Boisvert J, et al. Most azole resistance mutations in the Candida albicans drug target confer cross-resistance without intrinsic fitness cost [J]. Nat Microbiol, 2024, 9(11): 3025-3040.
[52]	Lin ZM, Akin H, Rao R, et al. Evolutionary-scale prediction of atomic-level protein structure with a language model [J]. Science, 2023, 379(6637): 1123-1130.
[53]	Jiang KY, Yan ZQ, Di Bernardo M, et al. Rapid in silico directed evolution by a protein language model with EVOLVEpro [J]. Science, 2025, 387(6732): eadr6006.
[54]	Tran VQ, Nemeth M, Bartie LJ, et al. Rapid directed evolution guided by protein language models and epistatic interactions [J]. Science, 2026, 392(6798): eaea1820.
[55]	Simon AJ, Ellington AD, Finkelstein IJ. Retrons and their applications in genome engineering [J]. Nucleic Acids Res, 2019, 47(21): 11007-11019.
[56]	Sharon E, Chen SA, Khosla NM, et al. Functional genetic variants revealed by massively parallel precise genome editing [J]. Cell, 2018, 175(2): 544-557.e16.
[57]	Schubert MG, Goodman DB, Wannier TM, et al. High-throughput functional variant screens via in vivo production of single-stranded DNA [J]. Proc Natl Acad Sci U S A, 2021, 118(18): e2018181118.
[58]	Weinstein JY, Martí-Gómez C, Lipsh-Sokolik R, et al. Designed active-site library reveals thousands of functional GFP variants [J]. Nat Commun, 2023, 14: 2890.
[59]	Binder S, Schendzielorz G, Stäbler N, et al. A high-throughput approach to identify genomic variants of bacterial metabolite producers at the single-cell level [J]. Genome Biol, 2012, 13(5): R40.
[60]	Della Corte D, van Beek HL, Syberg F, et al. Engineering and application of a biosensor with focused ligand specificity [J]. Nat Commun, 2020, 11: 4851.