Document [original]

Xenomicrobiology: a roadmap for genetic code

engineering

Carlos G. Acevedo-Rocha

1,2

and Nediljko Budisa

Biosyntia ApS, 2970 Hørsholm, Denmark.

Novo Nordisk Foundation Center for Biosustainability,

Technical University of Denmark, 2970 Hørsholm,

Denmark.

Department of Chemistry, Technical University Berlin,

M€

uller-Breslau-Str. 10, Berlin 10623, Germany.

Summary

Biology is an analytical and informational science

that is becoming increasingly dependent on chemi-

cal synthesis. One example is the high-throughput

and low-cost synthesis of DNA, which is a foundation

for the research ﬁeld of synthetic biology (SB). The

aim of SB is to provide biotechnological solutions to

health, energy and environmental issues as well as

unsustainable manufacturing processes in the frame

of naturally existing chemical building blocks. Xenobi-

ology (XB) goes a step further by implementing non-

natural building blocks in living cells. In this context,

genetic code engineering respectively enables the re-

design of genes/genomes and proteins/proteomes

with non-canonical nucleic (XNAs) and amino (ncAAs)

acids. Besides studying information ﬂow and evolu-

tionary innovation in living systems, XB allows the

development of new-to-nature therapeutic proteins/

peptides, new biocatalysts for potential applications

in synthetic organic chemistry and biocontainment

strategies for enhanced biosafety. In this perspective,

we provide a brief history and evolution of the genetic

code in the context of XB. We then discuss the latest

efforts and challenges ahead for engineering the

genetic code with focus on substitutions and addi-

tions of ncAAs as well as standard amino acid reduc-

tions. Finally, we present a roadmap for the directed

evolution of artiﬁcial microbes for emancipating rare

sense codons that could be used to introduce novel

building blocks. The development of such xenomi-

croorganisms endowed with a ‘genetic ﬁrewall’will

also allow to study and understand the relation

between code evolution and horizontal gene transfer.

Xenobiology research directions

Biology is an analytical and informational science that is

becoming increasingly dependent on chemical synthesis.

The high-throughput and low-cost synthesis of DNA, for

example, is the foundation of the ﬁeld of synthetic biol-

ogy (SB). The aim of SB is to provide urgent biotechno-

logical-based sustainable solutions to problems in the

health, energy and environmental sectors (Acevedo-

Rocha, 2016). This sort of biotechnology is performed

under the frame of naturally existing chemical building

blocks. Compared with the rich methods of synthetic

organic chemistry (Walsh et al., 2005), however, the

diversity of chemistries used by natural organisms is sur-

prisingly narrow; thus, limiting a wide range of useful

and chemically diverse biotransformations.

To address this issue, the ﬁeld of xenobiology (XB)

aims to endow biological systems with artiﬁcial chemistry

absent in natural organisms (Schmidt, 2010). XB is a

highly diverse area, which two main research directions

(Fig. 1). One involves the design and synthesis of alter-

native nucleic acids (xenonucleic acids, XNAs) based

upon new base pairs, sugars and modiﬁed backbones

(Benner and Sismour, 2005; Pinheiro and Holliger,

2012). The second major area concerns engineering the

genetic code of proteins and proteomes with non-canoni-

cal amino acids (ncAAs) (Bacher et al., 2004; Budisa,

2004).

Introducing new functional members of XNA into either

DNA or RNA and ncAAs into proteins breaks the univer-

sality of the genetic code of living organisms, alienating

these from other natural life forms (Budisa, 2014). This

alienation gives access to possible orthogonal life

because both living systems cannot share information

with each other, i.e. a ‘genetic ﬁrewall’is built (Acevedo-

Rocha and Budisa, 2011). Furthermore, these synthetic

organisms are not able to grow if they are not supplied

with synthetic nutrients (trophic containment) (Marliere,

2009). Combination of a genetic ﬁrewall with an ‘alien

ersatz’increases the degree of biocontainment (Herde-

wijn and Marliere, 2009).

Notably, the difference between SB and XB is that the

former generates genetically modiﬁed organisms (GMOs)

Received 12 July, 2016; accepted 12 July, 2016.

*For correspondence. E-mail: [email protected].de;

Tel. +49 30 314 28960; Fax: +49 30 314 28 279.

Microbial Biotechnology (2016) 9(5), 666–676

doi:10.1111/1751-7915.12398

Funding Information

No funding information provided.

ª2016 The Authors. Microbial Biotechnology published by John Wiley & Sons Ltd and Society for Applied Microbiology.

This is an open access article under the terms of the Creative Commons Attribution License, which permits use, distribution and

reproduction in any medium, provided the original work is properly cited.

bs_bs_banner

based on natural chemical building blocks [e.g. canonical

amino acids (cAAs)], whereas the latter aims at creating

chemically modiﬁed organisms (CMOs) in which the

chemical building blocks are artiﬁcial (Acevedo-Rocha,

2016). The likelihood of spreading genetic material from

GMOs/CMOs into the environment is still underestimated

and not well studied (Schafer et al., 2011). Thus, biocon-

tainment of GMOs/CMOs is an active research area for

enhancing biosafety in XB (Schmidt and de Lorenzo,

2016), among many other SB-based research endeavours

(Moe-Behrens et al., 2013). The idea of ‘biocontainment’

is of particular interest, as it represents a tool for under-

standing the relationship between the evolution of the

genetic code and horizontal gene transfer. This area is

well-debated in the ﬁeld of molecular evolution (Gogarten

and Townsend, 2005).

In this perspective, we brieﬂy discuss the origin and

evolution of the genetic code from the point of view of XB.

We then describe the various strategies for engineering

the genetic code, including the use of amino acid aux-

otrophic strains, suppression of stop codons, reassign-

ment of sense codons and reduction of the standard

amino acid repertoire. We also highlight the efforts on

metabolic, genome and strain engineering for improving

efforts for engineering the genetic code. Finally, we

discuss the importance and challenges ahead for the

directed evolution of xenomicroorganisms towards the

understanding of evolutionary innovation and for enhanc-

ing biocontainment strategies.

Origin and features of the genetic code

The genetic code is universal in all three domains of life

eukaryotes, bacteria and archaea (Woese et al., 1990). It

allows the transmission of genetic information stored in

DNA from one generation to the next one. Transcription of

protein-encoding genes into RNA enables the translation

machinery to convert this information into proteins, which

are the executors of the genetic information. A reliable

and accurate translation of the linear RNA-sequence into

a functional protein is ensured by tRNAs. These adaptor

molecules read the mRNA in a three-base letter sequence

code, i.e. in triplets (Crick et al., 1961).

The chemistry of the universal translational appara-

tus is highly standardized: DNA consists of four differ-

ent bases (A, T, G and C) and there are 4

=64

possible triplet combinations for encoding 20 cAAs. An

important feature of the genetic code is its codon

degeneracy, i.e., the genetic code is redundant

because several triplets can code the same amino

acid (Lagerkvist, 1978). Brieﬂy, there are 61 sense

codons coding for amino acids and 3 stop codon sig-

nals (UAA: ochre, UGA: opal and UAG: amber) that

are responsible for terminating protein biosynthesis at

the ribosome (Fig. 2). The assignment of one triplet to

one amino acid is not random: the degeneracy of the

genetic code enables an organism to choose between

abundant and rare codons. The number of triplets is

often directly proportional to the abundance of the

Fig. 1. Experimental approaches in xenobiology to re-design the ﬂow of genetic information. Left: the central dogma of molecular biology is

shown. DNA is replicated and transmitted to the descendants or transcribed into RNA. Translation subsequently gives rise to proteins, making

up the functional proteome. Recent advances show that the introduction of artiﬁcial base pairs as XNA can also replicate and proliferate (Maly-

shev et al., 2014). To date, however, this information cannot be transcribed or translated in vivo. Incorporation of ncAAs into the proteome was

shown to be feasible if evolutionary pressure is applied (Hoesl et al., 2015). Right: Methods and levels of the genetic code re-design that

enables the partial or full re-design of informational ﬂow in biology according to Budisa (2014).

ª2016 The Authors. Microbial Biotechnology published by John Wiley & Sons Ltd and Society for Applied Microbiology, Microbial

Biotechnology,9, 666–676

Biocontainment and evolutionary innovation of artiﬁcial microbes 667

amino acid in the proteome. For example, the amino

acid tryptophan, which is relatively rare (present in

approx. 1.5% in the proteome of Escherichia coli), is

only encoded by the single UGG codon, whereas argi-

nine (around 5.5% distribution in the E. coli proteome)

is decoded by six triplets (Lajoie et al., 2013).

In E. coli, out of the 61 sense codons, two are unique

(methionine and tryptophan) while the rest are degener-

ate, with amino acids being encoded by as little as two

and as much as six triplets (Fig. 2). In the latter case, sev-

eral of the codons are rare (e.g., arginine) and allow an

organism to control the expression rate of proteins, thus

providing a selective advantage when adapting to a

changing environment (Subramaniam et al., 2013a). For

example, protein translation of abundant and rare codons

seems to be equally fast when E. coli grows in medium rich

of amino acids. In nutrient-deﬁcient media, by contrast,

translation velocity of rare codons is substantially

decreased, suggesting that the intracellular amino acid

concentration is important for efﬁcient reading of rare

codons and not the isoacceptor tRNA abundance (Li et al.,

2012; Subramaniam et al., 2013b; Ling et al., 2015).

An unresolved question is why the 20 cAAs were

selected by evolution even though several other non-pro-

teinogenic amino acids were available in the ‘primordial

soup’(Kvenvolden et al., 1971) including norvaline and

norleucine? Special cAAs can be added to the standard

amino acid repertoire. For instance, many organisms

use the 21st (selenocysteine, SeCys, or Sec) and 22nd

(pyrrolysine, Pyl) amino acids (Fig. 2) for adaptive pur-

poses (Chambers et al., 1986; Srinivasan et al., 2002).

The evolution of the genetic code is a historical pro-

cess and we will most probably always remain agnostic

about its origin and sequence of events that lead to the

current codon–amino acid associations. Upon the eluci-

dation of the genetic code in 1964 (Nirenberg and Leder,

1964), Crick postulated a ‘frozen accident theory’which

assumes that the association of particular amino acids to

their codons was accidental until the point where any

further member expansion would be lethal to cells due to

functional protein destabilization (Crick, 1968). This con-

cept has been questioned for many times during the last

decades; the plasticity and evolvability of the standard

genetic code is postulated to be plausible, as supported

by the ﬁnding of distinct alternative genetic codes in

prokaryotes (Oba et al., 1991), eukaryotic nuclear gen-

omes (Sugita and Nakase, 1999) and mitochondrial gen-

omes (Inagaki et al., 1998). Additional widely accepted

concepts about the origin of the genetic code include

stereochemical (Pelc, 1965), adaptive (Woese, 1965)

and co-evolution (Wong, 1975, 2005) theory.

Before it was ‘frozen’, the early genetic code under-

went evolution and expansion, which is encompassed by

two main theories about codon reassignment. Figure 3

summarizes them and explains possible evolutionary

mechanisms for codon reassignment to other amino

Fig. 2. The genetic code structure in the RNA format in a radial representation. Chart from inside to outside: a triplet of mRNA (50?30)is

assigned to one of the 20 canonical amino acids or a stop codon. The natural expansion of the genetic code of selenocysteine (SeC) at opal

and pyrrolysine (Pyl) at the amber and opal stop codons is depicted (Courtesy provided by Dr Stefan Oehm) (Oehm, 2016).

ª2016 The Authors. Microbial Biotechnology published by John Wiley & Sons Ltd and Society for Applied Microbiology, Microbial

Biotechnology,9, 666–676

668 C. G. Acevedo-Rocha and N. Budisa

acids: codon capture (Osawa et al., 1992) and ambigu-

ous intermediate (Schultz and Yarus, 1994) theory. The

codon capture theory postulates that genomic composi-

tion changes due to GC pressure could impose alterna-

tive codon usage mainly due to changes at the third

base position of a codon. Owing to this pressure, a

potential codon can disappear, making a particular cog-

nate tRNA obsolete. Reappearance of the codon due to

random drift would lead to its recognition by another

tRNA (that could also charge another amino acid) lead-

ing to codon reassignment, i.e., change the identity of

the codon. Conversely, the ambiguous intermediate the-

ory postulates an appearance of mutant tRNA in compe-

tition with the cognate tRNA. In the course of selective

pressure, the latter might lose the race against the

mutant tRNA, hence giving rise to another amino acid

meaning [see Fig. 3 and (Budisa, 2006a)].

Code and metabolic engineering of auxotrophic

strains

Genetic code engineering aims to use ncAAs as building

blocks in proteins. It exploits the ﬂexibility of the compo-

nents involved in protein translation mainly aminoacyl-

tRNA synthetases (aaRSs) and tRNAs (Fig. 4). While

many aaRSs have evolved mechanisms to differentiate

between naturally abundant and structurally related

amino acids (especially metabolic intermediates) (Sch-

midt and Schimmel, 1994), most ncAA that can be

incorporated with this approach do not occur in natural

environments. In the laboratory, however, a codon can

be read ambiguously by mischarging tRNAs (i.e. a cog-

nate tRNA can be charged with two different yet similar

substrates). For example, methionyl-tRNA synthetase

charges its cognate tRNA

MetCAT

with a variety of

methionine (Met) analogues, homologues and surrogates

into heterologous-expressed proteins (Wiltschi et al.,

2009). The best know example is selenomethionine

(SeMet), which can almost quantitatively replace Met in

proteins and proteomes in suitable auxotrophic strains

(Cowie and Cohen, 1957; Budisa et al., 1995). We and

others have shown many examples of potential applica-

tions of such an approach for Met analogues and

beyond during the last two decades (Mohammadi et al.,

2001; Link and Tirrell, 2005; Acevedo-Rocha et al.,

2013; Bohlke and Budisa, 2014; Piotrowski et al., 2015).

In another example, the incorporation of norleucine (Nle)

into a lipase allow the engineering of a highly active

‘cold-wash’enzyme (Hoesl et al., 2011).

Till date, more than 50 ncAAs (mostly Met, Trp, Tyr,

Phe, Pro, Arg, Lys analogues) (Budisa, 2006b) (Fig. 4A)

have been incorporated into proteins using this approach.

Unfortunately, most of these experiments are limited to

individual xeno-proteins overexpressed from auxotrophic

strains. Although this is an important research area that

can give rise to potential applications in biocatalysis, it is

important to consider the costs of supplying expensive

ncAAs into a big fermenter. For this reason, genetic code

Fig. 3. The codon capture and ambiguous intermediate theory for codon reassignment in the historical development of the genetic code. While

in codon capture (left) the endogenous tRNA disappears due to the removal of its designated codon, in ambiguous intermediate theory (right)

the exogenous tRNA is in direct competition with the cognate one. In the end, both theories end up in a reassigned codon. Figure modiﬁed

according Santos et al. (2004) and Budisa (2006a).

ª2016 The Authors. Microbial Biotechnology published by John Wiley & Sons Ltd and Society for Applied Microbiology, Microbial

Biotechnology,9, 666–676

Biocontainment and evolutionary innovation of artiﬁcial microbes 669

engineering can be combined with metabolic engineering.

For example, we implemented a biosynthetic route in for

the biosynthesis of the ncAA azido homoalanine (Aha) by

expressing in trans O-acetylhomoserine sulfhydrylase

(cgOAHSS) from Corynebacterium glutamicum in the

Met-auxotroph E. coli strain B834 while supplying the

strain with O-acetyl-homoserine and inorganic sodium

azide (NaN

) as precursors (Ma et al., 2014). This engi-

neering effort resulted in an approach for a more econom-

ical production of Aha-labelled proteins, which can be

selectively modiﬁed via ‘click chemistry’with different

molecules for protein immobilization or labelling (De

Simone et al., 2016). Likewise, the group of Wiltschi

recently afforded a strain able to produce Nle-labelled

proteins in sufﬁcient economically amounts (Anderhuber

et al., 2016). This kind of experiments will become more

important if applications of ncAAs in biocatalysis are likely

to succeed. Ideally, the precursors of the ncAA should be

biosynthesized starting from a cheap carbon source such

as glucose. We envision more synergies between meta-

bolic and genetic code engineering in the years to come.

Code engineering by stop codon suppression

In contrast to the ambiguous decoding of sense codons

by ncAAs as explained above, the meaning of a stop

codon can be suppressed by introducing engineered

orthogonal pairs (Fig. 4B). The most common pairs are

the Methanocaldococcus jannaschii TyrRS:tRNA

TyrCUA

pair, and PylRS:tRNA

PylCUA

-pairs from Methanosarcina

barkeri and Methanosarcina mazei. This stop codon sup-

pression (SCS) approach allows the site-speciﬁc incor-

poration of ncAAs into a target protein, which is usually

overexpressed (Hoesl and Budisa, 2012). Because alter-

ation of the protein structure is minimal, it is possible to

study protein function with an unprecedented level of

accuracy. For example, it is possible to study protein

conformation by F€

orster resonance energy transfer, fol-

low intracellular protein localization (using ﬂuorescent

ncAAs as probes) or determine protein–protein interac-

tions by using cross-linking ncAAs (Lang et al., 2015).

The SCS methodologies have matured in the last

years, offering researchers with a plethora of ncAAs to

(A)

(B)

Fig. 4. Flow chart presentations of two basic in vivo approaches for incorporating non-canonical amino acids into proteins. (A) Auxotrophy-

based selective pressure incorporation (SPI) method exploits the endogenous translational system to load an isostructural analogue onto a

canonical tRNA leading to residue-speciﬁc incorporation, i.e. cAA?ncAA substitution. (B) Stop codon suppression (SCS) usually endows the

host with a new orthogonal aaRS:tRNA pair in charge of introducing a given ncAA or more at one or more stop codons. For more details see

Hoesl and Budisa (2012).

ª2016 The Authors. Microbial Biotechnology published by John Wiley & Sons Ltd and Society for Applied Microbiology, Microbial

Biotechnology,9, 666–676

670 C. G. Acevedo-Rocha and N. Budisa

incorporate site-speciﬁcally (Wright et al., 2016). How-

ever, the SCS is usually limited because the release fac-

tors (RFs), which are responsible for terminating protein

biosynthesis at the ribosome, compete with the tRNAs,

resulting in a mixture of labelled and truncated protein.

Although the efﬁciency of ncAA incorporation can be

improved with RF-free engineered strains (Amiram et al.,

2015), additional ‘context effect’issues persist, which is

still the most probably cause of obtaining low protein

yields. Recently, Ignatova and others provided a more

precise experimental description of these phenomena.

The insertion of a stop codon in a particular sequence

can cause large deviations from natural mRNA folding

energy. Furthermore, it can affect the binding site interac-

tion strength within the ribosome (see e.g. Gorochowski

et al., 2015). In the years to come, we expect to obtain

better in silico models that could predict what the most

suitable sites for inserting stop codons are. Hence, pro-

moting synergies between computational chemistry and

code engineering is an important research endeavour.

Code engineering by sense codon reassignment

Codon reassignment mainly includes the removal of the

endogenous translation components like tRNA and

aaRS, synonymous codon swaps throughout the chro-

mosome or essential genes (Lajoie et al., 2015). For

example, the groups of Isaacs (Rovner et al., 2015) and

Church (Mandell et al., 2015) engineered the genetic

code of E. coli C321.DA (which lacks amber stop

codons) by re-introducing amber stop codons into essen-

tial genes (devising a sort of experimental ‘codon cap-

ture’approach). Because the UAG codon lacks any

meaning in that distinct strain, these otherwise lethal

stop codons have to be rescued with M. jannaschii

TyrRS:tRNA

TyrCUA

for p-azido-L-phenylalanine (pAzF) or

biphenyl-L-alanine (bipA) (Xie et al., 2007) respectively.

As the UAG codon has no meaning in strain C321.DA,

both studies converted (‘captured’) this codon into a

sense one by supplying orthogonal pairs. Although both

experiments differ in their strategy and their execution,

they yield the same result: a GMO that cannot grow

without supplying the ncAA, therefore becoming trophi-

cally biocontained. Those synthetic auxotrophic strains

show less than 6 910

12

escape mutants per colony

forming unit in various media. Most other techniques for

biocontainment purposes yield higher escape frequen-

cies and suffer from the possibility of cross-feeding of

metabolites produced from other organisms (Schmidt

and de Lorenzo, 2016). Both ncAAs used (pAzF and

bipA) are not available in natural environments. How-

ever, the synthetic auxotrophy only relies on the orthogo-

nal translation system. This system might be stable

under laboratory conditions, but there are many places

where the system could lose its orthogonality, e.g.,

mutating TyrRS to accept any other cAA, introduction of

a natural amber suppressors (supE, supD, supF, supZ,

etc.) or simply reverting the amber sites back to their

original meaning. Thus, long-term evolution experiments

of strain C321.DA in more diverse and rich media are

needed to ensure that it can be biocontained without

reversing its genotype.

While ambiguous decoding and SCS approaches are

always in competition with the original codon meaning

and function of the corresponding amino acid, sense

codon reassignment aims to completely remove the pre-

existent functional information of a certain codon. For

example, isoleucine can be coded by codon AUA, but

there is no tRNA

IleUAU

gene in E. coli because the third

base of the tRNA

MetUAC

is modiﬁed by tRNA

Ile

-lysidine

synthetase (TilS) to lysidine, yielding tRNA

IleLAU

that is

read by wobbling at the third base position (Bohlke and

Budisa, 2014). We generated a tilS knockout strain (this

mutation is lethal to the cell leaving more than 5000

codons unassigned) by using a rescue plasmid express-

ing the Mycoplasma mobile IleRS:tRNA

IleUAU

pair to sus-

tain cell viability. The rescue plasmid can be easily

removed in E. coli DtilS and can be exploited to apply

evolutionary pressure to reassign the Ile codon AUA to

another cAA (Doring and Marliere, 1998) or even ncAA.

Another strategy was reported by Sakamoto and co-

workers (Mukai et al., 2015) who reassigned the rare

AGG codon of arginine (Arg) to L-homoarginine (hArg)

and L-N6-(1-iminoethyl)lysine (L-NIL) by using PylRS:

tRNA

PylCUA

-pair. Cells can survive the deletion of Arg

rare codons (argW and argU) only if the evolved PylRS-

system is co-expressed and AGG positions in most

essential genes are exchanged by synonymous codons

(to avoid detrimental effects of AGG-reading as hArg or

L-NIL). Although analytical data were not presented (e.g.

MS/MS experiments in essential genes to verify Arg-

replacement), the exchange of Arg at remaining AGG

positions most probably took place given that both

ncAAs used (hArg, L-NIL) are similar to the cAA Arg

(‘similar replaces similar’) (Moroder and Budisa, 2010).

Code engineering via reduction of the standard

amino acid repertoire

Widely accepted co-evolution theory about the origin of

the genetic code postulates its evolution from a simpli-

ﬁed set of amino acid and extension to recent form with

the expansion of cellular metabolism (Wong, 1981).

Therefore, an experimental model with the reduced

genetic code could provide some insight into evolution-

ary trajectories of amino acids. This was accomplished

by the construction of a functional enzyme using a nine-

amino acid alphabet (Walter et al., 2005). Namely,

ª2016 The Authors. Microbial Biotechnology published by John Wiley & Sons Ltd and Society for Applied Microbiology, Microbial

Biotechnology,9, 666–676

Biocontainment and evolutionary innovation of artiﬁcial microbes 671

Hilvert and colleagues demonstrated that E. coli lacking

the aroQ gene for chorismate mutase is unable to grow

in media where all sources for tyrosine (Tyr) and pheny-

lalanine (Phe) are withdrawn. This synthetic auxotrophy

can be exploited to screen a library of artiﬁcial choris-

mate mutase analogues, in which aroQ is only com-

posed from a nine-amino acid alphabet. It was found

that even this reduced set of amino acids can lead to a

functional, metabolically competent and near to wild-type

growing strain. However, the reduction led to a restricted

diversity ending up with a destabilization of overall pro-

tein structure. These results indicate the need of more

amino acids for robust proteins and explain the evolu-

tionary force towards the increased diversity. A more

general approach towards this issue was reported by

Marli

ere and co-workers: they showed the possibility of

removing Trp from the genetic code (Pezo et al., 2013).

Taking advantage of the conserved and essential his-

tidine (His) residue in the active site of yeast transketo-

lase gene tktY, this position was mutated to the Trp-

codon UGG in trans in a suitably conﬁgured E. coli

MG1655 DtktAB strain. Supplying this strain with the

missense tRNA

HisCCA

, the UGG codon in the essential

TktY is read as a His leading to a functional gene, which

is the driving force to keep the missense tRNA. Subse-

quently, the system was proliferated in a continuous

pulse feeding regime, yielding 30 times higher mis-incor-

poration of His at Trp positions in proteins. This experi-

ment suggests the possibility of engineering the genetic

code applying evolutionary pressure using a reduced

amino acid alphabet. This is an important research area

whose potentials have not yet been extensively studied

(Oehm, 2016).

Code engineering via experimental evolution

Over more than three billion years, the standard genetic

standard code has not undergone any signiﬁcant

changes besides some local alterations (Knight et al.,

2001). Therefore, an utmost goal of XB is to introduce

new biochemical building blocks into an organism at

both the genomic and proteomic level. One option is by

performing experimental evolution, which deals with the

long-term cultivation of organisms under controlled con-

ditions. Doubtless, adaptive evolution experiments with

suitably designed metabolic prototypes could accelerate

the propagation, manipulation and analysis of organisms

in a controlled environment. Furthermore, such evolution-

ary adaptation is the most expedient route to generate

artiﬁcial biocontained microbes. Because microbes have

fast generation times and large population sizes in a rea-

sonable and manageable dimension of cultivation space,

they are ideal candidates for such experimental setups.

Additionally, populations can be easily frozen for later

analysis (‘fossils’). Adaptive evolution experiments have

been designed to cultivate E. coli to adapt on glucose,

glycerol and lactate minimal medium, high temperature

irradiation as well as high salt or ethanol concentration

(Blount et al., 2012; Wiser et al., 2013).

Already in the late 1950s, Cowie and Cohen (1957)

demonstrated that Met auxotrophic E. coli grow in Met-

free medium when SeMet is supplied. This proteome

wide exchange of Met by producing functionally essential

enzymes for cell proliferation illustrates how the amino

acid composition of a proteome can be inﬂuenced by

external pressure. Another report from 1963 by Rennert

and Anker claims that E. coli cells able to use 50,50,50-tri-

ﬂuoroleucine (TFL) as an analogue for leucine (Leu) can

be isolated from a suitably designed cultivation experi-

ment (Rennert and Anker, 1963). Wong (1983) was able

to adapt a Trp deﬁcient Bacillus subtilis with some few

single-cell transfers to grow on 4-ﬂuoro-L-tryptophan (4F-

Trp) as Trp source. Using the mutagenic agent it was

possible to obtain a derivative exhibiting preference

towards the ncAA. Nevertheless, those old experiments

were not accompanied by precise analytic measure-

ments to quantify the extent of ncAA incorporation (con-

taminations of cAA from media preparation cannot be

excluded –most likely, those sources of cAA were ele-

mental for the observed growth). Moreover, those experi-

ments were conducted in ‘rich’medium in which

standard amino acids, vitamins and nucleobases were

supplemented. Therefore, only a small portion of the cel-

lular machinery had to adapt to the incorporation of the

ncAA. Bacher and Ellington (2001) adapted the E. coli

strain C600 DtrpE towards 4F-Trp by serially transferring

it to minimal medium while gradually increasing the 4F-

Trp-to-Trp ratio. Unfortunately, the commercial source of

4F-Trp was contaminated with 0.03% Trp (the authors

were not able to withdraw all cAA-source from the

media). Yet they found a strain tolerating high levels of

4F-Trp to a level at which it was toxic for the parental

strain. Due to the lack of next generation sequencing,

only a few relevant genes were sequenced but these

were unchanged (tnaA, trpR, mtr,trpT, bla,gapA), while

others like trpS, aroP and tyrR exhibited single nucleo-

tide polymorphisms. Expression of trpS and aroP mutant

proteins in trans in the ancestral strain led to a growth

advantage in the presence of 4F-Trp. However, those

mutations are not sufﬁcient to let the C600 strain incor-

porate the ncAA at high levels (deep analytics were not

performed). Nevertheless, this experiment improved our

understanding of serial evolution experiments to change

the chemical composition of the proteome.

Based on the experiences gathered on earlier long-

term E. coli cultivation experiments by the groups of

Lenski, Ellington, Bacher and Wong, we also set out to

adapt this bacterium towards an ncAA. In particular, our

ª2016 The Authors. Microbial Biotechnology published by John Wiley & Sons Ltd and Society for Applied Microbiology, Microbial

Biotechnology,9, 666–676

672 C. G. Acevedo-Rocha and N. Budisa

group completely substituted Trp by the analogue

Thienopyrrole-alanine ([3,2]Tpa) in a Trp auxotrophic

strain (Hoesl et al., 2015). To avoid contamination of

Trp traces in commercial ncAA preparations, the strains

were engineered to produce the surrogate by using the

enzyme tryptophan synthase (TrpBA), which conden-

sates the indole analogue [3,2]Thienopyrrole with serine

yielding [3,2]Tpa (Budisa et al., 2001). Rigorous and

high-precision analytics conﬁrmed the almost completely

quantitative replacement of Trp by the surrogate at

20 899 sense codons in E. coli W3110 genome.

Hence, the engineered strain can be deﬁned as ‘trophi-

cally reassigned’(i.e., the meaning of a codon is rede-

ﬁned throughout the whole proteome). However,

supplementing these cells with Trp reverses them to

‘natural’ones given that they still favour the incorpora-

tion of the canonical building block. To achieve a Trp-

independent reassignment (i.e. ‘real’codon reassign-

ment) at all UGG codons across E. coli’s genome –an

experimental strategy for biocontainment still needs to

be developed and executed (Acevedo-Rocha and

Schulze-Makuch, 2015).

Roadmap to equip xenomicroorganisms with a

‘genetic ﬁrewall’

In the previous sections, we highlighted various

approaches for genetic code engineering. The most pop-

ular one is perhaps the genome-wide replacements of

stop or rare sense codons with synonymous alternatives

in order to design genomes with radically altered genetic

codes, which can be used for biocontainment purposes.

However, owing to the complexity of experimental gen-

ome re-design, it is difﬁcult to avoid changes that impair

cellular ﬁtness, which creates fragile strains. Among

other features, rare codons ﬁne tune translation to facili-

tate the biogenesis of the encoded protein and their syn-

onymous replacements can lethally impact mRNA

secondary structure or ribosome binding sites which is

expected to result in dramatically decreased ﬁtness dur-

ing the genome assembly process (Plotkin and Kudla,

2011). Furthermore, widely used orthogonal pairs are

not as active and accurate as natural aaRSs (Nehring

et al., 2012). An alternative strategy for experimental

genetic code evolution relies on the global substitution

of cAAs with ncAAs (or an addition of ncAAs) assisted

with simple metabolic optimization following an ‘ambigu-

ous intermediate’engineering concept. In other words,

we propose a novel strategy that relies on liberation of

rare sense codons of the genetic code (i.e. ‘codon

emancipation’) from their natural decoding functions

(Bohlke and Budisa, 2014). This approach consists of

long-term cultivation of bacterial strains coupled with the

design of orthogonal pairs for sense codon decoding. In

particular, directed evolution of bacteria should be

designed to enforce ambiguous decoding of target

codons using genetic selection. In this system, viable

mutants with improved ﬁtness towards missense sup-

pression can be selected from large bacterial popula-

tions that can be automatically cultivated in suitably

designed turbidostat devices. Once ‘emancipation’is

performed, full codon reassignment can be achieved

with suitably designed orthogonal pairs. Codon emanci-

pation will likely induce compensatory adaptive muta-

tions that will yield robust descendants tolerant to

disruptive amino acid substitutions in response to

codons targeted for reassignment. We envision this

strategy as a promising experimental road to achieve

sense codon reassignment –the ultimate prerequisite to

achieve stable ‘biocontainment’as an emergent feature

of xenomicroorganisms equipped with a ‘genetic ﬁre-

wall’.

Conclusions

In summary, genetic code engineering with ncAA by

using amino acid auxotrophic strains, SCS and sense

codon reassignment has provided invaluable tools to

study accurately protein function as well as many possi-

ble applications in biocatalysis. Nevertheless, to fully

realize the power of synthetic organic chemistry in bio-

logical systems, we envision synergies with metabolic,

genome and strain engineering in the next years to

come. In particular, we believe that the experimental

evolution of strains with ncAAs will allow the develop-

ment of ‘genetic ﬁrewall’that can be used for enhanced

biocontainment and for studying horizontal gene transfer.

Additionally, these efforts could allow the production of

new-to-nature therapeutic proteins and diversiﬁcation of

difﬁcult-to-synthesize antimicrobial compounds for ﬁght-

ing against ‘super’pathogens (McGann et al., 2016).

Yet the most fascinating aspect of XB is perhaps to

understand the genotype–phenotype changes that lead

to artiﬁcial evolutionary innovation. To what extent is

innovation possible? What emergent properties are

going to appear? Will these help us to re-examine the

origin of the genetic code and life itself? During evolu-

tion, the choice of the basic building blocks of life was

dictated by (i) the need for speciﬁc biological functions;

(ii) the abundance of elements and precursors in past

habitats on earth and (iii) the nature of existing solvent

(s) and available energy sources in the prebiotic environ-

ment (Budisa, 2014). Thus far, there are no detailed

studies on proteomics and metabolomics of engineered

xenomicrobes, let alone systems biology models that

could integrate the knowledge from such efforts. In

2020, we expect to have such kind of analysis in order

to have a clearer picture of the possibilities that artiﬁcial

ª2016 The Authors. Microbial Biotechnology published by John Wiley & Sons Ltd and Society for Applied Microbiology, Microbial

Biotechnology,9, 666–676

Biocontainment and evolutionary innovation of artiﬁcial microbes 673

evolution can offer. We believe that expanding the reper-

toire of the genetic code beyond the canonical 20 (+2)

amino acids will not only change the chemical makeup

of life, but also allow to re-examine our understanding

and current concepts about the origin of the genetic

code and life itself.

Acknowledgements

We thank Dr Stefan Oehm and Ms. Federica Agostini for

help in the preparation of ﬁgures and scientiﬁc discus-

sions during manuscript preparation.

Conﬂict of interest

None declared.

References

Acevedo-Rocha, C. (2016) The synthetic nature of biology.

In Ambivalences of Creating Life Societal and Philosophi-

cal Dimensions of Synthetic Biology. Kristin Hagen, Mar-

gret Engelhard, Georg Toepfer (eds.). Switzerland:

Springer, pp. 9–53.

Acevedo-Rocha, C.G., and Budisa, N. (2011) On the road

towards chemically modiﬁed organisms endowed with a

genetic ﬁrewall. Angew Chem Int Ed Engl 50: 6960–6962.

Acevedo-Rocha, C.G., and Schulze-Makuch, D. (2015) How

many biochemistries are available to build a cell? Chem-

BioChem 16: 2137–2139.

Acevedo-Rocha, C.G., Hoesl, M.G., Nehring, S., Royter, M.,

Wolschner, C., Wiltschi, B., et al. (2013) Non-canonical

amino acids as a useful synthetic biological tool for

lipase-catalysed reactions in hostile environments. Catal

Sci Technol 3: 1198–1201.

Amiram, M., Haimovich, A.D., Fan, C.G., Wang, Y.S., Aerni,

H.R., Ntai, I., et al. (2015) Evolution of translation

machinery in recoded bacteria enables multi-site incorpo-

ration of nonstandard amino acids. Nat Biotechnol 33:

1272–1275.

Anderhuber, N., Fladischer, P., Gruber-Khadjawi, M., Stried-

ner, G., Mairhofer, J., and Wiltschi, B. (2016) High-level

biosynthesis of norleucine in E. coli for the economic

labeling of proteins. J Bacteriol (in press). doi: 10.1016/

j.jbiotec.2016.04.033

Bacher, J.M., and Ellington, A.D. (2001) Selection and char-

acterization of Escherichia coli variants capable of growth

on an otherwise toxic tryptophan analogue. J Bacteriol

183, 5414–5425.

Bacher, J.M., Hughes, R.A., Wong, J.T.F., and Ellington,

A.D. (2004) Evolving new genetic codes. Trends Ecol

Evol 19: 69–75.

Benner, S.A., and Sismour, A.M. (2005) Synthetic biology.

Nat Rev Genet 6: 533–543.

Blount, Z.D., Barrick, J.E., Davidson, C.J., and Lenski, R.E.

(2012) Genomic analysis of a key innovation in an experi-

mental Escherichia coli population. Nature 489: 513–518.

Bohlke, N., and Budisa, N. (2014) Sense codon emancipation

for proteome-wide incorporation of noncanonical amino

acids: rare isoleucine codon AUA as a target for genetic

code expansion. FEMS Microbiol Lett 351: 133–144.

Budisa, N. (2004) Prolegomena to future experimental efforts

on genetic code engineering by expanding its amino acid

repertoire. Angew Chem Int Ed Engl 43: 6426–6463.

Budisa, N. (2006a) Amino acids and codons –code organi-

sation and protein structure. In Engineering the Genetic

Code. Nediljko Budisa (ed.). Weinheim, Germany:

WILEY-VHC, pp. 66–85.

Budisa, N. (2006b) Reprogramming the cellular translation

machinery. In Engineering the Genetic Code. Nediljko

Budisa (ed.). Weinheim, Germany: WILEY-VHC, pp. 90–

165.

Budisa, N. (2014) Xenobiology, new-to-nature synthetic cells

and genetic ﬁrewall. Curr Org Chem 18: 936–943.

Budisa, N., Steipe, B., Demange, P., Eckerskorn, C., Keller-

mann, J., and Huber, R. (1995) High-level biosynthetic

substitution of methionine in proteins by Its analogs 2-

aminohexanoic acid, selenomethionine, telluromethionine

and ethionine in Escherichia coli.Eur J Biochem 230:

788–796.

Budisa, N., Alefelder, S., Bae, J.H., Golbik, R., Minks, C.,

Huber, R., and Moroder, L. (2001) Proteins with beta-

(thienopyrrolyl)alanines as alternative chromophores and

pharmaceutically active amino acids. Protein Sci 10:

1281–1292.

Chambers, I., Frampton, J., Goldfarb, P., Affara, N., McBain,

W., and Harrison, P.R. (1986) The structure of the mouse

glutathione peroxidase gene: the selenocysteine in the

active site is encoded by the ‘termination’codon, TGA.

EMBO J 5: 1221–1227.

Cowie, D.B., and Cohen, G.N. (1957) Biosynthesis by sof

active altered proteins containing selenium instead of sul-

fur. Biochim Biophys Acta 26: 252–261.

Crick, F.H. (1968) The origin of the genetic code. J Mol Biol

38: 367–379.

Crick, F.H., Barnett, L., Brenner, S., and Watts-Tobin, R.J.

(1961) General nature of the genetic code for proteins.

Nature 192: 1227–1232.

De Simone, A., Hoesl, M.G., and Budisa, N. (2016) Engineer-

ing lipases with an expanded genetic code. In Applied Bio-

catalysis –From Fundamental Science to Industrial

Applications. Liese, A.H., Kettling, U. and Antranikian, G.

(eds). Weinheim, Germany: WILEY-VHC, pp. 3–12.

Doring, V., and Marliere, P. (1998) Reassigning cysteine in

the genetic code of Escherichia coli.Genetics 150:

543–551.

Gogarten, J.P., and Townsend, J.P. (2005) Horizontal gene

transfer, genome innovation and evolution. Nat Rev

Microbiol 3: 679–687.

Gorochowski, T.E., Ignatova, Z., Bovenberg, R.A.L., and

Roubos, J.A. (2015) Trade-offs between tRNA abundance

and mRNA secondary structure support smoothing of

translation elongation rate. Nucleic Acids Res 43: 3022–

3032.

Herdewijn, P., and Marliere, P. (2009) Toward safe geneti-

cally modiﬁed organisms through the chemical diversiﬁca-

tion of nucleic acids. Chem Biodivers 6: 791–808.

ª2016 The Authors. Microbial Biotechnology published by John Wiley & Sons Ltd and Society for Applied Microbiology, Microbial

Biotechnology,9, 666–676

674 C. G. Acevedo-Rocha and N. Budisa

Hoesl, M.G., and Budisa, N. (2012) Recent advances in

genetic code engineering in Escherichia coli.Curr Opin

Biotechnol 23: 751–757.

Hoesl, M.G., Acevedo-Rocha, C.G., Nehring, S., Royter, M.,

Wolschner, C., Wiltschi, B., et al. (2011) Lipase con-

geners designed by genetic code engineering. Chem-

CatChem 3: 213–221.

Hoesl, M.G., Oehm, S., Durkin, P., Darmon, E., Peil, L.,

Aerni, H.-R., et al. (2015) Chemical s. Angew Chem Int

Ed Engl 54: 10030–10034.

Inagaki, Y., Ehara, M., Watanabe, K.I., Hayashi-Ishimaru,

Y., and Ohama, T. (1998) Directionally evolving genetic

code: the UGA codon from stop to tryptophan in mito-

chondria. J Mol Evol 47: 378–384.

Knight, R.D., Freeland, S.J., and Landweber, L.F. (2001)

Rewiring the keyboard: evolvability of the genetic code.

Nat Rev Genet 2: 49–58.

Kvenvolden, K.A., Lawless, J.G., and Ponnamperuma, C.

(1971) Nonprotein Amino Acids in the Murchison Mete-

orite. Proc Natl Acad Sci USA 68: 486–490.

Lagerkvist, U. (1978) “Two out of three”: an alternative

method for codon reading. Proc Natl Acad Sci USA 75:

1759–1762.

Lajoie, M.J., Kosuri, S., Mosberg, J.A., Gregg, C.J., Zhang,

D., and Church, G.M. (2013) Probing the limits of genetic

recoding in essential genes. Science 342: 361–363.

Lajoie, M.J., S€

oll, D., and Church, G.M. (2015) Overcoming

challenges in engineering the genetic code. J Mol Biol

428: 1004–1021.

Lang, K., Davis, L., and Chin, J.W. (2015) Genetic encoding

of unnatural amino acids for labeling proteins. Methods

Mol Biol 1266: 217–228.

Li, G.W., Oh, E., and Weissman, J.S. (2012) The anti-

Shine-Dalgarno sequence drives translational pausing

and codon choice in bacteria. Nature 484: 538–541.

Ling, J., O’Donoghue, P., and S€

oll, D. (2015) Genetic code

ﬂexibility in microorganisms: novel mechanisms and

impact on physiology. Nat Rev Microbiol 13: 707–721.

Link, A.J., and Tirrell, D.A. (2005) Reassignment of sense

codons in vivo. Methods 36: 291–298.

Ma, Y., Biava, H., Contestabile, R., Budisa, N., and Di

Salvo, M.L. (2014) Coupling bioorthogonal chemistries

with artiﬁcial metabolism: Intracellular biosynthesis of azi-

dohomoalanine and its incorporation into recombinant

proteins. Molecules 19: 1004–1022.

Malyshev, D.A., Dhami, K., Lavergne, T., Chen, T., Dai, N.,

Foster, J.M., et al. (2014) A semi-synthetic organism with

an expanded genetic alphabet. Nature 509: 385–388.

Mandell, D.J., Lajoie, M.J., Mee, M.T., Takeuchi, R., Kuznet-

sov, G., Norville, J.E., et al. (2015) Biocontainment of

genetically modiﬁed organisms by synthetic protein

design. Nature 518: 55–60.

Marliere, P. (2009) The farther, the safer: a manifesto for

securely navigating synthetic species away from the old

living world. Syst Synth Biol 3: 77–84.

McGann, P., Bunin, J.L., Snesrud, E., Singh, S., Maybank,

R., Ong, A.C., et al. (2016) Real time application of whole

genome sequencing for outbreak investigation –What is

an achievable turnaround time? Diagn Microbiol Infect Dis

85: 277–282.

Moe-Behrens, G.G., Davis, R., and Haynes, K.A. (2013)

Preparing synthetic biology for the world. Front Microbiol

4: 10.

Mohammadi, F., Prentice, G.A., and Merrill, A.R. (2001) Pro-

tein-protein interaction using tryptophan analogues: novel

spectroscopic probes for toxin-elongation factor-2 interac-

tions. Biochemistry 40: 10273–10283.

Moroder, L., and Budisa, N. (2010) Synthetic biology of pro-

tein folding. ChemPhysChem 11: 1181–1187.

Mukai, T., Hoshi, H., Ohtake, K., Takahashi, M., Yamaguchi,

A., Hayashi, A., et al. (2015) Highly reproductive Escheri-

chia coli cells with no speciﬁc assignment to the UAG

codon. Sci Rep 5: 9699.

Nehring, S., Budisa, N., and Wiltschi, B. (2012) Perfor-

mance analysis of orthogonal pairs designed for an

expanded eukaryotic genetic code. PLoS ONE 7: 13.

Nirenberg, M., and Leder, P. (1964) RNA codewords

and protein synthesis. The effect of trinucleotides upon

the binding of sRNA to ribosomes. Science 145:

1399–1407.

Oba, T., Andachi, Y., Muto, A., and Osawa, S. (1991) CGG:

an unassigned or nonsense codon in Mycoplasma capri-

colum.Proc Natl Acad Sci USA 88: 921–925.

Oehm, S. (2016) Adaptation of E. coli towards tryptophan

analogue usage. PhD Thesis. Berlin, Germany: Depart-

ment of Chemistry of TU Berlin.

Osawa, S., Jukes, T.H., Watanabe, K., and Muto, A. (1992)

Recent evidence for evolution of the genetic code. Micro-

biol Rev 56: 229–264.

Pelc, S.R. (1965) Correlation between coding-triplets and

amino-acids. Nature 207: 597–599.

Pezo, V., Louis, D., Gu

erineau, V., Le Caer, J.-P., Gaillon,

L., Mutzel, R., and Marli

ere, P. (2013) A metabolic proto-

type for eliminating tryptophan from the genetic code. Sci

Rep 3: 1359.

Pinheiro, V.B., and Holliger, P. (2012) The XNA world: pro-

gress towards replication and evolution of synthetic

genetic polymers. Curr Opin Chem Biol 16: 245–252.

Piotrowski, C., Ihling, C.H., and Sinz, A. (2015) Extending

the cross-linking/mass spectrometry strategy: facile incor-

poration of photo-activatable amino acids into the model

protein calmodulin in Escherichia coli cells. Methods 89:

121–127.

Plotkin, J.B., and Kudla, G. (2011) Synonymous but not the

same: the causes and consequences of codon bias. Nat

Rev Genet 12: 32–42.

Rennert, O., and Anker, H. (1963) On the incorporation of

5’,5’,5’-triﬂuoroleucine into proteins of E. coli.Biochem-

istry 2: 471–476.

Rovner, A.J., Haimovich, A.D., Katz, S.R., Li, Z., Grome,

M.W., Gassaway, B.M., et al. (2015) Recoded organisms

engineered to depend on synthetic amino acids. Nature

518: 89–93.

Santos, M., Moura, G., Massey, S.E., and Tuite, M.F.

(2004) Driving change: the evolution of alternative genetic

codes. Trends Genet 20: 95–102.

Schafer, M.G., Ross, A.A., Londo, J.P., Burdick, C.A., Lee,

E.H., Travers, S.E., et al. (2011) The establishment of

genetically engineered canola populations in the U.S.

PLoS ONE 6, 8–11.

ª2016 The Authors. Microbial Biotechnology published by John Wiley & Sons Ltd and Society for Applied Microbiology, Microbial

Biotechnology,9, 666–676

Biocontainment and evolutionary innovation of artiﬁcial microbes 675

Schmidt, M. (2010) Xenobiology: a new form of life as the

ultimate biosafety tool. BioEssays 32: 322–331.

Schmidt, M., and de Lorenzo, V. (2016) Synthetic bugs on

the loose: containment options for deeply engineered (mi-

cro)organisms. Curr Opin Biotechnol 38: 90–96.

Schmidt, E., and Schimmel, P. (1994) Mutational isolation of

a sieve for editing in a transfer RNA synthetase. Science

264: 265–267.

Schultz, D.W., and Yarus, M. (1994) Transfer RNA mutation

and the malleability of the genetic code. J Mol Biol 235:

1377–1380.

Srinivasan, G., James, C.M., and Krzycki, J.A. (2002) Pyrro-

lysine encoded by UAG in Archaea: charging of a UAG-

decoding specialized tRNA. Science 296: 1459–1462.

Subramaniam, A.R., Deloughery, A., Bradshaw, N., Chen,

Y., O’Shea, E., Losick, R., and Chai, Y. (2013a) A serine

sensor for multicellularity in a bacterium. eLife 2, e01501.

Subramaniam, A.R., Pan, T., and Cluzel, P. (2013b) Envi-

ronmental perturbations lift the degeneracy of the genetic

code to regulate protein levels in bacteria. Proc Natl Acad

Sci USA 110: 2419–2424.

Sugita, T., and Nakase, T. (1999) Non-universal usage of

the leucine CUG codon and the molecular phylogeny of

the genus Candida. Syst Appl Microbiol 22: 79–86.

Walsh, C.T., Garneau-Tsodikova, S., and Gatto, G.J. (2005)

Protein posttranslational modiﬁcations: the chemistry of

proteome diversiﬁcations. Angew Chem Int Ed Engl 44:

7342–7372.

Walter, K.U., Vamvaca, K., and Hilvert, D. (2005) An active

enzyme constructed from a 9-amino acid alphabet. J Biol

Chem 280: 37742–37746.

Wiltschi, B., Merkel, L., and Budisa, N. (2009) Fine tuning

the N-terminal residue excision with methionine ana-

logues. ChemBioChem 10: 217–220.

Wiser, M.J., Ribeck, N., and Lenski, R.E. (2013) Long-term

dynamics of adaptation in asexual populations. Science

342: 1364–1367.

Woese, C.R. (1965) On the evolution of the genetic code.

Proc Natl Acad Sci USA 54: 1546–1552.

Woese, C.R., Kandler, O., and Wheelis, M.L. (1990)

Towards a natural system of organisms: proposal for the

domains Archaea, Bacteria, and Eucarya. Proc Natl Acad

Sci USA 87: 4576–4579.

Wong, J. (1975) A co-evolution theory of the genetic code.

Proc Natl Acad Sci USA 72: 1909–1912.

Wong, J. (1981) Coevolution of genetic code and amino

acid biosynthesis. TIBS 6: 33–36.

Wong, J. (1983) Membership mutation of the genetic code:

loss of ﬁtness by tryptophan. Proc Natl Acad Sci USA 80:

6303–6306.

Wong, J. (2005) Coevolution theory of the genetic code at

age thirty. BioEssays 27: 416–425.

Wright, T.H., Vallee, M.R.J., and Davis, B.G. (2016) From

chemical mutagenesis to post-expression mutagenesis: a

50 year odyssey. Angew Chem Int Ed Engl 55: 5896–

5903.

Xie, J., Liu, W., and Schultz, P.G. (2007) A genetically

encoded bidentate, metal-binding amino acid. Angew

Chem Int Ed Engl 46: 9239–9242.

ª2016 The Authors. Microbial Biotechnology published by John Wiley & Sons Ltd and Society for Applied Microbiology, Microbial

Biotechnology,9, 666–676

676 C. G. Acevedo-Rocha and N. Budisa