10506–10519 Nucleic Acids Research, 2019, Vol. 47, No. 20 Published online 4 October 2019 doi: 10.1093/nar/gkz831 SURVEY AND SUMMARY Synonymous genome recoding: a tool to explore microbial biology and new therapeutic strategies Miguel Angel Martı́nez* , Ana Jordan-Paiz, Sandra Franco and Maria Nevot IrsiCaixa, Hospital Universitari Germans Trias i Pujol, Universitat Autònoma de Barcelona (UAB), Badalona, Spain Received August 13, 2019; Revised September 12, 2019; Editorial Decision September 16, 2019; Accepted September 30, 2019 ABSTRACT and the fact that there are four different nucleotides, only 43 = 64 different codons are possible. Of these possible 64 Synthetic genome recoding is a new means of gen- codons, 61 are sense codons, coding for 20 amino acids erating designed organisms with altered pheno- and 3 are stop codons, coding for termination of transla- types. Synonymous mutations introduced into the tion. Eighteen of these amino acids are encoded by more protein coding region tolerate modifications in DNA than one synonymous codon. Consequently, a given amino or mRNA without modifying the encoded proteins. acid might be encoded by one, two, three, four or six dif- Synonymous genome-wide recoding has allowed ferent codons (Figure 1). Importantly, the observed ratios the synthetic generation of different small-genome of synonymous codons are highly non-random (1). This viruses with modified phenotypes and biological phenomenon is termed codon usage bias. Foreign protein properties. Recently, a decreased cost of chemically expression in Escherichia coli is well-known to be influ- synthesizing DNA and improved methods for assem- enced significantly by the presence or absence of uncommon codons, and heterologous protein expression often requires bling DNA fragments (e.g. lambda red recombination coding sequence modification (2). and CRISPR-based editing) have enabled the con- A suggested model of codon usage bias is that abun- struction of an Escherichia coli variant with a 4-Mb dant codons correlate with the abundances of isoaccept- synthetic synonymously recoded genome with a re- ing tRNAs, which correlate with protein production levels duced number of sense codons (n = 59) encoding (3). However, tRNA availability is not the sole determinant the 20 canonical amino acids. Synonymous genome of protein production (4). Synonymous codon mutations recoding is increasing our knowledge of microbial in- can impact many molecular phenotypes, including tran- teractions with innate immune responses, identifying scription modifications (5,6), translation initiation (7–9), functional genome structures, and strategically ame- translation elongation (10,11), translation accuracy (12,13), liorating cis-inhibitory signaling sequences related RNA stability (14), RNA structure and folding (8,15), RNA to splicing, replication (in eukaryotes), and complex splicing (16,17), RNA toxicity (18) and cotranslational fold- ing (23). In the case of eukaryotes, synonymous substi- microbe functions, unraveling the relevance of codon tutions can affect chromatin organization (19), enhancer usage for the temporal regulation of gene expression functions (20,21) and microRNA targeting (20,22). Due to and the microbe mutant spectrum and adaptability. the impact of synonymous mutations on molecular pheno- New biotechnological and therapeutic applications types, it is not difficult to assume that they are under se- of this methodology can easily be envisaged. In this lective pressure. Moreover, natural selection has generated review, we discuss how synonymous genome recod- unequal codon usage across genes and genomes (24,25). ing may impact our knowledge of microbial biology Accordingly, synonymous nucleotide positions have been and the development of new and better therapeutic modelled by genome-wide mutational processes during evo- methodologies. lution (26,27). Synonymous codon usage is neither random nor neutral (28–30). Organisms and different genes from the same organism include specific codons at different fre- SYNONYMOUS MUTATIONS AND GENOME CODING quencies (31). Human codon frequencies for synonymous The genetic code is common to all organisms and exten- codons are not equal (Figure 1). Synonymous codon fre- sively redundant. Based on the nature of triplet encoding quencies varies among different taxa, with codon usage in * To whom correspondence should be addressed. Tel: +34 934656374; Fax: +34 934653968; Email: mmartinez@irsicaixa.es C The Author(s) 2019. Published by Oxford University Press on behalf of Nucleic Acids Research. This is an Open Access article distributed under the terms of the Creative Commons Attribution Non-Commercial License (http://creativecommons.org/licenses/by-nc/4.0/), which permits non-commercial re-use, distribution, and reproduction in any medium, provided the original work is properly cited. For commercial re-use, please contact journals.permissions@oup.com Nucleic Acids Research, 2019, Vol. 47, No. 20 10507 Figure 1. The standard genetic code and codon bias. Each amino acid is encoded by multiple codons, except methionine and tryptophan. On the right are the human frequencies per thousand for each codon. The Kazusa database (http://www.kazusa.or.jp/codon/) was used to compile human codon usage. Codons carrying CpG and UpA dinucleotides are underrepresented. mammals differing from that in bacteria or insects. Synony- sented, as it is used only one-seventh as often as GCAGAG. mous codons may also play distinct roles at different sites Analysis of an extended data set of 35 770 human mRNA in the genome, and epistatic interactions may occur among sequences had similar findings (37). Whether codon-pair us- codons within and between genes (2). Although this review age bias is shaped by selection is controversial. Looking mainly deals with prokaryotes and viruses, it is worthy to at the underrepresented codon-pair GCCGAA, we easily mention that genome-wide association studies and ulterior notice that a CpG dinucleotide is generated between the functional analyses have recently shown that synonymous third base of the first codon and the first base of the ac- substitutions can impact complex human diseases, such as companying codon. Many of the underrepresented human cancer (32) and autism (33). codon-pairs generate CpG dinucleotides (38). CpG dinu- Synonymous genome recoding alters not only codon us- cleotides are less frequent than expected in both humans age, but also codon-pair usage, and dinucleotide frequency and other mammals (Figure 1) (39). The mechanistic rea- (e.g. CpG/UpA) (reviewed in (34)). Recently, it has been de- sons for this underrepresentation are unclear. Compared scribed that efficient mRNA translation is also determined to other dinucleotides, the mutation rate of the CpG mo- by a triplet-of-triplet genetic code, that is, translation of tif is at least 10-fold higher, especially in higher primates a given codon is influenced by the two preceding codons (40), likely due to cytosine methylation resulting in deam- and that given codon (56). In addition to codon usage bias, ination to thymine (40). Moreover, 5-methylcytosine in the which has been thoroughly studied in many species, codon- context of the CpG sequence is an epigenetic mechanism as- pair frequencies (termed codon-pair bias) are different in sociated with transcriptional repression (41,42). These find- different organisms. In any given genome, different frequen- ings strongly suggest that it is very difficult to distinguish be- cies of codon-pairs are found than would be expected based tween codon-pair usage bias and dinucleotide frequencies. on the individual codon usage bias of that genome (re- Because codon-pair bias and dinucleotide frequency are viewed in (31)). Codon-pair usage bias was first described inevitably connected, whether one is generating the other in bacteria (35), and then in all studied organisms (30). An is still in question (37,38,43,44). Interestingly, most mam- analysis of 14 795 human genes showed that some synony- malian RNA viruses have low frequencies of CpGs (45,46). mous codon-pairs are used more or less frequently than ex- Furthermore, viruses with high CpG frequencies may be pected (36). For example, based on the codon frequencies, more recognizable by pathogen innate immune sensors (47– the amino acid pair Ala–Glu would be expected to be en- 50). A broad spectrum antiviral protein as the zinc-finger coded equally as frequently by GCCGAA and GCAGAG. antiviral protein (ZAP) has been shown to inhibit HIV- However, even though the codon-pair GCCGAA contains 1 production by cells infected with CpG-enriched virus the most frequent Ala codon, it is strongly underrepre- genomes (42). Nevertheless, there are other hypotheses to 10508 Nucleic Acids Research, 2019, Vol. 47, No. 20 Table 1. Methods for large-scale genome synonymous recoding SYNTHETIC GENOMICS Method References The ability to read (sequencing) and write (synthesis) DNA Codon bias usage (68,72,74,77,79,83,85,152) sequences has greatly increased our capability to under- Codon-pair usage (64,69,70,71,73,75, stand and manipulate biological systems. Moreover, DNA 78,80,82,84,88,89,153) sequencing and synthesis are the fundamental technologies Dinucleotide frequency (37,86,130) in the development of synthetic genomics, which aims to de- (e.g. CpG/UpA) Codon-triplet bias usage (56) termine the fundamentals and design of a genome and engi- Codons that could generate stop (136) neer novel biological functions. A genome is defined as syn- mutations after a single nucleotide thetic when all building blocks of the final DNA molecule substitution are generated by chemical synthesis (90,91). Synthetic tech- niques for designing and manufacturing DNA offer the promise of completely controlling an organism’s genetic in- formation (92). In the last decade, sequencing technologies have evolved explain the codon-pair bias. In particular, work with E. coli considerably. Following the development of several novel has proposed that the size of the tRNA variable loop de- DNA sequencing methods in the late 1990s and early 2000s termines whether the 3 nucleotide adjacent to the codon (93,94), different technologies have become commercially has a context effect (51), but others have claimed that the available for so-called next-generation sequencing (NGS). P-site wobble position within the codon–anticodon inter- NGS in all branches of the tree of life has increased our action, and the A-site anticodon loop and acceptor stem capability to understand biological systems and the interre- have a decisive effect on the observed genomic codon-pair lated genetic nature of different organisms. Knowledge re- patterns (52). More recent studies in yeast have shown that garding the genetic diversity of microorganisms, the micro- the wobble base pairing plays a critical role in whether a biome, and human disease has also increased due to NGS. codon-pair has an inhibitory effect on translation (53). This Seven years ago, we started to sequence >15 petabases per work is in agreement with former studies, which would sup- year (95). This wealth of sequence data also enhanced the port the role of codon-pair usage bias. In addition, numer- discipline of synthetic biology and genomics. In order to ous studies have offered evidence of the effect of codon- better understand DNA sequence information, we need to pair usage bias on translational fidelity and rate (29,54,55). manipulate it and create new synthetic DNA reads. In con- Both codon-pair usage bias and dinucleotide content have trast to the rapid development of DNA sequencing tech- been incorporated into algorithms for gene optimization niques, the capability to synthesize DNA has not increased or deoptimization (Table 1) (31). Furthermore, work per- at the same pace (90). Writing is much more difficult than formed with the Salmonella flgM gene, which coordinates reading. flagellar gene expression and flagellum assembly, has shown Synthetic genomics starts with short synthetic single- that mRNA translation is determined by a triplet-of-triplet stranded DNA fragments called oligonucleotides. Develop- genetic code (56). That is, and similar to codon-pair us- ments in the 1980s in polymerase chain reaction (PCR) al- age, the efficiency of translating a particular codon is in- lowed the rapid assembly of overlapping oligonucleotides. fluenced by the sequence of the immediately adjacent flank- Kary Mullis and Michael Smith were awarded with the No- ing codons. A model explains these codon-context effects bel Prize in Chemistry in 1993 for the invention of PCR and by suggesting that codon recognition by elongation factor- the fundamental application of synthetic oligonucleotides bound aminoacyl-tRNA is initiated by hydrogen bond in- in DNA manipulation. In addition to other applications in teractions between the first two nucleotides of the codon molecular biology, the invention of PCR and the chemical and anticodon, and then stabilized by base-stacking energy synthesis of DNA oligonucleotides opened the door >30 over three successive codons (54). These data suggest that years ago to synthetic genomics. the context of the genome sequence should consider codon DNA chemical synthesis allows the synthesis of oligonu- triplets in addition to codon-pairs and highlight the chal- cleotides no more than 200 nucleotides in length (reviewed lenge of designing genes for maximal expression, whether in (90)). This is an intrinsic length limit based on phos- by natural selection or in the laboratory. phoramidite chemistry coupling efficiency. Yields of full- These studies illustrate the possible fitness costs of syn- length oligonucleotides can be affected by spurious depuri- onymous genome recoding, which can be under strong se- nation of the synthesized oligonucleotide during the acid lection. As suggested previously, their importance in evo- deprotection steps of the synthesis cycle and becomes es- lution may be under-appreciated (57). Synonymous mu- pecially problematic for longer oligonucleotide sequences. tations influence fitness in bacteria (8,11,18,57–67) and The cost of column-based oligonucleotide synthesis has not viruses (36,68–89). In addition, synonymous rewriting of decreased substantially in the last decade. Custom com- protein-coding regions allows the encoded amino acid se- mercial oligonucleotide prices range from $0.05 to $0.50 quences to be retained, but it may remove other genetic per base depending on the scale of synthesis, the length of information, including hidden control elements embedded the oligonucleotide, and the supplier. Moreover, in contrast within amino acid coding regions (56). We and others hy- to custom pricing for sequencing, oligonucleotide prices pothesize that synonymous genome recoding offers an op- have increased in the last few years. Alternatively, the prices portunity to manipulate the microbial replication capacity have decreased for DNA synthesis via microarray-based and biology. oligonucleotide synthesis. Compared to column-based tech- Nucleic Acids Research, 2019, Vol. 47, No. 20 10509 niques, array oligonucleotide synthesis can be consider- scribed RNA generated from 7.5 kb of synthetic poliovirus ably cheaper. The problem with making oligonucleotides DNA representing its genomic sequence yielded infectious from these array platforms for gene synthesis applications is virus particles after transfection of HeLa cell-free extract. that they produce many more synthesis-related errors than One year later, Smith and colleagues were able to gen- column-synthesized oligonucleotides (96,97). Minimizing erate the complete infectious genome of bacteriophage synthesis errors is crucial for large-scale genome synthesis. X174 (5386 b) from a single pool of chemically synthesized One wrong base out of >1 million in an essential gene ren- oligonucleotides in 14 days (104). Recently, additional vi- ders the genome of Mycoplasma mycoides (1.1 Mb) inactive ral RNA and DNA genomes were synthesized, up to a size (98). Both the length and price of oligonucleotide synthesis of 212 kb with the DNA genome of horsepox virus (105). remain the main technical challenges to speeding up syn- Horsepox virus is a DNA virus belonging to the genus Or- thetic genomics development. thopoxvirus and is closely related to a group of viruses in- Overlapping single-stranded DNA oligonucleotides 100– cluding variola virus, the etiological agent of smallpox, and 200 nucleotides in length are designed to facilitate the as- vaccinia virus, the vaccine virus used to eradicate smallpox. sembly and generation of double-stranded DNA. Several Variola virus no longer exists in nature, and the last nat- oligonucleotides can easily be assembled by PCR or isother- urally occurring case of smallpox was in 1977. Synthetic mal amplification to generate 1000–2000 base pair (bp) horsepox virus was generated by purchasing DNA frag- DNA fragments (Figure 2). Using conventional cloning ments of ∼30 kb each that collectively represent the entire methods, these synthetic DNA fragments can be cloned 212 kb genome with some overlap between the fragments. into a bacterial plasmid vector and individual clones iso- The most time-consuming part of the project was construc- lated and sequenced by Sanger sequencing. Importantly, tion of the requested DNA fragments by the DNA synthesis both chemical DNA synthesis and PCR are error prone. company (105). The probability of any given sequence containing a muta- The synthetic 30 kb DNA fragments of horsepox virus tion increases with the length of the DNA, which means that were co-transfected into cells already infected with Shope more clones have to be sequenced to obtain a completely fibroma virus, which served as a helper virus to provide the correct DNA clone. Five to six overlapping PCR fragments horsepox virus with the viral proteins it needed to conduct 1000–2 000bp long can be assembled and cloned in a plas- DNA transcription and replication (106). However, we can- mid vector using the Gibson protocol to obtain up to 10– not utilize this methodology to construct larger genomes 15 kb DNA fragments (99). Via PCR recombination and of other microorganisms, such as bacteria. An approach the isothermal cloning reaction, it has been shown to be taken by the Venter group was to reduce the genome of possible to synthesize the entire 16.3 kb mouse mitochon- M. mycoides by approximately 50% (107). Three rounds drial genome from 600 overlapping 60-nucleotide synthetic of design, synthesis, and testing were necessary to pro- oligonucleotides (100). Cloned fragments of this size can be duce JCVI-syn3.0 (531 kb), which has a genome smaller bought from commercial vendors at $0.10 to $0.50 per bp than that of any autonomously replicating cell found in na- depending on the fragment size and cloning vector. Com- ture. Gene-deletion methods allowed reduction of the E. paring the prices for custom oligonucleotides and gene syn- coli K-12 genome by 15% (108). Using this reduced E. coli thesis, it may be worthwhile to use commercial DNA frag- genome, Wang, Chin and colleagues developed a method to ment synthesis platforms. The Gibson protocol is proba- replace genomic DNA with long (∼100 kb) synthetic DNA bly the most commonly used method for assembling mul- through the in vivo excision of double-stranded DNA from tiple pieces of DNA into larger constructs. This method an episomal replicon by CRISPR/Cas9, coupled to lambda uses a one-pot isothermal technique, which utilizes an en- red-mediated recombination and simultaneous positive and zyme mixture containing a thermostable DNA polymerase, negative selection (Figure 3) (109). Using this approach, this DNA ligase and exonuclease to chew-back, anneal, and re- group was the first to complete a 4-Mb synthetic version of pair adjacent overlapping DNA sequences in order to as- the E. coli genome (65). Synthetic DNA fragments approxi- semble the required DNA construct. More recent work has mately 10 kb in length were purchased from a DNA synthe- updated and improved this method to further expand it sis company. Between 9 and 14 of these synthetic fragments (101). Alternatively, the introduction of 1557 synonymous were assembled by homologous recombination in Saccha- TTA/TTG to CTA/CTG leucine codon changes across romyces cerevisiae using bacterial artificial chromosomes 200 kb of Salmonella typhimurium genomic DNA was per- (BACs) to generate 37 fragments 91–136 kb in length. These formed using stepwise integration of rolling circle amplified 37 BAC fragments contained Cas9 cleavage sites that al- segments (SIRCAS) (102). Genome recoding by SIRCAS lowed its introduction into the previously reduced E. coli involves iterative recombination of recoded DNA segments K-12 genome. containing a selection marker (typically kanamycin or chlo- The declared cost of the synthesis of the 212 kb genome ramphenicol resistance cassettes). After each step, one se- of horsepox virus was $100 000 (105). An international con- lection marker is gained while the other is lost, providing a sortium of 21 different institutions is synthetically recon- readout for successful recoding. structing all 16 chromosomes of S. cerevisiae. The consor- The first full-genome synthetic reconstructions were per- tium has already described the synthesis of artificial yeast formed with RNA viruses because of their smaller genomes chromosome synIII (273 kb) (110). Successively, five addi- (∼10 kb). RNA virus genomes can be constructed easily tional chromosomes have been reported (111–115). Similar and reasonably inexpensively using the method described to E. coli, the redesigned yeast chromosomes had repetitive above (Figure 2). In 2002, the first complete synthesis of sequences (tRNA genes, introns and transposons) removed a viral RNA genome was performed (103); in vitro tran- to allow genome reduction. In contrast, CRISPR has not 10510 Nucleic Acids Research, 2019, Vol. 47, No. 20 Figure 2. Synthetic DNA oligonucleotide assembly and cloning. Overlapping oligonucleotides encoding a DNA duplex are successively assembled via polymerase chain reaction (PCR) or isothermal amplification. Designed DNA duplex are now cloned in yeast or bacterial plasmid that after E. coli transformation can be profusely amplified, sequenced and linearized for further cloning. been used for progressive genome reduction. At the be- 57-codon E. coli genome is also under way (119). NcAAs ginning of this century, the first individual human genome are those not naturally encoded or found in the genetic was sequenced (116,117). The Human Genome Project was code of any organism. Reassigned tRNAs are being used time-consuming and very expensive, but it accelerated the to expand the genetic code by enabling the co-translational development of new sequencing methods. Lowering DNA and site-specific incorporation of diverse ncAAs into pro- synthesis costs should be one of the major goals in the field teins synthesized in cells (120). To expand the genetic code, of synthetic genomics (118). an orthogonal aminoacyl-tRNA synthetase–tRNA pair is used to direct the incorporation of an ncAA into proteins SYNONYMOUS MUTATIONS AND MICROBE BIOL- through an unassigned codon (e.g. the amber stop codon OGY UAG or the two serine codons UCG and UCA) introduced at the desired site in a gene of interest (121–123). The pos- Aiming the construction of synthetic genomes will not only sibility of integrating ncAAs with diverse structures and enable the development of new technologies, but will also properties into proteins in diverse microorganisms will al- allow an immense increase in knowledge in the biological low proteins to be engineered with alternative structures sciences. Large-scale synonymous recoding has broad ap- and functions, and to develop new therapeutic agents (see plications for modulating diverse biological systems (34). below) (121). Synonymous recoding of synthetic genomes is an effective The combination of DNA synthesis, lambda red recom- alternative to conventional mutational analyses involving a bination (a technique based on homologous recombina- limited segment rather than broad-based genome recoding. tion), and CRISPR-based editing, which allowed the gen- Furthermore, genome recoding may allow us to approach eration of a full synthetic E. coli genome, can be used to biological questions that are very difficult to answer through explore other aspects of bacterial biology. The introduc- conventional methodologies (Table 2). tion in Caulobacter ethensis of 133 313 synonymous sub- Recently, life was elegantly demonstrated to function stitutions that recoded 123 562 codons resulted in the in- with a reduced number of synonymous sense codons (65). activation of 98 genes (124). Interestingly, 27 of the for- An E. coli was created in which the entire 4-Mb genome mer genes contained non-coding control features embedded was replaced with synthetic DNA that no longer uses the within protein-coding sequences. The functional basis of stop codon UAG (amber stop codon) and the two serine this gene inactivation is still poorly understood. Transcrip- codons UCG and UCA. By recoding 18 2014 codons, an tional and/or translational control signals at the termini of organism was created with 59 codons to encode the 20 coding sequences, as well as imprecise genome annotations, amino acids. The previously essential bacterial tRNAs that may be the underlying cause of gene inactivation. The ex- recognize these three codons can now either be deleted or pression of several hundred synonymous mutants of green reassigned to recruit non-canonical amino acids (ncAAs) fluorescent protein (GFP) in E. coli revealed that many syn- into proteins synthesized in cells (109). The synthesis of a onymous variants reduce bacterial fitness (18). The underly- Nucleic Acids Research, 2019, Vol. 47, No. 20 10511 Figure 3. Synthetic DNA generation of a full-length bacterial genome. (A) Oligonucleotides (100–200 nucleotides long) are recombined by PCR, cloned in a bacterial plasmid, and sequenced. Synthetic DNA fragments from plasmids of ∼10 kb (pink, green, and blue) are assembled by homologous recombi- nation in Saccharomyces cerevisiae using bacterial artificial chromosomes (BACs) to generate a length of ∼100 kb. BAC fragments contain Cas9 cleavage sites (black arrows) that allow them to be liberated and to replace genomic bacterial DNA by homologous recombination. The double-selection cassette (−1, +1) (e.g. rpsL-kanR) ensures the integration of the synthetic DNA, and the double-selection cassette (−2, +2) on the genome ensures the removal of the corresponding wild-type DNA. (B) Different BAC fragments have homologous overlapping sequences and corresponding selection cassettes for sequentially and iteratively replace the bacterial genome in the precise genome location. ing mechanism is independent of translation but depends on factor that suppresses HIV-1 protein synthesis in a codon the production of toxic mRNA molecules, suggesting that usage-dependent manner (128). SLFN11 utilizes the unique translation-independent RNA toxicity may be an unrecog- viral codon bias towards A/U nucleotides. Consequently, nized obstacle in bacterial gene expression. wild-type (WT) HIV-1 is affected by SLFN11, but not re- Due to the shorter genomes of RNA viruses, synthetic coded viruses containing HIV-1 structural gag sequences genomics has been employed much more often to explore optimized for human cell expression (i.e. without A/U in diverse aspects of virus biology. Similar to what has been the third position) (128). Using synonymous mutagenesis, found in C. crescentus, large-scale synonymous recoding CpG suppression was recently shown to be essential for has allowed the identification of diverse features encoded HIV-1 (50). The effect of CpG dinucleotides is exerted post- within virus genomes (Table 2). Large-scale synonymous re- transcriptionally and independent of translation efficiency. coding of viral genomes has uncovered new antiviral mech- A zinc-finger antiviral protein (ZAP) has been identified anisms within the innate immune response (125). Com- that inhibits virion production by cells infected with CpG- pared to the host cell preference, the human immunode- enriched HIV-1. These results suggest that ZAP utilizes ficiency virus type 1 (HIV-1) genome contains extremely host CpG suppression to identify non-self RNA. Similarly, high frequencies of A nucleotides (between 31.7% and 38%), large-scale synonymous recoding of viral genomes allows low G/C content, and suboptimal codon usage, partic- the identification of viral sequence patterns, including CpG ularly within the gag (virus structural proteins) and pol content, detected by innate immune pathways in different (virus enzymes) sequences (126,127). This uncommon rare viral systems, such as influenza virus (48,129), echovirus codon bias favors A/U in the third codon position. Schlafen (130), foot and mouth disease virus (131), and simian im- (SLFN) genes contain a subgroup of interferon-stimulated munodeficiency virus (SIV) (77). early-response genes that are directly induced by pathogens RNA virus populations are genetically very heteroge- via the interferon regulatory factor 3 pathway (54). Hu- neous. Genetic variation is generated by the accumulation man SLFN11 has been described as a novel host restriction of mutations during replication and their rearrangement by 10512 Nucleic Acids Research, 2019, Vol. 47, No. 20 Table 2. Synonymous recoding as a new tool for exploring microbial biology Explored features Microbe Targeted genome region References Reduced number of synonymous sense E. coli Full-genome (65,121) codons Specific mRNA structures required for Poliovirus P2 and P3 (138,139) replication Virus genome cis-inhibitory signal Respiratory Syncytial Virus NS1 and NS2 (153) sequences relevant to complex viral functions Non-coding control features embedded C. ethensis Full-genome (124) within protein-coding sequences Immune innate response Influenza Virus Segment 5 (48,50,77,128–130) HIV-1 Env and Gag SIV Gag and Pol Echovirus 7 VP3, 3C, and 3D Codon usage and the temporal regulation SIV Herpesvirus Env ORF57 (141) of viral gene expression Genome position in sequence space, Poliovirus Capsid (134–137) evolutionary trajectory, and pathogenicity Coxsackie virus Capsid Influenza Virus Polymerase and Haemagglutinin HIV-1 Protease genetic recombination, and genome segment reassortment tease inhibitors (PIs). Ultradeep sequence clonal analysis in the case of segmented genomes (132). Unlike DNA poly- revealed that both viruses harbor previously described mu- merases, RNA viruses lack proofreading activity, resulting tations conferring resistance to PIs. However, the WT and in an estimated error rate during replication of 10−3 to 10−5 recoded HIV-1 proteases had different resistance variant mutations per nucleotide per replication cycle. This high repertoires, indicating that the HIV-1 protease position in mutation rate generates a virus population composed of sequence space delineates the evolution of its mutant spec- closely related mutant spectra or mutant clouds termed vi- trum. Importantly, in both poliovirus and HIV-1, synony- ral quasispecies (133). The quasispecies dynamics of RNA mously recoded viruses did not have a reduced replication viruses are intimately related to both viral disease and an- capacity compared to the WT virus. tiviral treatment strategies. One unexplored aspect of the Adenoviruses are DNA viruses broadly employed as vec- genetic architecture of RNA viruses is how codon choice tors for human gene therapy and vaccines. In particular, influences population diversity and evolvability. In order to single-stranded adenovirus-associated virus (AAV) has the explore this aspect of RNA virus biology, the quasispecies unique capacity for site-specific integration into a transcrip- structure of a poliovirus containing 934 synonymous sub- tionally silent region of the human genome. However, AAV stitutions randomly introduced in the 2643-nucleotide-long Rep 78 protein has an inhibitory effect on adenovirus repli- capsid coding region (72) was compared to the WT virus cation. In order to explore the inhibitory mechanism of (134). As expected, the recoded virus explores a different Rep78, synonymous recoding was performed on the 1866- sequence space and gives rise to a distinct spectrum of mu- nucleotide-long Rep genomic segment (76). The genetically tants. Interestingly, mice infected with the synonymously re- recoded AAV Rep gene permitted the identification of a coded poliovirus exhibit delayed and decreased mortality 135-bp long cis-acting element in the Rep sequence. Reengi- compared to mice inoculated with the same doses of WT neering this cis-acting element overcame Rep’s inhibitory ef- virus, suggesting that a virus quasispecies structure and mu- fects on adenovirus replication. Reengineered Rep is readily tant spectrum impact viral pathogenesis. This attenuated tolerated, facilitating the development of improved adeno- phenotype exhibited by the recoded virus suggests a pos- based gene-therapy vectors. Large-scale codon optimiza- sible new strategy for the generation of live attenuated vac- tion of two DNA viruses, adenovirus and papilloma virus, cines (34,135). Another strategy explored to attenuate cox- has also aided in our understanding of how these viruses sackie B3 and influenza A viruses has been to redirect their increase their fitness, immunogenicity, and oncogenic po- evolutionary trajectories towards detrimental regions of se- tential (81,83). The utility of synonymous recoding to iden- quence space (136). Viral genomes were synonymously re- tify the locations of critical DNA or RNA sequence signals coded to harbor more serine and leucine codons that could was also demonstrated with poliovirus. The introduction of generate stop mutations after a single nucleotide substi- 2256 synonymous substitutions into the 6 606-nucleotide- tution (136). These engineered viruses generate more stop long non-structural protein-coding region of poliovirus al- codons and result in significant losses in viral fitness. More lowed the discovery of two 75-nucleotide-long unique, func- recently, synonymous virus genome recoding has been used tionally redundant RNA elements, which are separated by to investigate how the HIV-1 position in sequence space 150 nucleotides, in the 3 -terminal coding sequence of po- defines its mutant spectrum. To this end, a synthetic virus liovirus RNA polymerase (138). Similarly, massively re- was generated carrying a reengineered protease sequence in- coded poliovirus has allowed the identification of conserved cluding 13% of synonymous mutations (137). WT and re- synonymous sites and local secondary RNA that play a role coded viruses were propagated in the presence of HIV-1 pro- in determining virus packaging (139). Nucleic Acids Research, 2019, Vol. 47, No. 20 10513 Another example of the utility of large-scale synony- terial lysis adjuvants to stimulate antitumor immunity and mous recoding is the identification of how codon usage promote tumor regression. is implicated in the temporal regulation of gene expres- An obvious application of synonymous nucleotide recod- sion. A number of viruses are able to maintain life-long in- ing is codon optimization to increase protein expression fections in their natural host despite strong host immune- and improve the production of therapeutic antibodies, cy- surveillance measures. Persisting viruses can be found in the tokines, or fusion proteins (reviewed in (143)). However, lenti-retroviruses (e.g. HIV-1/SIV), herpes, polyoma, pa- in mammalian cells, codon optimization is not evident be- pilloma, parvo, and adenovirus families. Persisting viruses cause, as noted previously in this review, not all synonymous exquisitely regulate their gene expression in a time-based codon mutations are neutral. manner. In many cases, such as HIV-1, structural proteins The approach of reducing expression by intentional in- are made late in the viral replication cycle, and expression of troduction of suboptimal synonymous codons has been these late gene products is typically induced by viral transin- used extensively with viruses (Table 3). The first com- ducers made earlier in the viral replication cycle (140). In- plete synthesis of an organism’s genome was performed triguingly, structural gene products of different members of with poliovirus (103). Similarly, the first large-scale syn- human persisting viruses exhibit a marked bias of codon onymous recoding of a genome was carried out with po- usage compared to the codon usage of the human exome liovirus. Simultaneously, Burns et al. and Mueller et al. (141). In HIV-1/SIV, the viral protein Rev acts as a transin- demonstrated the viability of synonymously recoding the ducer. Rev protein interacts with a highly structured ele- poliovirus genome (72,79). Using the Sabin type 2 oral po- ment in the envelope gene (env), referred to as the Rev- liovirus vaccine strain, Burns et al. reengineered the capsid responsive element (RRE). In the absence of Rev, mRNAs coding region of poliovirus by replacing degenerate codons of the HIV-1 late structural genes (e.g. Env protein) are re- for nine amino acids with corresponding non-preferred syn- tained in the nucleus, preventing their translation. Env has onymous codons. Up to 542 synonymous substitutions were unusual codon usage that is different from that of the host introduced among the 2 555 nucleotides of the virus capsid cell. Synonymous recoding of env has demonstrated that region. Quantifying virus yields in HeLa cells and plaque Rev induction of Env protein expression depends on its bi- areas (visible structures formed within a cell culture by virus ased codon usage. Rev inducibility was lost when 10.5% of cell destruction), these authors demonstrated that the repli- the env codons were synonymously replaced (141). Anal- cation capacity of the virus was reduced up to 98%. When ogous results were obtained when this approach was used recoded viruses were serially propagated in HeLa cells to with rhesus monkey rhadinovirus (RRV), a member of the test their stability, viral evolution was demonstrated by in- herpesvirus family, ORF57 transinducer (141). ORF57 in- creased virus yields, increased plaque sizes, and mutated ge- duces the late expression of RRV structural glycoprotein nomic sequences (79). Mueller et al. were also able to reduce H (gH), but this inducibility was lost when 14.3% the gH the replication capacity of the Mahoney strain of poliovirus codons were replaced with synonymous codons. Interest- by introducing suboptimal synonymous codon changes in ingly, the transinducibility of Env expression can be regu- the virus capsid coding region (72). When this suboptimal lated by ORF57 instead of Rev by modifying env to make recoded virus was inoculated in poliovirus CD155 trans- it better resemble the codon use of gH. Similarly, the de- genic mice, the variant presented a neuroattenuated pheno- pendence of gH on ORF57 expression can be changed to type. Attenuation of the classical oral poliovirus vaccine is Rev dependence by changing the nature of its codon usage. based on very few point mutations, which can revert to vir- These results suggest that persisting viruses use poor codon ulence after a few rounds of viral replication (144). These usage that is biased in a distinctive way to temporally reg- pioneering results obtained with recoded polioviruses sug- ulate late expression of structural gene products. Coupling gest that codon-usage in recoded viruses may be much more synthetic genomics and synonymous recoding is an inter- stable than most RNA virus point mutants, and could pos- esting and effective strategy for modulating diverse biolog- sibly enable the development of live attenuated RNA virus ical systems, and a possible alternative to conventional mu- vaccines with superior genetic stability. Moreover, synony- tational analysis involving a limited number of nucleotide mous recoding does not affect the viral amino acid sequence substitutions. and its subsequent antigenicity. Large-scale synonymous codon usage recoding has also been used to obtain safe live attenuated vaccines for ar- SYNONYMOUS MUTATIONS AND MICROBE THERA- boviruses, such as Chikungunya virus (74) and tick-borne PEUTICS encephalitis virus (85). An interesting approach to develop- Synthetic genomics is pushing a new era of medicine via ing new arbovirus vaccines is to codon-optimize the virus the genetic programming of living cells. A recent exam- to grow in insect cells and deoptimize it to reduce its repli- ple is bacteria. An E. coli strain containing a synchronized cation capacity to non-viability in mammalian cells (75). lysis circuit that colonizes tumors and undergoes intratu- Viruses that are non-viable in mammalian cells could still be moral quorum lysis was engineered to locally release an en- grown in insect cells, implying that one might design a vac- coded nanobody antagonist of CD47, an anti-phagocytic cine that combines features of live attenuated viruses with receptor overexpressed in several human cancers (142). This features of inactivated viruses. Analogously, work with SIV construct allows for the combined local delivery of an im- has shown that codon optimization of gag and pol results munotherapeutic agent along with immunestimulatory bac- in a 100-fold decrease in virus replicative capacity (77). Re- 10514 Nucleic Acids Research, 2019, Vol. 47, No. 20 Table 3. Examples of microbe attenuation by synonymous genome recoding Microbe Targeted genome region Method References Streptococcus pneumoniae Pneumolysin Codon-pair usage (64) Methylobacterium extorquens Formaldehyde-activating enzyme Codon bias usage (62) Salmonella enterica Ribosomal protein S20, L-␥ -glutamyl Codon bias usage (57,58) phosphate reductase Escherichia coli Lac-Z, green florescence protein, Codon bias usage (11,18,60,61) Beta-lactamase Poliovirus Capsid Codon bias usage, (36,68,72,79) Codon-pair usage and Dinucleotide frequency Echovirus 7 VP1, VP2, 3C, 3D Dinucleotide frequency (37,86) Influenza virus Polymerase subunit PB1, PB2, PA, Codon-pair usage (71,80,151) nucleoprotein, matrix, nonstructural protein, neuraminidase, hemagglutinin Human immunodeficiency virus type 1 Protease Codon-pair usage (69) Simian immunodeficiency virus Gag, Pol Codon bias usage (77) Chikungunya virus Non-structural proteins: nsP1, nsP4 Codon bias usage (74) Structural proteins: E2, E1 Dengue virus E, NS3, NS5 Codon-pair usage (75) Zika virus E, NS1 Codon-pair usage (89) Tick-borne encephalitis virus NS5 Codon bias usage (85) Respiratory syncytial virus NS1, NS2, N, P, M, SH, G, F, L Codon-pair usage (87) Porcine reproductive and respiratory Envelope GP5 Codon-pair usage (73) syndrome virus Vesicular stomatitis virus L1 Codon-pair usage (78) Papillomavirus Oncogenes E6, E7 Codon bias usage (83) Marek’s disease virus UL54/ICP27, UL49/VP22, UL30 Codon-pair usage (82,84) Adenovirus Fiber Codon bias usage (81) Cucumber mosaic virus Coat protein Codon bias usage (70) markably, the recoded virus has a reduced ability to stim- virus and echovirus 7 (another picornavirus like poliovirus) ulate type I interferon and possibly attenuated pathogenic (37,48,68,86). As discussed previously, sometimes increased potential. CpG and UpA dinucleotide frequencies and codon-pair de- Wimmer et al. demonstrated that genome recoding using optimization are not easily distinguished (43). underrepresented codon pairs can also result in poliovirus Recoding viral genomes by numerous synonymous but virus attenuation (36). The introduction of 407 synonymous suboptimal substitutions can be a new source of live at- substitutions into the poliovirus (Mahoney) capsid coding tenuated vaccine candidates. The hypothesis is that large- region, which generates underrepresented codon pairs, re- scale recoded viruses should be genetically stable because sulted in decreased rates of virus protein translation, re- deattenuation may require dozens or hundreds of reversions duced replication capacity in HeLa cells, and attenuation in (36). However, their genetic stability under selective pres- mice (36). Similar results were obtained with another vari- sure is largely unknown. Attempts to evaluate phenotype ant carrying 224 synonymous changes in the same region stability have resulted in interesting results. Codon pair de- of the viral genome. These polioviruses were used to im- optimized versions of RSV were generated and presented an munize mice and provided protective immunity after chal- attenuated and temperature-sensitive phenotype (88). When lenge. Codon-pair deoptimization has been performed in recoded viruses were serially propagated in tissue culture at a number of viruses, including negative RNA viruses such progressively increasing temperatures, an RSV variant con- as influenza virus (71, 80) and respiratory syncytial virus taining 2 692 synonymous mutations in 9 of 11 open read- (RSV) (87), vesicular stomatitis virus (78), positive RNA ing frames (ORFs) did not lose temperature sensitivity, re- viruses such as porcine reproductive and respiratory syn- mained genetically stable, and was restricted at tempera- drome virus (73), dengue virus (75), zika virus (89), HIV- tures of 34/35◦ C and above. However, a variant contain- 1 (69), the plant cucumber mosaic virus (70), and DNA ing 1 378 synonymous mutations solely in the polymerase L viruses such as Marek’s disease herpesvirus (82,84). Syn- ORF quickly lost substantial attenuation. Sequence analy- onymous codon pair deoptimization has not been restricted sis of virus populations after tissue culture passages iden- to viruses and has been applied to other microorganisms tified many different potentially deattenuating mutations in (Table 3) (64). the L ORF, as well as many others appearing in other ORFs Synonymously increasing the frequencies of CpG and (88). These results emphasize, again, the rapid evolution of UpA dinuclotides has also been explored as an alter- RNA viruses and the necessity of stability studies to obtain native method of virus attenuation. Most mammalian improved synthetic virus vaccine candidates. RNA and small DNA viruses suppress genomic CpG The recent generation of a full synthetic E. coli genome and UpA dinucleotide frequencies, apparently mimicking (65) has opened up unexplored possibilities for designing a host mRNA composition. Increasing the CpG and UpA new generation of bacterial live attenuated vaccines. Proof dinucleotide frequencies attenuated poliovirus, influenza of the concept has been shown with Streptococcus pneu- Nucleic Acids Research, 2019, Vol. 47, No. 20 10515 moniae (64). Codon pair recoding of S. pneumonia pneu- and recombinant protocols). These new methodologies will molysin gene has resulted in strains that are less virulent be necessary if we want to synthetically engineer eukaryotic in mice than the WT. One particular aspect of the new syn- genomes. Without a doubt, synthetic genomics will greatly thetic E. coli genome is that it uses only 61 of the 64 available impact new diagnostics and therapies, as well as basic re- codons in its protein-coding sequences, replacing two ser- search. ine codons and one stop codon with synonyms (65). These deleted codons can now be reassigned to recruit ncAAs beyond the usual 20 used by most living cells (145). The ACKNOWLEDGEMENTS genome of S. typhimurium has also been found to have two We apologize to those whose works have not been cited in fewer codons; 1 557 synonymous leucine replacements were this article due to a lack of space. performed to delete two leucine codons (102). Engendered S. typhimurium can be used as a cell-based technology in the human gut. Now ncAAs can take part in chemical re- FUNDING actions that natural proteins cannot perform. An expanded Spanish Ministry of Science, Innovation and Universi- genetic code might engender next-generation biopharma- ties [SAF2016–75277-R]; Spanish Ministry of Innova- ceuticals (146). An enlarged amino acid repertoire will al- tion and Universities [BES-2014-069931 to A.J.-P.]; Gri- low proteins to have new or modified functions, providing fols Grant [miRNA 2018]; Spanish AIDS network (RIS) ample opportunities for innovative or improved medicines. [RD16/0025/0041 to M.N.]. Funding for open access Current pipelines involve the areas of half-life extension, charge: Spanish Ministry of Science, Innovation and Uni- antibody-drug conjugates (ADCs), bi-specific antibodies, versities [SAF2016-75277-R]. modular chimeric antigen receptor T cell therapy (CAR-T), Conflict of interest statement. None declared. vaccine development, and expanded applications for diag- nosis and bio-compatible materials (146). These new bio- pharmaceuticals are being explored in cancer and inflam- REFERENCES matory, autoimmune, and metabolic diseases (121). 1. Grantham,R., Gautier,C. and Gouy,M. (1980) Codon frequencies Synonymous codon compression and ncAAs can also be in 119 individual genes confirm consistent choices of degenerate employed to develop viral attenuated vaccines (121). The in- bases according to genome type. Nucleic Acids Res., 8, 1893–1912. 2. Quax,T.E., Claassens,N.J., Soll,D. and van der Oost,J. (2015) Codon troduction of amber codons into the genomes of hepatitis D bias as a means to Fine-Tune gene expression. Mol. Cell, 59, virus, HIV-1 and influenza A has allowed the generation of 149–161. viral genomes that replicate only in cells that can decode the 3. Ikemura,T. (1985) Codon usage and tRNA content in unicellular amber codon (147–149). These viruses can be replicated in and multicellular organisms. Mol. Biol. Evol., 2, 13–34. 4. Brule,C.E. and Grayhack,E.J. (2017) Synonymous Codons: Choose cells that contain orthogonal aminoacyl-tRNA synthetase wisely for expression. Trends Genet., 33, 283–297. and orthogonal tRNACUA . However, they are attenuated in 5. Zhou,Z., Dang,Y., Zhou,M., Li,L., Yu,C.H., Fu,J., Chen,S. and cells that cannot suppress the amber codon. With the in- Liu,Y. (2016) Codon usage is an important determinant of gene troduction of several amber codons throughout the viral expression levels largely through its effects on transcription. Proc. genome, the frequency of revertants can be greatly reduced Natl. Acad. Sci. U.S.A., 113, E6117–E6125. 6. Findlay,G.M., Daza,R.M., Martin,B., Zhang,M.D., Leith,A.P., (148). However, safety studies should be performed because Gasperini,M., Janizek,J.D., Huang,X., Starita,L.M. and Shendure,J. a read-through of stop codons can occur in some cell types (2018) Accurate classification of BRCA1 variants with saturation (150). genome editing. Nature, 562, 217–222. 7. Goodman,D.B., Church,G.M. and Kosuri,S. (2013) Causes and effects of N-terminal codon bias in bacterial genes. Science, 342, OUTLOOK FOR THE FUTURE 475–479. 8. Kudla,G., Murray,A.W., Tollervey,D. and Plotkin,J.B. (2009) Synthetic genomics and synonymous genome-wide rewrit- Coding-sequence determinants of gene expression in Escherichia ing will decode essential genome functions and allow the de- coli. Science, 324, 255–258. 9. Stergachis,A.B., Haugen,E., Shafer,A., Fu,W., Vernot,B., sign of improved microorganisms for industrial and health Reynolds,A., Raubitschek,A., Ziegler,S., LeProust,E.M., Akey,J.M. applications. Although significant advances have been made et al. (2013) Exonic transcription factor binding directs codon in this field, it is still necessary to elucidate the exact mecha- choice and affects protein evolution. Science, 342, 1367–1372. nistic foundations underlying the contributions of synony- 10. Boel,G., Letso,R., Neely,H., Price,W.N., Wong,K.H., Su,M., Luff,J., mous substitutions to phenotype. Unraveling the funda- Valecha,M., Everett,J.K., Acton,T.B. et al. (2016) Codon influence on protein expression in E. coli correlates with mRNA levels. mental mechanistic impact of synonymous substitutions is Nature, 529, 358–363. essential to support the use of synonymous recoding in an 11. Sorensen,M.A., Kurland,C.G. and Pedersen,S. (1989) Codon usage array of biological and clinical contexts. Despite this re- determines translation rate in Escherichia coli. J. Mol. Biol., 207, view focusing on viruses and bacteria, the work can pio- 365–377. 12. Akashi,H. (1994) Synonymous codon usage in Drosophila neer the study of the impact of synonymous substitution melanogaster: natural selection and translational accuracy. Genetics, on human diseases. As exemplified by the development of 136, 927–935. new and massive DNA sequencing methodologies, cheap 13. Drummond,D.A. and Wilke,C.O. (2008) Mistranslation-induced and accurate DNA synthesis will be necessary to achieve protein misfolding as a dominant constraint on coding-sequence ambitious objectives. The synthesis of relatively small DNA evolution. Cell, 134, 341–352. 14. Presnyak,V., Alhusaini,N., Chen,Y.H., Martin,S., Morris,N., fragments (<10 kb) is currently straightforward. However, Kline,N., Olson,S., Weinberg,D., Baker,K.E., Graveley,B.R. et al. new and easier methodologies should be developed to as- (2015) Codon optimality is a major determinant of mRNA stability. semble larger pieces of synthetic DNA (e.g. DNA cloning Cell, 160, 1111–1124. 10516 Nucleic Acids Research, 2019, Vol. 47, No. 20 15. Shabalina,S.A., Ogurtsov,A.Y. and Spiridonov,N.A. (2006) A 35. Gutman,G.A. and Hatfield,G.W. (1989) Nonrandom utilization of periodic pattern of mRNA secondary structure created by the codon pairs in Escherichia coli. Proc. Natl. Acad. Sci. U.S.A., 86, genetic code. Nucleic Acids Res., 34, 2428–2437. 3699–3703. 16. Pagani,F., Raponi,M. and Baralle,F.E. (2005) Synonymous 36. Coleman,J.R., Papamichail,D., Skiena,S., Futcher,B., Wimmer,E. mutations in CFTR exon 12 affect splicing and are not neutral in and Mueller,S. (2008) Virus attenuation by genome-scale changes in evolution. Proc. Natl. Acad. Sci. U.S.A., 102, 6368–6372. codon pair bias. Science, 320, 1784–1787. 17. Takata,M.A., Soll,S.J., Emery,A., Blanco-Melo,D., Swanstrom,R. 37. Tulloch,F., Atkinson,N.J., Evans,D.J., Ryan,M.D. and Simmonds,P. and Bieniasz,P.D. (2018) Global synonymous mutagenesis identifies (2014) RNA virus attenuation by codon pair deoptimisation is an cis-acting RNA elements that regulate HIV-1 splicing and artefact of increases in CpG/UpA dinucleotide frequencies. Elife, 3, replication. PLoS Pathog., 14, e1006824. e04531. 18. Mittal,P., Brindle,J., Stephen,J., Plotkin,J.B. and Kudla,G. (2018) 38. Kunec,D. and Osterrieder,N. (2016) Codon pair bias is a direct Codon usage influences fitness through RNA toxicity. Proc. Natl. consequence of dinucleotide bias. Cell Rep., 14, 55–67. Acad. Sci. U.S.A., 115, 8639–8644. 39. Simmonds,P., Xia,W., Baillie,J. and McKinnon,K. (2013) Modelling 19. Warnecke,T., Weber,C.C. and Hurst,L.D. (2009) Why there is more mutational and selection pressures on dinucleotides in eukaryotic to protein evolution than protein function: splicing, nucleosomes phyla –selection against CpG and UpA in cytoplasmically expressed and dual-coding sequence. Biochem. Soc. Trans., 37, 756–761. RNA and in RNA viruses. BMC Genomics, 14, 610. 20. Birnbaum,R.Y., Clowney,E.J., Agamy,O., Kim,M.J., Zhao,J., 40. Hodgkinson,A. and Eyre-Walker,A. (2011) Variation in the Yamanaka,T., Pappalardo,Z., Clarke,S.L., Wenger,A.M., mutation rate across mammalian genomes. Nat. Rev. Genet., 12, Nguyen,L. et al. (2012) Coding exons function as tissue-specific 756–766. enhancers of nearby genes. Genome Res., 22, 1059–1068. 41. Jones,P.A. (2012) Functions of DNA methylation: islands, start sites, 21. Lin,M.F., Kheradpour,P., Washietl,S., Parker,B.J., Pedersen,J.S. and gene bodies and beyond. Nat. Rev. Genet., 13, 484–492. Kellis,M. (2011) Locating protein-coding sequences under selection 42. Takahashi,Y., Wu,J., Suzuki,K., Martinez-Redondo,P., Li,M., for additional, overlapping functions in 29 mammalian genomes. Liao,H.K., Wu,M.Z., Hernandez-Benitez,R., Hishida,T., Genome Res., 21, 1916–1928. Shokhirev,M.N. et al. (2017) Integration of CpG-free DNA induces 22. Brest,P., Lapaquette,P., Souidi,M., Lebrigand,K., Cesaro,A., de novo methylation of CpG islands in pluripotent stem cells. Vouret-Craviari,V., Mari,B., Barbry,P., Mosnier,J.F., Hebuterne,X. Science, 356, 503–508. et al. (2011) A synonymous variant in IRGM alters a binding site for 43. Futcher,B., Gorbatsevych,O., Shen,S.H., Stauft,C.B., Song,Y., miR-196 and causes deregulation of IRGM-dependent xenophagy in Wang,B., Leatherwood,J., Gardin,J., Yurovsky,A., Mueller,S. et al. Crohn’s disease. Nat. Genet., 43, 242–245. (2015) Reply to Simmonds et al.: Codon pair and dinucleotide bias 23. Pechmann,S. and Frydman,J. (2013) Evolutionary conservation of have not been functionally distinguished. Proc. Natl. Acad. Sci. codon optimality reveals hidden signatures of cotranslational U.S.A., 112, E3635–E3636. folding. Nat. Struct. Mol. Biol., 20, 237–243. 44. Simmonds,P., Tulloch,F., Evans,D.J. and Ryan,M.D. (2015) 24. Plotkin,J.B. and Kudla,G. (2011) Synonymous but not the same: the Attenuation of dengue (and other RNA viruses) with codon pair causes and consequences of codon bias. Nat. Rev. Genet., 12, 32–42. recoding can be explained by increased CpG/UpA dinucleotide 25. Chamary,J. V, Parmley,J.L. and Hurst,L.D. (2006) Hearing silence: frequencies. Proc. Natl. Acad. Sci. U.S.A., 112, E3633–E3634. non-neutral evolution at synonymous sites in mammals. Nat. Rev. 45. Rima,B.K. and McFerran,N. V (1997) Dinucleotide and stop codon Genet., 7, 98–108. frequencies in single-stranded RNA viruses. J. Gen. Virol., 78, 26. Chen,S.L., Lee,W., Hottes,A.K., Shapiro,L. and McAdams,H.H. 2859–2870. (2004) Codon usage between genomes is constrained by 46. van der Kuyl,A.C. and Berkhout,B. (2012) The biased nucleotide genome-wide mutational processes. Proc. Natl. Acad. Sci. U.S.A., composition of the HIV genome: a constant factor in a highly 101, 3480–3485. variable virus. Retrovirology, 9, 92. 27. Plotkin,J.B., Robins,H. and Levine,A.J. (2004) Tissue-specific codon 47. Kumagai,Y., Takeuchi,O. and Akira,S. (2008) TLR9 as a key usage and the expression of human genes. Proc. Natl. Acad. Sci. receptor for the recognition of DNA. Adv. Drug. Deliv. Rev., 60, U.S.A., 101, 12588–12591. 795–804. 28. Hunt,R.C., Simhadri,V.L., Iandoli,M., Sauna,Z.E. and 48. Gaunt,E., Wise,H.M., Zhang,H., Lee,L.N., Atkinson,N.J., Kimchi-Sarfaty,C. (2014) Exposing synonymous mutations. Trends Nicol,M.Q., Highton,A.J., Klenerman,P., Beard,P.M., Dutia,B.M. Genet., 30, 308–321. et al. (2016) Elevation of CpG frequencies in influenza A genome 29. Moura,G.R., Pinheiro,M., Freitas,A., Oliveira,J.L., Frommlet,J.C., attenuates pathogenicity but enhances host response to infection. Carreto,L., Soares,A.R., Bezerra,A.R. and Santos,M.A. (2011) Elife, 5, e12735. Species-specific codon context rules unveil non-neutrality effects of 49. Zhang,L., Li,H., Hai,Y., Yin,W., Li,W., Zheng,B., Du,X., Li,N., synonymous mutations. PLoS One, 6, e26817. Zhang,Z., Deng,Y. et al. (2017) CpG in combination with an 30. Tats,A., Tenson,T. and Remm,M. (2008) Preferred and avoided inhibitor of notch signaling suppresses Formalin-Inactivated codon pairs in three domains of life. BMC Genomics, 9, 463. respiratory syncytial Virus-Enhanced airway hyperresponsiveness 31. Alexaki,A., Kames,J., Holcomb,D.D., Athey,J., and inflammation by inhibiting Th17 memory responses and Santana-Quintero,L. V, Lam,P.V.N., Hamasaki-Katagiri,N., promoting Tissue-Resident memory cells in lung. J. Virol., 91, Osipova,E., Simonyan,V., Bar,H. et al. (2019) Codon and e02111-16. Codon-Pair Usage Tables (CoCoPUTs): facilitating genetic 50. Takata,M.A., Goncalves-Carneiro,D., Zang,T.M., Soll,S.J., variation analyses and recombinant gene design. J. Mol. Biol., 431, York,A., Blanco-Melo,D. and Bieniasz,P.D. (2017) CG dinucleotide 2434–2441. suppression enables antiviral defence targeting non-self RNA. 32. Biasoli,D., Compston-Garnett,L., Ricketts,S.L., Birand,Z., Nature, 550, 124–127. Courtay-Cahen,C., Fineberg,E., Arendt,M., Boerkamp,K., 51. Curran,J.F., Poole,E.S., Tate,W.P. and Gross,B.L. (1995) Selection of Melin,M., Koltookian,M. et al. (2019) A synonymous germline aminoacyl-tRNAs at sense codons: the size of the tRNA variable variant in a gene encoding a cell adhesion molecule is associated loop determines whether the immediate 3 nucleotide to the codon with cutaneous mast cell tumour development in Labrador and has a context effect. Nucleic Acids Res., 23, 4104–4108. Golden Retrievers. PLos Genet., 15, e1007967. 52. Buchan,J.R., Aucott,L.S. and Stansfield,I. (2006) tRNA properties 33. Zhou,J., Park,C.Y., Theesfeld,C.L., Wong,A.K., Yuan,Y., help shape codon pair preferences in open reading frames. Nucleic Scheckel,C., Fak,J.J., Funk,J., Yao,K., Tajima,Y. et al. (2019) Acids Res., 34, 1015–1027. Whole-genome deep-learning analysis identifies contribution of 53. Gamble,C.E., Brule,C.E., Dean,K.M., Fields,S. and Grayhack,E.J. noncoding mutations to autism risk. Nat. Genet., 51, 973–980. (2016) Adjacent codons act in concert to modulate translation 34. Martinez,M.A., Jordan-Paiz,A., Franco,S. and Nevot,M. (2016) efficiency in yeast. Cell, 166, 679–690. Synonymous virus genome recoding as a tool to impact viral fitness. 54. Chevance,F.F., Le Guyon,S. and Hughes,K.T. (2014) The effects of Trends Microbiol., 24, 134–147. codon context on in vivo translation speed. PLos Genet., 10, e1004392. Nucleic Acids Research, 2019, Vol. 47, No. 20 10517 55. Irwin,B., Heck,J.D. and Hatfield,G.W. (1995) Codon pair utilization 75. Shen,S.H., Stauft,C.B., Gorbatsevych,O., Song,Y., Ward,C.B., biases influence translational elongation step times. J. Biol. Chem., Yurovsky,A., Mueller,S., Futcher,B. and Wimmer,E. (2015) 270, 22801–22806. Large-scale recoding of an arbovirus genome to rebalance its insect 56. Chevance,F.F.V. and Hughes,K.T. (2017) Case for the genetic code versus mammalian preference. Proc. Natl. Acad. Sci. U.S.A., 112, as a triplet of triplets. Proc. Natl. Acad. Sci. U.S.A., 114, 4745–4750. 4749–4754. 57. Kristofich,J., Morgenthaler,A.B., Kinney,W.R., Ebmeier,C.C., 76. Sitaraman,V., Ward,C. and Skiena,S. (2012) Synthetic sequence Snyder,D.J., Old,W.M., Cooper,V.S. and Copley,S.D. (2018) design for signal location search. Lect. Notes Comput. Sci., 7262, Synonymous mutations make dramatic contributions to fitness when 165–179. growth is limited by a weak-link enzyme. PLos Genet., 14, e1007615. 77. Vabret,N., Bailly-Bechet,M., Lepelley,A., Najburg,V., Schwartz,O., 58. Knoppel,A., Nasvall,J. and Andersson,D.I. (2016) Compensating Verrier,B. and Tangy,F. (2014) Large-scale nucleotide optimization the fitness costs of synonymous mutations. Mol. Biol. Evol., 33, of simian immunodeficiency virus reduces its capacity to stimulate 1461–1477. type I interferon in vitro. J. Virol., 88, 4161–4172. 59. Raghavan,R., Kelkar,Y.D. and Ochman,H. (2012) A selective force 78. Wang,B., Yang,C., Tekes,G., Mueller,S., Paul,A., Whelan,S.P.J. and favoring increased G+C content in bacterial genes. Proc. Natl. Acad. Wimmer,E. (2015) Recoding of the vesicular stomatitis virus L gene Sci. U.S.A., 109, 14504–14507. by Computer-Aided design provides a live, attenuated vaccine 60. Ballard,A., Bieniek,S. and Carlini,D.B. (2019) The fitness candidate. MBio, 6, e00237-15. consequences of synonymous mutations in Escherichia coli: 79. Burns,C.C., Shaw,J., Campagnoli,R., Jorba,J., Vincent,A., Quay,J. Experimental evidence for a pleiotropic effect of translational and Kew,O. (2006) Modulation of poliovirus replicative fitness in selection. Gene, 694, 111–120. HeLa cells by deoptimization of synonymous codon usage in the 61. Yannai,A., Katz,S. and Hershberg,R. (2018) The codon usage of capsid region. J. Virol., 80, 3259–3272. lowly expressed genes is subject to natural selection. Genome Biol. 80. Yang,C., Skiena,S., Futcher,B., Mueller,S. and Wimmer,E. (2013) Evol., 10, 1237–1246. Deliberate reduction of hemagglutinin and neuraminidase 62. Agashe,D., Martinez-Gomez,N.C., Drummond,D.A. and Marx,C.J. expression of influenza virus leads to an ultraprotective live vaccine (2013) Good codons, bad Transcript: Large reductions in gene in mice. Proc. Natl. Acad. Sci. U.S.A., 110, 9481–9486. expression and fitness arising from synonymous mutations in a key 81. Villanueva,E., Martı́-Solano,M. and Fillat,C. (2016) Codon enzyme. Mol. Biol. Evol., 30, 549–560. optimization of the adenoviral fiber negatively impacts structural 63. Brandis,G. and Hughes,D. (2016) The Selective Advantage of protein expression and viral fitness. Sci. Rep., 6, 27546. Synonymous Codon Usage Bias in Salmonella. PLos Genet., 12, 82. Eschke,K., Trimpert,J., Osterrieder,N. and Kunec,D. (2018) e1005926. Attenuation of a very virulent Marek’s disease herpesvirus (MDV) 64. Coleman,J.R., Papamichail,D., Yano,M., Garcia-Suarez Mdel,M. by codon pair bias deoptimization. PLoS Pathog., 14, e1006857. and Pirofski,L.A. (2011) Designed reduction of Streptococcus 83. Cladel,N.M., Budgeon,L.R., Hu,J., Balogh,K.K. and pneumoniae pathogenicity via synthetic changes in virulence factor Christensen,N.D. (2013) Synonymous codon changes in the codon-pair bias. J. Infect. Dis., 203, 1264–1273. oncogenes of the cottontail rabbit papillomavirus lead to increased 65. Fredens,J., Wang,K., de la Torre,D., Funke,L.F.H., Robertson,W.E., oncogenicity and immunogenicity of the virus. Virology, 438, 70–83. Christova,Y., Chia,T., Schmied,W.H., Dunkelmann,D.L., 84. Conrad,S.J., Silva,R.F., Hearn,C.J., Climans,M. and Dunn,J.R. Beránek,V. et al. (2019) Total synthesis of Escherichia coli with a (2018) Attenuation of Marek’s disease virus by codon pair recoded genome. Nature, 569, 514–518. deoptimization of a core gene. Virology, 516, 219–226. 66. Frumkin,I., Schirman,D., Rotman,A., Li,F., Zahavi,L., Mordret,E., 85. de Fabritus,L., Nougairede,A., Aubry,F., Gould,E.A. and de Asraf,O., Wu,S., Levy,S.F. and Pilpel,Y. (2017) Gene architectures Lamballerie,X. (2015) Attenuation of tick-borne encephalitis virus that minimize cost of gene expression. Mol. Cell, 65, 142–153. using large-scale random codon re-encoding. PLoS Pathog., 11, 67. Kelsic,E.D., Chung,H., Cohen,N., Park,J., Wang,H.H. and e1004738. Kishony,R. (2016) RNA structural determinants of optimal codons 86. Fros,J.J., Dietrich,I., Alshaikhahmed,K., Passchier,T.C., Evans,D.J. revealed by MAGE-Seq. Cell Syst., 3, 563–571. and Simmonds,P. (2017) CpG and UpA dinucleotides in both 68. Burns,C.C., Campagnoli,R., Shaw,J., Vincent,A., Jorba,J. and coding and non-coding regions of echovirus 7 inhibit replication Kew,O. (2009) Genetic inactivation of poliovirus infectivity by initiation post-entry. Elife, 6, e29112. increasing the frequencies of CpG and UpA dinucleotides within 87. Le Nouen,C., Brock,L.G., Luongo,C., McCarty,T., Yang,L., and across synonymous capsid region codons. J. Virol., 83, Mehedi,M., Wimmer,E., Mueller,S., Collins,P.L., Buchholz,U.J. 9957–9969. et al. (2014) Attenuation of human respiratory syncytial virus by 69. Martrus,G., Nevot,M., Andres,C., Clotet,B. and Martinez,M. genome-scale codon-pair deoptimization. Proc. Natl. Acad. Sci. (2013) Changes in codon-pair bias of human immunodeficiency U.S.A., 111, 13169–13174. virus type 1 have profound effects on virus replication in cell culture. 88. Le Nouen,C., McCarty,T., Brown,M., Smith,M.L., Lleras,R., Retrovirology, 10, 78. Dolan,M.A., Mehedi,M., Yang,L., Luongo,C., Liang,B. et al. 70. Mochizuki,T., Ohara,R. and Roossinck,M.J. (2018) Large-Scale (2017) Genetic stability of genome-scale deoptimized RNA virus synonymous substitutions in cucumber mosaic virus RNA 3 vaccine candidates under selective pressure. Proc. Natl. Acad. Sci. facilitate amino acid mutations in the coat protein. J. Virol., 92, U.S.A., 114, E386–E395. e01007-18. 89. Li,P., Ke,X., Wang,T., Tan,Z., Luo,D., Miao,Y., Sun,J., Zhang,Y., 71. Mueller,S., Coleman,J.R., Papamichail,D., Ward,C.B., Nimnual,A., Liu,Y., Hu,Q. et al. (2018) Zika virus attenuation by codon pair Futcher,B., Skiena,S. and Wimmer,E. (2010) Live attenuated deoptimization induces sterilizing immunity in mouse models. J. influenza virus vaccines by computer-aided rational design. Nat. Virol., 92, e00701-18. Biotechnol., 28, 723–726. 90. Hughes,R.A. and Ellington,A.D. (2017) Synthetic DNA synthesis 72. Mueller,S., Papamichail,D., Coleman,J.R., Skiena,S. and and assembly: putting the synthetic in synthetic biology. Cold Spring Wimmer,E. (2006) Reduction of the rate of poliovirus protein Harb. Perspect. Biol., 9, a023812. synthesis through large-scale codon deoptimization causes 91. Schindler,D., Dai,J. and Cai,Y. (2018) Synthetic genomics: a new attenuation of viral virulence by lowering specific infectivity. J. venture to dissect genome fundamentals and engineer new Virol., 80, 9687–9696. functions. Curr. Opin. Chem. Biol., 46, 56–62. 73. Ni,Y.-Y., Zhao,Z., Opriessnig,T., Subramaniam,S., Zhou,L., 92. Cameron,D.E., Bashor,C.J. and Collins,J.J. (2014) A brief history of Cao,D., Cao,Q., Yang,H. and Meng,X.-J. (2014) Computer-aided synthetic biology. Nat. Rev. Microbiol., 12, 381–390. codon-pairs deoptimization of the major envelope GP5 gene 93. Brenner,S., Johnson,M., Bridgham,J., Golda,G., Lloyd,D.H., attenuates porcine reproductive and respiratory syndrome virus. Johnson,D., Luo,S., McCurdy,S., Foy,M., Ewan,M. et al. (2000) Virology, 450–451, 132–139. Gene expression analysis by massively parallel signature sequencing 74. Nougairede,A., De Fabritus,L., Aubry,F., Gould,E.A., Holmes,E.C. (MPSS) on microbead arrays. Nat. Biotechnol., 18, 630–634. and de Lamballerie,X. (2013) Random codon re-encoding induces 94. Ronaghi,M., Karamohamed,S., Pettersson,B., Uhlen,M. and stable reduction of replicative fitness of Chikungunya virus in Nyren,P. (1996) Real-time DNA sequencing using detection of primate and mosquito cells. PLoS Pathog., 9, e1003172. pyrophosphate release. Anal. Biochem., 242, 84–89. 10518 Nucleic Acids Research, 2019, Vol. 47, No. 20 95. Schatz,M.C. and Phillippy,A.M. (2012) The rise of a digital immune chromosome V and behavior of a ring derivative. Science, 355, system. Gigascience, 1, 4. eaaf4704. 96. Kosuri,S. and Church,G.M. (2014) Large-scale de novo DNA 116. Lander,E.S., Linton,L.M., Birren,B., Nusbaum,C., Zody,M.C., synthesis: technologies and applications. Nat. Methods, 11, 499–507. Baldwin,J., Devon,K., Dewar,K., Doyle,M., FitzHugh,W. et al. 97. Wan,W., Li,L., Xu,Q., Wang,Z., Yao,Y., Wang,R., Zhang,J., Liu,H., (2001) Initial sequencing and analysis of the human genome. Gao,X. and Hong,J. (2014) Error removal in microchip-synthesized Nature, 409, 860–921. DNA using immobilized MutS. Nucleic Acids Res., 42, e102. 117. Venter,J.C., Adams,M.D., Myers,E.W., Li,P.W., Mural,R.J., 98. Gibson,D.G., Glass,J.I., Lartigue,C., Noskov,V.N., Chuang,R.Y., Sutton,G.G., Smith,H.O., Yandell,M., Evans,C.A., Holt,R.A. et al. Algire,M.A., Benders,G.A., Montague,M.G., Ma,L., Moodie,M.M. (2001) The sequence of the human genome. Science, 291, 1304–1351. et al. (2010) Creation of a bacterial cell controlled by a chemically 118. Boeke,J.D., Church,G., Hessel,A., Kelley,N.J., Arkin,A., Cai,Y., synthesized genome. Science, 329, 52–56. Carlson,R., Chakravarti,A., Cornish,V.W., Holt,L. et al. (2016) 99. Gibson,D.G., Young,L., Chuang,R.Y., Venter,J.C., Hutchison,C.A. GENOME ENGINEERING. The Genome Project-Write. Science, and Smith,H.O. (2009) Enzymatic assembly of DNA molecules up 353, 126–127. to several hundred kilobases. Nat. Methods, 6, 343–345. 119. Ostrov,N., Landon,M., Guell,M., Kuznetsov,G., Teramoto,J., 100. Gibson,D.G., Smith,H.O., Hutchison,C.A. 3rd, Venter,J.C. and Cervantes,N., Zhou,M., Singh,K., Napolitano,M.G., Merryman,C. (2010) Chemical synthesis of the mouse Moosburner,M. et al. (2016) Design, synthesis, and testing toward a mitochondrial genome. Nat. Methods, 7, 901–903. 57-codon genome. Science, 353, 819–822. 101. Torella,J.P., Lienert,F., Boehm,C.R., Chen,J.H., Way,J.C. and 120. Mukai,T., Lajoie,M.J., Englert,M. and Söll,D. (2017) Rewriting the Silver,P.A. (2014) Unique nucleotide sequence-guided assembly of Genetic Code. Annu. Rev. Microbiol., 71, 557–577. repetitive DNA parts for synthetic biology applications. Nat. 121. Chin,J.W. (2017) Expanding and reprogramming the genetic code. Protoc., 9, 2075–2089. Nature, 550, 53–60. 102. Lau,Y.H., Stirling,F., Kuo,J., Karrenbelt,M.A.P., Chan,Y.A., 122. Chin,J.W. (2014) Expanding and reprogramming the genetic code of Riesselman,A., Horton,C.A., Schafer,E., Lips,D., Weinstock,M.T. cells and animals. Annu. Rev. Biochem., 83, 379–408. et al. (2017) Large-scale recoding of a bacterial genome by iterative 123. Liu,C.C. and Schultz,P.G. (2010) Adding new chemistries to the recombineering of synthetic DNA. Nucleic Acids Res., 45, genetic code. Annu. Rev. Biochem., 79, 413–444. 6971–6980. 124. Venetz,J.E., Del Medico,L., Wolfle,A., Schachle,P., Bucher,Y., 103. Cello,J., Paul,A.V. and Wimmer,E. (2002) Chemical synthesis of Appert,D., Tschan,F., Flores-Tinoco,C.E., van Kooten,M., poliovirus cDNA: generation of infectious virus in the absence of Guennoun,R. et al. (2019) Chemical synthesis rewriting of a natural template. Science, 297, 1016–1018. bacterial genome to achieve design flexibility and biological 104. Smith,H.O., Hutchison,C.A. 3rd, Pfannkoch,C. and Venter,J.C. functionality. Proc. Natl. Acad. Sci. U.S.A., 116, 8070–8079. (2003) Generating a synthetic genome by whole genome assembly: 125. Vabret,N., Bhardwaj,N. and Greenbaum,B.D. (2017) phiX174 bacteriophage from synthetic oligonucleotides. Proc. Natl. Sequence-Specific sensing of Nucleic Acids. Trends Immunol., 38, Acad. Sci. U.S.A., 100, 15440–15445. 53–65. 105. Koblentz,G.D. (2017) The de novo synthesis of horsepox Virus: 126. Kypr,J. and Mrazek,J. (1987) Unusual codon usage of HIV. Nature, Implications for biosecurity and recommendations for preventing 327, 20. the reemergence of smallpox. Heal. Secur, 15, 620–628. 127. Meintjes,P.L. and Rodrigo,A.G. (2005) Evolution of relative 106. Yao,X.D. and Evans,D.H. (2003) High-frequency genetic synonymous codon usage in Human Immunodeficiency Virus recombination and reactivation of orthopoxviruses from DNA type-1. J. Bioinform. Comput. Biol., 3, 157–168. fragments transfected into leporipoxvirus-infected cells. J. Virol., 77, 128. Li,M., Kao,E., Gao,X., Sandig,H., Limmer,K., Pavon-Eternod,M., 7281–7290. Jones,T.E., Landry,S., Pan,T., Weitzman,M.D. et al. (2012) 107. Hutchison,C.A., Chuang,R.Y., Noskov,V.N., Assad-Garcia,N., Codon-usage-based inhibition of HIV protein synthesis by human Deerinck,T.J., Ellisman,M.H., Gill,J., Kannan,K., Karas,B.J., Ma,L. schlafen 11. Nature, 491, 125–128. et al. (2016) Design and synthesis of a minimal bacterial genome. 129. Jimenez-Baranda,S., Greenbaum,B., Manches,O., Handler,J., Science, 351, aad6253. Rabadan,R., Levine,A. and Bhardwaj,N. (2011) Oligonucleotide 108. Posfai,G., Plunkett,G., Feher,T., Frisch,D., Keil,G.M., motifs that disappear during the evolution of influenza virus in Umenhoffer,K., Kolisnychenko,V., Stahl,B., Sharma,S.S., de humans increase alpha interferon secretion by plasmacytoid Arruda,M. et al. (2006) Emergent properties of reduced-genome dendritic cells. J. Virol., 85, 3893–3904. Escherichia coli. Science, 312, 1044–1046. 130. Atkinson,N.J., Witteveldt,J., Evans,D.J. and Simmonds,P. (2014) 109. Wang,K., Fredens,J., Brunner,S.F., Kim,S.H., Chia,T. and Chin,J.W. The influence of CpG and UpA dinucleotide frequencies on RNA (2016) Defining synonymous codon compression schemes by virus replication and characterization of the innate cellular pathways genome recoding. Nature, 539, 59–64. underlying virus attenuation and enhanced replication. Nucleic 110. Annaluru,N., Muller,H., Mitchell,L.A., Ramalingam,S., Acids Res., 42, 4527–4545. Stracquadanio,G., Richardson,S.M., Dymond,J.S., Kuang,Z., 131. Diaz-San Segundo,F., Medina,G.N., Ramirez-Medina,E., Scheifele,L.Z., Cooper,E.M. et al. (2014) Total synthesis of a Velazquez-Salinas,L., Koster,M., Grubman,M.J. and de los functional designer eukaryotic chromosome. Science, 344, 55–58. Santos,T. (2016) Synonymous deoptimization of foot-and-mouth 111. Mitchell,L.A., Wang,A., Stracquadanio,G., Kuang,Z., Wang,X., disease virus causes attenuation in vivo while inducing a strong Yang,K., Richardson,S., Martin,J.A., Zhao,Y., Walker,R. et al. neutralizing antibody response. J. Virol., 90, 1298–1310. (2017) Synthesis, debugging, and effects of synthetic chromosome 132. Andino,R. and Domingo,E. (2015) Viral quasispecies. Virology, consolidation: synVI and beyond. Science, 355, eaaf4831. 479–480, 46–51. 112. Richardson,S.M., Mitchell,L.A., Stracquadanio,G., Yang,K., 133. Mas,A., Lopez-Galindez,C., Cacho,I., Gomez,J. and Dymond,J.S., DiCarlo,J.E., Lee,D., Huang,C.L., Martinez,M.A. (2010) Unfinished stories on viral quasispecies and Chandrasegaran,S., Cai,Y. et al. (2017) Design of a synthetic yeast Darwinian views of evolution. J. Mol. Biol., 397, 865–877. genome. Science, 355, 1040–1044. 134. Lauring,A.S., Acevedo,A., Cooper,S.B. and Andino,R. (2012) 113. Shen,Y., Wang,Y., Chen,T., Gao,F., Gong,J., Abramczyk,D., Codon usage determines the mutational robustness, evolutionary Walker,R., Zhao,H., Chen,S., Liu,W. et al. (2017) Deep functional capacity, and virulence of an RNA virus. Cell Host Microbe., 12, analysis of synII, a 770-kilobase synthetic yeast chromosome. 623–632. Science, 355, eaaf4791. 135. Lauring,A.S., Frydman,J. and Andino,R. (2013) The role of 114. Wu,Y., Li,B.-Z., Zhao,M., Mitchell,L.A., Xie,Z.-X., Lin,Q.-H., mutational robustness in RNA virus evolution. Nat. Rev. Microbiol., Wang,X., Xiao,W.-H., Wang,Y., Zhou,X. et al. (2017) Bug mapping 11, 327–336. and fitness testing of chemically synthesized chromosome X. 136. Moratorio,G., Henningsson,R., Barbezange,C., Carrau,L., Science, 355, eaaf4706. Borderı́a,A. V., Blanc,H., Beaucourt,S., Poirier,E.Z., Vallet,T., 115. Xie,Z.-X., Li,B.-Z., Mitchell,L.A., Wu,Y., Qi,X., Jin,Z., Jia,B., Boussier,J. et al. (2017) Attenuation of RNA viruses by redirecting Wang,X., Zeng,B.-X., Liu,H.-M. et al. (2017) “Perfect” designer their evolution in sequence space. Nat. Microbiol., 2, 17088. Nucleic Acids Research, 2019, Vol. 47, No. 20 10519 137. Nevot,M., Jordan-Paiz,A., Martrus,G., Andrés,C., Garcı́a-Cehic,D., 146. Kang,M., Lu,Y., Chen,S. and Tian,F. (2018) Harnessing the power Gregori,J., Franco,S., Quer,J. and Martinez,M.A. (2018) HIV-1 of an expanded genetic code toward next-generation protease evolvability is affected by synonymous nucleotide recoding. biopharmaceuticals. Curr. Opin. Chem. Biol., 46, 123–129. J. Virol., 92, e00777-18. 147. Lin,S., Yan,H., Li,L., Yang,M., Peng,B., Chen,S., Li,W. and 138. Song,Y., Liu,Y., Ward,C.B., Mueller,S., Futcher,B., Skiena,S., Chen,P.R. (2013) Site-specific engineering of chemical Paul,A.V. and Wimmer,E. (2012) Identification of two functionally functionalities on the surface of live hepatitis D virus. Angew. Chem. redundant RNA elements in the coding sequence of poliovirus using Int. Ed. Engl., 52, 13970–13974. computer-generated design. Proc. Natl. Acad. Sci. U.S.A., 109, 148. Si,L., Xu,H., Zhou,X., Zhang,Z., Tian,Z., Wang,Y., Wu,Y., 14301–14307. Zhang,B., Niu,Z., Zhang,C. et al. (2016) Generation of influenza A 139. Song,Y., Gorbatsevych,O., Liu,Y., Mugavero,J., Shen,S.H., viruses as live but replication-incompetent virus vaccines. Science, Ward,C.B., Asare,E., Jiang,P., Paul,A. V, Mueller,S. et al. (2017) 354, 1170–1173. Limits of variation, specific infectivity, and genome packaging of 149. Wang,N., Li,Y., Niu,W., Sun,M., Cerny,R., Li,Q. and Guo,J. (2014) massively recoded poliovirus genomes. Proc. Natl. Acad. Sci. Construction of a live-attenuated HIV-1 vaccine through genetic U.S.A., 114, E8731–E8740. code expansion. Angew. Chem. Int. Ed. Engl., 53, 4867–4871. 140. Malim,M.H., Hauber,J., Le,S.Y., Maizel,J.V. and Cullen,B.R. (1989) 150. Loughran,G., Chou,M.Y., Ivanov,I.P., Jungreis,I., Kellis,M., The HIV-1 rev trans-activator acts through a structured target Kiran,A.M., Baranov,P.V. and Atkins,J.F. (2014) Evidence of sequence to activate nuclear export of unspliced viral mRNA. efficient stop codon readthrough in four mammalian genes. Nucleic Nature, 338, 254–257. Acids Res., 42, 8928–8938. 141. Shin,Y.C., Bischof,G.F., Lauer,W.A. and Desrosiers,R.C. (2015) 151. Fan,R.L., Valkenburg,S.A., Wong,C.K., Li,O.T., Nicholls,J.M., Importance of codon usage for the temporal regulation of viral gene Rabadan,R., Peiris,J.S. and Poon,L.L. (2015) Generation of live expression. Proc. Natl. Acad. Sci. U.S.A., 112, 14030–14035. attenuated influenza virus by using codon usage bias. J. Virol., 89, 142. Chowdhury,S., Castro,S., Coker,C., Hinchliffe,T.E., Arpaia,N. and 10762–10773. Danino,T. (2019) Programmable bacteria induce durable tumor 152. Meng,J., Lee,S., Hotard,A.L. and Moore,M.L. (2014) Refining the regression and systemic antitumor immunity. Nat. Med., 25, balance of attenuation and immunogenicity of respiratory syncytial 1057–1063. virus by targeted codon deoptimization of virulence genes. MBio, 5, 143. Mauro,V.P. and Chappell,S.A. (2014) A critical analysis of codon e01704-14. optimization in human therapeutics. Trends Mol. Med., 20, 604–613. 153. Sitaraman,V., Hearing,P., Ward,C.B., Gnatenko,D. V, Wimmer,E., 144. Minor,P.D. (1993) Attenuation and reversion of the Sabin vaccine Mueller,S., Skiena,S. and Bahou,W.F. (2011) Computationally strains of poliovirus. Dev. Biol. Stand., 78, 17–26. designed adeno-associated virus (AAV) Rep 78 is efficiently 145. Blount,B.A. and Ellis,T. (2019) Construction of an Escherichia coli maintained within an adenovirus vector. Proc. Natl. Acad. Sci. genome with fewer codons sets records. Nature, 569, 492–494. U.S.A., 108, 14294–14299.