development of a cell-free protein synthesis system for

Review

Development of a cell-free protein synthesis systemfor practical use

By Yaeta ENDO*1,†

(Edited by Shigekazu NAGATA, M.J.A.)

Abstract: Conventional cell-free protein synthesis systems had been the major platform tostudy the mechanism behind translating genetic information into proteins, as proven in the centraldogma of molecular biology. Albeit being powerful research tools, most of the in vitro methods atthe time failed to produce enough protein for practical use. Tremendous efforts were being madeto overcome the limitations of in vitro translation systems, though mostly with limited success.While great knowledge was accumulated on the translation mechanism and ribosome structure,researchers rationalized that it may be impossible to fully reconstitute such a complex molecularprocess in a test tube. This review will examine how we have solved the difficulties holding backprogress. Our newly developed cell-free protein synthesis system is based on wheat embryos and hasmany excellent characteristics, in addition to its high translation activity and robustness. Combinedwith other novel elementary technologies, we have established cell-free protein synthesis systems forpractical use in research and applied sciences.

Keywords: ribosome, ribotoxin, wheat embryos, cell-free protein synthesis

Introduction

The concept of cell-free protein synthesis is basedon the idea of carrying out translation reactions in atest tube using the translation machinery extractedfrom living cells. In 1961, Nirenberg and colleaguesreported mRNA-dependent polypeptide synthesisusing an extract from Escherichia coli.1) This cell-free system enabled deciphering of the genetic code.Furthermore, the method revealed a full picture ofthe translation mechanism and opened the door tomolecular biology as we know it today. Because thetranslation apparatus in a cell-free system retainsthe activity, speed, and accuracy of polypeptidesynthesis present in living cells, such a system canbe used as an excellent and essential analytical tool.However, the productivity of these conventionalsystems was too low, and protein products wereonly detected by incorporation of radio-labeled aminoacids in tracer experiments. Higher protein yields

were thought to be unachievable because; 1) lowproductivity had been generally observed in allreported systems; and 2) it was impossible to workon a high throughput for analyzing hundreds oftranscribed elements in a test tube. Therefore, theidea of robust protein synthesis in a cell-free systemwas essentially dismissed. As a graduate student inthe 1970s, while seeking to understand the structureand function of the ribosome, I had often thoughtabout potential ways to overcome the above-men-tioned limitations of classical cell-free systems.

Ribosomes are known as large molecular ma-chines (a complex of 980 proteins (r-proteins) and 4RNAs (rRNA) in eukaryotes, and 955 r-proteins and3 rRNAs in prokaryotes) which catalyze polypeptidesynthesis with accuracy and high speed. However,the structure and function of ribosomes were largelyelusive in those days. Nomura and colleagues madegreat strides in ribosome research using theirreconstitution technique to determine which moiety,the proteins or the RNA, might contribute themajor function in protein synthesis. They successfullyreconstituted fully active hybrid 30S subunits byassembling E. coli 30S r-proteins and 16S rRNA fromphylogenetically distant Bacillus subtilis. Moreover,

*1 Ehime Prefectural University of Health Sciences, Tobe-cho, Iyo-gun, Ehime, Japan.

† Correspondence should be addressed: Y. Endo, EhimePrefectural University of Health Sciences, 543, Takoda, Tobe-cho,Iyo-gun, Ehime 791-2101, Japan (e-mail: [email protected]).

Proc. Jpn. Acad., Ser. B 97 (2021)No. 5] 261

doi: 10.2183/pjab.97.015©2021 The Japan Academy

http://dx.doi.org/10.2183/pjab.97.015

they found that introduction of multiple nicks intothe rRNA in E. coli ribosomes did not affect anyactivity. Based on their findings, it was concludedthat the key players in the ribosomes were the r-proteins rather than rRNA.2) Although lines ofgenetic and biochemical evidence for mechanisms ofaction for some antibiotics had suggested that theirtargets were on rRNA, these antibiotics were thoughtto inhibit ribosomal function by disturbing thescaffold formed by the rRNA chain leading toimproper arrangement of the functional r-proteins.With this idea, Wittmann and colleagues madeconsiderable efforts to find r-proteins that carry outany elementary function of ribosomes.3) Around thattime, we had started investigating the mechanisms ofaction of ribotoxins, known as ribosomal inactivatingproteins (RIPs), in order to understand the structureand function of ribosomes. One of such RIP, ,-Sarcin(SAR), is a cytotoxic protein originally found as ananticancer protein isolated from cultured Aspergillusgiganteous. SAR was known to inhibit proteinsynthesis enzymatically by inactivating ribosomefunctions during peptide formation, and we eluci-dated the mode of action of this RIP. We found that

SAR is a ribonuclease (RNase) that specificallyhydrolyzes a single phosphodiester bond betweenG4325 and A4326 in an evolutionally conservedsequence at the 3B-end of eukaryotic 28S rRNA (orin 23S rRNA in bacteria). This produced a 394-nucleotide fragment (,-fragment) with a hydroxylgroup at the 5B-end of the downstream rRNA(Fig. 1A).4)–6) We also described another type ofribotoxin, ricin A-chain, found in the seeds of castorbeans, Ricinus communis. Despite the beautifulflower it comes from, the protein is known as adeadly toxin. The A-chain of ricin (RCA) enzymati-cally inactivates the elongation-factor (EF-1 and EF-2)-dependent transpeptidation reaction of ribosomes,whereas the B-chain is a lectin that delivers RCAinto animal cells where the inactivation of ribosomestakes place. Although the molecular mechanism ofaction of RCA had been unknown for a long time,ricin had been used for assassinations because of itsdeadly toxic action, where tens of micrograms wereenough to kill a human. We investigated the mode ofaction of RCA and successfully solved the mystery ofthe mechanism.7)–11) The target site of RCA wasdetermined to be A4324 in the rRNA, adjacent to the

Fig. 1. Modes of action of two families of RIPs. (A) ,-Sarcin catalyzes the hydrolysis of one phosphodiester bond (indicated with anarrow) between G4325 and A4326 in rat 28S rRNA and produces an ,-fragment (3B side of 28S, 394 nucleotides). (B) Chemistry ofRNGase- and RALyase-catalyzed reactions. RNGase hydrolyzes an adjacent N-glycosidic bond at A4324 in the same loop as shownin (A), releasing adenine from ribosomes. RALyase in wheat embryos cleaves the phosphate backbone at the apurinic site throughits lyase activity, yielding a fragment of 395 nucleotides containing the 3B-end with a 5B-terminal guanosine-5B-phosphate residue.A simple depurination assay (RNGase activity) can be done by means of gel-electrophoresis to see the formation of the fragment(395 nucleotides) after the treatment of extracted RNA with acid-aniline.7),8)

Y. ENDO [Vol. 97,262

specific cleavage site by SAR, and it was shown thatthe protein catalyzes the hydrolysis of the N-glycosidic bond between adenine and ribose. Thecatalytic mechanism was unique, and we named thisenzyme an “RNA N-glycosidase” or RNGase for short(Fig. 1B). Meanwhile, we also found that the A-chains of both Shiga toxin and Vero toxin, had anidentical mode of action and functioned in the sameway as RCA.12),13) Thus far, two distinct families ofribotoxins irreversibly inactivate the peptide elonga-tion activity of ribosomes solely by hydrolyzing asingle phosphodiester bond (SAR) or performing asingle deadenylation (RCA) at one contiguous siteamong 97,000 nucleotides in eukaryotic ribosomes(or 95,000 nucleotides of in E. coli ribosomes).Kinetic parameters of RCA activity on ribosomeswere analyzed, with apparent Km and Kcat deter-mined to be 92 µM and 92,000 s!1, respectively.9) Ithas been shown that about 15 ribosomes on averageare functioning on a single mRNA forming polysomesin translationally active cells. Upon inactivation ofpeptide elongation by a RCA, freezing of ribosomestakes place on polysomes, which physically blocks themovement of all following intact ribosomes. Thus, intheory, the depurination (inactivation) of about 7%of the entire ribosome population in a cell is enoughto completely terminate protein synthesis, thusleading to cell-death, which turned out to be thereason for the extreme toxicity of ricin to humans. Asshown in Fig. 2A, both SAR and RCA recognition-sites are in the same small stem-loop on the specific2-dimensional (2D) Watson–Crick structure modelconsisting of 12 nucleotides (indicated in red). Thissequence is universally conserved among organisms.We named this small, but known to be the longestconserved nucleotide sequence in rRNA, the “Sarcin-Ricin Loop” (SRL). Our findings suggested that theSRL has a structurally and functionally importantrole in ribosomes.14) The 3D structure of SRL inthe naked form and in ribosome particles wasdetermined using NMR15),16) and X-ray crystallog-raphy.17) Figure 2B shows the structure of a syn-thetic 29-mer representing the eukaryotic SRL, asdetermined using X-ray crystallography. The adeno-sine residue, the target nucleotide of the RNGase,sticks out from the helix, making it more accessible.A matching GAGA, tetra-nucleotide loop, and fivebase pairs formed practically the same 3B structurein the structure of 50S ribosomes, as solved by X-raycrystallography (Fig. 2B and D), thus confirming ourresults obtained with a synthetic oligonucleotide.Noller and others pointed out that the SRL plays

a central role for two elongation factors, therebytriggering the dynamic structure of the 50S unit infunctioning 70S ribosomes.17),18) Together with thesefindings and others, it is now widely accepted that“The ribosome is a ribozyme”.19)–21) However, in thosedays, we were asking, “What is the biologicalsignificance of RNGase in nature?” To obtain cluesto answer this question, we first surveyed thedistribution of RNGase among plants22)–25) using asimple acid-aniline assay, which we had developed.From the initial screening, we demonstrated that theRIPs and some of the so-called antiviral proteins werealmost ubiquitously distributed within the plantkingdom, including grains. However, it should benoted, most of these single-chain ribotoxins lackedthe B chain, so they are not toxic to animals. Usingimmunoelectron microscopy, RNGases, includingthe pokeweed antiviral protein, were found to belocalized in the extracellular space of the cotyle-don.26)–28) Moreover, we found another novel enzyme,a ribosomal RNA apurinic site-specific lyase(RALyase), in wheat embryos, which cleaved aphosphodiester bond at the apurinic site within 28SrRNA in ribosomes, producing a 395 nucleotidefragment leaving a 5B phosphate in the rRNA(Fig. 1B).29) Deadenylated ribosomes have littleactivity in the two elongation factor-dependenttranslation reaction, as mentioned above. However,they showed full peptide synthesis activity withribosomes fully possessing poly-phenylalanine syn-thesis activity when measured in artificial conditionsusing high magnesium ion concentrations in theabsence of the above-mentioned elongation factors.Those ribosomes, however, were totally inactivatedonce the chain scission was introduced byRALyase,13),29) just as SAR-treated ribosomes were.These finding suggested that there might be acooperative role between tritin (wheat RNGaselater), localized in the outer space of cells, andRALyase inside the embryos to achieve the termi-nation of ribosome function. Considering the func-tions of RNGase and RALyase, their cellular local-ization, and the antiviral action altogether, wespeculated that the system might be functioning asa plant self-defense mechanism through killing theirown ribosomes to stop viruses. In this scenario, oncethe cuticle of the plant cell wall is damaged by insectbites, RNGase flows into the cytoplasm of embryoniccells and cleaves the N-glycosidic bond of ribosomes.The depurinated ribosomes are then completelyterminated by RALyase in embryonic cells, thuspreventing the proliferation of viruses. Hence, we

Cell-free protein synthesis systemNo. 5] 263

speculated that the programmed suicidal systemmight have evolved together with the translationsystem not only in the plant kingdom but may alsobe used as a general defense system. Such antiviralself-defensive mechanisms are well known in animalsbased on research on interferon.30)

Having this defense mechanism in mind, wereconsidered the reasons for the common observationof “the low productivity of classical cell-free systems”.We hypothesized that the low productivity ofreported systems may, in part, be caused by anartificial event during the homogenization of theraw materials, which would induce a malfunction ofthe self-defensive system, resulting in killing of thetranslation machinery. The most well-studied plant-based cell-free protein synthesis system was thewheat germ cell-free system originally reported-byRoberts and colleagues. This wheat germ cell-freesystem had the highest translation activity with aprolonged reaction time of an hour, whereas theE. coli system worked only for 10–15min.31) Weexpected that the wheat system might be furtherimproved by eliminating the self-defensive systemand other endosperm-originating contaminants. Dor-mant embryos are known to store a large amount ofthe translation machinery for upcoming germination,although tritin, a single-chain RIP,32) was alsoreported in wheat seeds and leaves. Initially, it wasreported in a functional analysis that wheat embry-onic ribosomes are resistant to tritin.32)–34) However,we directly demonstrated that tritin indeed possessesa depurination activity at A4324 of the 28S rRNA ofembryonic ribosomes. Furthermore, we were able tosolve this problem by simply washing crude embryosbefore the homogenization step during wheat extractpreparation. The effect was far more successful thanwe expected, and this became the first essential pointtoward a practical cell-free expression system describ-ed in the later section.

The second crucial point we addressed was therobustness of the reaction, a mode of the translationreaction set up to meet the needs of a practical use.Spirin and colleagues had been working on improvingthe productivity of cell-free systems mainly byfocusing on the reaction mode. They believed thatthe short reaction time, which was generally observedfor in vitro translation reactions, must be caused bythe consumption of substrates (amino acids, ATP,GTP) during the reaction. Therefore, they intro-duced a continuous-exchange cell-free (CECF) reac-tion method to replace the commonly used regularbatch-wise reaction, which significantly improved the

capability of their new system.35) However, theiroriginal CECF method, using an ultra-filtrationmembrane, was not very reproducible mainly dueto clogging of the membrane, and thus their methodcould not be fully utilized. However, it was fortunatethat we were able to reproduce a part of theirprotocol.36) Later, we applied the two differentreaction procedures, using the dialysis reaction modefor large-scale production and a bilayer reactionmode for the parallel production of small proteinquantities.

Inevitably, the quality of synthesized proteins iscritical, and the ability to produce a properly foldedactive protein has been the third essential need for apractical system. The folding mechanisms for newlysynthesized proteins are entirely different in prokar-yotes and eukaryotes, which seemed to be coupledwith the speed of peptide bond formation, where therate in E. coli is about 10 times faster than that ineukaryotic cells. It was assumed that protein foldingin bacterial cells mainly takes place post-translation-ally with the help of chaperones, whereas ineukaryotes co-translational folding plays a key rolein protein folding with the help of partner moleculessuch as prosthetic groups.37),38) As a result, theE. coli-based expression methods, whether in vivo orin vitro, undoubtedly have limitations in synthesiz-ing eukaryotic proteins in their active forms, espe-cially for larger polypeptides commonly found ineukaryotes. In fact, it is quite often observed that theproducts are segregated as an insoluble inclusionbody, when eukaryotic proteins are expressed inE. coli cells. In this respect, a wheat-based expressionsystem can better address co-translational foldingbecause of its lower translation speed. In addition,our wheat cell-free technology may solve otherproblems, such as biohazardous, ethical issues, andcost-effectiveness. It was fortunate for us that wehad chosen the wheat platform for improving in vitrotranslation systems.

Hereafter, this review will introduce the elemen-tary technologies that we have developed; 1)preparation of a highly efficient and robust extractfrom wheat embryos, 2) an optimized mRNA design,3) customized DNA template design for in vitrotranscription, 4) a platform for a simple CECF modetranslation reaction, and 5) the world’s first fullyautomated protein synthesizers were invented com-bining all these technologies into a comprehensivesystem.


1. Preparation of an extract with a hightranslation activity from purified

wheat embryos.

(1) Preparation of a high-quality extract. Asshown in Fig. 3A, whitish powder (wheat flour) couldbe seen on the surface of crude embryo preparations,which indicated that isolated embryos from the seedsare contaminated with the remaining components ofthe endosperm. The extract was directly prepared,and cell-free translation reactions were performed inbatch-wise reactions following the protocol reportedby Robert and others.31) We began by examining theoccurrence of any depurinated ribosomes caused bycontaminating tritin using the acid-aniline assay andfound that the number of depurinated ribosomesincreased overtime, as shown by the formation of a385-nucleotide fragment on the gel (arrow in Fig. 3B,lanes 1–5). Furthermore, even before the incubation,about 7% of the ribosome population had already

been depurinated. When RNA was directly extractedfrom embryos using the guanidine isothiocyanate-phenol method, a harsh condition to denature tritinand other proteins, little formation of the aniline-induced fragment was observed, indicating that thedepurination reaction had already taken place duringthe extraction process. Because we knew that even asmall portion of depurinated ribosomes led to severeinhibition of protein synthesis, we decided to removetritin from the extract. Attempts were made toneutralize the enzyme with chemical components,including synthetic RNA aptamers that boundtightly to the RIP,39) but these approaches did not

Fig. 3. Robust protein synthesis using wheat extract preparedfrom washed wheat embryos. (A) Extracts (S30, 240A260 ODunits/ml) were prepared from unwashed and washed embryos,respectively. (B) The depurination assay used to monitor tritinactivity in unwashed or washed embryos. Translation mixtureswere constructed with extracts from unwashed or washedembryos and were incubated in a batch-wise reaction for theindicated times, and phenol-extracted RNA samples from eachtime point were treated with acid-aniline to cleave the phosphatebackbone at the depurinated site. Lanes 1–5 contain samplesafter incubation of the mixture with unwashed embryos for 0, 1,2, 3, and 4 h, and lanes 7–9 contain samples after incubation ofthe mixture with washed embryos for 0, 2, and 4 h, respectively.Lane 6 shows the reaction marker, in which a fragment wasproduced by incubation in the presence of gypsophilin, a highlyactive RNA N-glycosidase from Gypsophila elegance.28) Thearrow indicates the aniline-induced fragment of the 3B-side ofwheat 28S rRNA. (C) Protein synthesis capacities of an extractfrom washed embryos using the DHFR mRNA with a 5B-Capstructure and 100-nucleotide poly(A) tails, transcribed frompSP65 plasmid. Translation reactions were carried out in testtubes for the indicated time course. Protein synthesis wasmeasured using hot-trichloroacetic acid-insoluble radioactivityof incorporated 14C-labeled leucine. (D) The same experimentwas carried out as in (C) using an extract from unwashedembryos. Arrows indicate the time points of supplementation ofthe substrate solution.49) Copyright (2002) National Academy ofSciences, U.S.A.

Fig. 2. Sites of action of SAR and RCA in a conserved nucleotidesequence (SRL) in rRNAs and their structures. (A) The sites on2D Watson–Crick base-pair models. Arrowheads and arrowsmark RCA and SAR sites, respectively, in the universallyconserved nucleotide sequence (highlighted in red color). (B)Schematic representation of the X-ray crystal structure of asynthetic 29-mer of rat SRL.17) (C) The crystal structure ofE. coli 50S subunits and the SRL (dotted circle) is shown witha magnified view (D). Ribosomal proteins and RNA-chain,including SRL, are marked in blue and red, respectively.Copyright (1998) National Academy of Sciences, U.S.A. (forFig. 2A and B). Figure 2C and D are a gift from H. F. Noller,who kindly created the picture based on the published data-bank.20)


work. Instead, careful selection and subsequentextensive washing of isolated embryos resulted inan excellent result. The depurination assay showedan undetectable level of fragment formation in theacid-aniline assay (Fig. 3B, lanes 7–9), indicatingthat these washed embryos contained little tritin,and other contaminants had probably also beenremoved. As shown in Fig. 3C and D, the cell-freesystem prepared from washed embryos had muchhigher translation activity than the conventionalsystem.40) When programmed with an mRNAencoding dihydrofolate reductase (DHFR) as amodel, the reaction continued for at least 4 hours ina system containing 24% extract (Fig. 3C), asopposed to the original system (unwashed), whichceased activity after 1.5 hours (Fig. 3D). In thereaction containing 48% of the new extract, theinitial reaction rate was increased two-fold comparedwith that using 24% of the extract; however, thetranslation reaction stopped after 1 hour. Halting ofthe reaction (arrows) was overcome with kineticssimilar to the initial rate after supplementing freshsubstrate, including amino acids, ATP, and GTP. Incontrast, 48% of the extract from unwashed embryosshowed even lower activity than 24% of the extract,and the addition of fresh substrate after the reactionhad stopped did not have any effect (Fig. 3D). Theseobservations were consistent with the results of theabove-mentioned depurination assays, indicatingthat there was irreversible damage done to theribosomes by tritin and probably other contaminantssuch as RNase(s) from the endosperm. A hightranslation activity of the system with washedembryos was also demonstrated using a sucrosedensity gradient experiment, in which polysomeformation can be monitored. Significant numbers ofnewly formed polysomes were observed after incuba-tion for 1 hour, and a shift to heavier polysomes witha concomitant decrease in 80S monosomes was foundafter incubation for 2 hours.40) Moreover, electronmicroscopic examination revealed, for the first time,in vitro formation of circular-shaped polyribosomes,which are well known from living cells with highprotein synthesis activity.41) There is an additionalexplanation for the dramatic improvement in proteinsynthesis after washing of the embryos. Thioninsare a group of small, basic, cysteine-rich proteins,originally purified as antifungal proteins from variousplants, including wheat seeds.42) Wheat .-thionin isknown to be in the endosperm of seeds,43) andBrummer and colleagues showed that ,- and O-thionins from barley endosperm are potent inhibitors

of protein synthesis initiation in a wheat germtranslation system.44) In addition, several nucleaseshave been reported in the endosperm of seeds.45)

Thus, it was thought to be possible that washing ofthe embryos also resulted in the elimination ofthionin and ribonucleases together with tritin.

(2) Performance of the new extract. Afterestablishing a procedure for the preparation of highlyactive wheat embryo extract, we examined itspossible application for protein production. For thispurpose, we adapted a dialysis system as a simplifiedCECF mode translation reaction. With cappedmRNA encoding DHFR and with a poly(A) tail,protein synthesis worked efficiently. The reactionproceeded for up to 60 hours, yielding 4mg of activeenzyme in a 1ml reaction when supplemented withfresh mRNA every 24 hours.40) This result directlyindicated the robustness of the system and its hightranslational activity. In theory, the strategy we haveintroduced for making our system work is applicableto other cell-free systems; however, it may not be atrivial task. For example, an RNase1 deletion mutanthas been used for constructing an E. coli cell-freesystem, though there is still a non-negligible amountof RNase 1 together with other cellular RNaseactivities in the extract. Rabbit reticulocyte lysatealso contains RNase M attached to the plasmamembrane, which cannot be eliminated by anyavailable procedure. These RNases, even occurringat low concentrations in the translation mixture,bring the system a drawback to complete translationreactions for high molecular weight proteins becausethe mRNA template will always be extremelysensitive to RNases. Thus, it is hard to expect anyformation of large, circular polysomes in thesesystems. In general, to optimize the reconstitutionof any given biochemical reaction in vitro, limitingfactors must be found and removed from theartificially created systems. Hence, we carefullyoptimized each of the ingredients and the translationconditions, including the incubation temperature forthe new extract.

2. Development of other elementary technologiesfor improving the wheat germ cell-free

protein synthesis system.

(1) Designing an efficient mRNA and con-struction of a DNA template for in vitro tran-scription. The 5B-Cap (7-mG-5B-ppp-5BG) plays acrucial role in eukaryotic translation initiation, butit was initially problematic in cell-free translationsystems. When RNA templates are prepared in vitro,


the cap structure can be introduced using an RNApolymerase that incorporates the modified freedinucleotide at the 5B-end of an RNA polynucleotidechain. However, the efficiency of this incorporationwas generally low, and there was inevitably an excessof free modified nucleotide remaining in the reactionmixture. Moreover, the free modified nucleotidebinds competitively to the cap-binding protein,eIF4E, thereby inhibiting translation. Completeremoval of the free nucleotide is tedious and noteasily accomplished in the context of a high-throughput system. Furthermore, there is a verynarrow range for suitable concentrations of cappedRNA that give efficient translation; this range is bestdetermined empirically because the exact concen-tration of effectively capped RNA is difficult todetermine.40) This optimization is problematic whenproteins are to be synthesized from many genes(RNAs) in parallel. The poly(A) tail at the 3B-end ofmRNA is also a problem when preparing templatesfor in vitro translation reactions because plasmidswith long poly dT/dA sequences are unstable duringreplication in E. coli cells. To solve these problems,we developed new 5B- and 3B-UTRs (untranslatedregions) that enhance mRNA translation in theabsence of a cap structure and a poly(A) tail.

We designed novel enhancer sequences bymimicking the +-sequence from tobacco mosaicvirus, a 67-base naturally occurring translationalenhancer.46) A pool of DNA fragments composed ofrandom sequences (73 nucleotides long) was chemi-cally synthesized in the absence of dGTP in order notto create an undesirable start codon (AUG). Theserandom sequences were inserted behind an SP6promoter sequence and followed by a luciferase geneas a reporter open reading frame (ORF). The DNApool was transcribed using SP6 RNA polymerasein vitro. The resulting mRNA pool with manydifferent 5B-UTRs was combined with the new wheatgerm extract described above, and a ribosome displayexperiment was performed to select an optimalinitiation site for cap-free translation. One of theselected 73-mer sequences showed 970% translationinitiation activity compared with 5B-capped mRNA,and we named this sequence enhancer-01 of EhimeUniversity (E01).47) We next designed a 3B UTRelement to replace the poly(A) tail, because it wasknown that the poly(A) tail is dispensable fortranslation in yeast.48) By varying the 3B-UTR ofmRNAs with the E01 5B-enhancer and encoding theluciferase reporter gene, we found that translationdoes not depend on any specific sequence but solely

relates to the length of the 3B-UTR, where about1,500 nucleotides added downstream of a given geneshowed a comparable activity with that of apolyadenylated mRNA49),50) consistent with an ear-lier report on mRNA stability. Later, we learned thateven shorter 3B-UTRs of 150–500 nucleotides possessadequate template activity compared with that ofthe longer template with the cap structure presentedin Fig. 4A. Based on these findings, we constructeda new expression vector, pEU (plasmid of EhimeUniversity, Matsuyama, Japan; Fig. 4B, left) toprepare optimal templates for in vitro translation.When the DHFR gene cloned into the pEU vector(circular form) was transcribed using SP6 RNApolymerase, three major RNA products were ob-

Fig. 4. Two methods to construct DNA templates for in vitrotranscription. (A) 5B- and 3B-end structures of a typicaleukaryotic mRNA (upper row) and the newly designed one(lower) are shown. (B) The plasmid, originating from pSP65,possesses an SP6 polymerase promoter, a newly developedenhancer sequence (E01), multi-cloning sites, and a replicationsite (Ori) (left). Circular pEU carrying a luciferase gene with E01gave 3 mRNA bands when transcribed (arrows on the gel)(right). (C) A split-primer PCR method was developed for theparallel transcription of cDNAs from a cDNA library, minimiz-ing primer-dimer byproducts (left). A set of primers in which thecomplete promoter sequence was not contained in any one of theprimers but was generated only when the sequences of twoprimers were joined correctly, leading to the production of onlycomplete mRNA (arrow, right).49) Copyright (2002) NationalAcademy of Sciences, U.S.A.


tained (Fig. 4B, right). Assuming that transcriptionfails at the replication origin (Ori), the lowest band(1.5 kb, about halfway around the pEU vector)corresponded to the expected size from the E01enhancer, including the DHFR mRNA with the laternucleotide sequence of the plasmid up to the Ori. Themiddle size band (3 kb) and the upper band tran-scripts (5.5 kb) were considered to be products oftranscription reactions that went around the circularplasmid one and two times, respectively. Of note,subsequent translation assays showed that thesemRNAs had practically the same template activityas the one transcribed from a linear pEU vectoryielding one RNA product of defined length. Werecognized this result as a significant advantagebecause this may allow skipping of the linearizationstep using a restriction enzyme, which mightrecognize additional cleavage sites within the ORF.As shown in Fig. 5A, when a GFP mRNA tran-scribed from a circular pEU vector was translated ina dialysis bag with our new wheat germ extract, alarge amount of protein (9.7mg of GFP in 1ml) wasproduced.49) The translation reaction continued to

produce protein for a surprisingly long period, up to14 days in this experiment. Although supplementa-tion of mRNA was required every second day, theamount of protein produced was higher than that ofproteins prepared using endogenous protein synthesisin cellular systems. This result supported our notionthat the translation machinery itself is inherentlyrobust and stable. Thus, the pEU vector-based cell-free system is able to produce a large amount ofprotein products; however, in this form the platformis not yet suitable for high-throughput approachesin the post-genome era.

Meanwhile, we had sought to develop a methodto prepare mRNAs in a high-throughput manner fora genome-wide in vitro protein synthesis system. APCR-based method was thought to be less laboriousand more cost-effective. In the first attempt, a setof four different primers was used to obtain DNAtemplates from a cDNA library. The resultingtemplates consisted of a 5B-SP6 promoter sequence,an E01 element, followed by a given ORF. However,this primer set produced a large amount of shortDNA fragments due to primer-dimers, and the PCR

Fig. 5. Performance of a cell-free system. (A) SDS-PAGE analysis of GFP produced during reactions lasting for 14 days. mRNAproduced by transcription of circular pEU was used for the translation reaction in a dialysis membrane system and was added every48 h. A 0.1 µl aliquot of the mixture was run on a gel, and protein bands were stained with CBB. The arrow shows GFP, which appearsas a prominent band among the endogenous protein bands (lanes, 0–14 days). “st” designates a 0.5 µg equivalent of an authentic GFPband. (B) In small-scale protein synthesis, the translation solution was layered under the buffered substrate mixture. The twosolutions (usually 25 µl on the bottom and 125µl on the top in a 96 microtiter plate, the bilayer mode reaction) mix together graduallyduring incubation period, enabling continuous substrate supply and dilution of byproducts, which may otherwise suppress thetranslation reaction (left). The translation reaction continued for about 20h at 15 °C, and time-dependent synthesis of GFP anddiffusion of two layers are shown (right). (C) Five mRNAs, transcribed from PCR-generated DNA templates with designed C-termHis-tag sequences, were translated overnight. The crude reaction products (lanes with even numbers) and Ni-column purifiedproducts (lanes with odd numbers) were separated by SDS-PAGE and stained with CBB. Lane 1 contains a negative control of thecell-free synthesis reaction lacking mRNA.49) Copyright (2002) National Academy of Sciences, U.S.A.


products yielded only multiple small RNA fragmentsin SP6 polymerase driven transcription experiments.Therefore, we designed a set of primers, in which thecomplete promoter sequence was not contained inany single primer but was only generated when thesequences of two primers were joined correctly(Fig. 4C, left), leading to the production of onlycomplete mRNAs (Fig. 4C, right).49) We confirmedthe mRNA transcribed from these PCR-generatedDNA fragments, which showed a similar translationkinetic as shown in Fig. 3C for working with anexpression vector. Thus, the split-primer PCRmethod enables genome-wide construction of tran-scription templates.49),50)

(2) Development of a bilayer reaction modefor cell-free translation reactions. Although theCECF method was quite useful for producing a largeamount of protein as described above, it was notconvenient for genome-wide parallel protein produc-tion and subsequent characterization, where a rathersmall quantity of each protein is preferable in a high-throughput system. To this end, we developed asimple bilayer-based continuous translation reactionmode, which satisfied the principle of CECF reac-tions, with continuous supply of the substrates anddilution of the byproducts, which may potentiallyinhibit translation (Fig. 5B, left). A similar trans-lation mixture as used for the batch-mode and theCECF reactions was placed at the bottom of thetube, and the substrate solution containing all thenecessary ingredients was then loaded as an upperlayer with the translation mixture. During incuba-tion, diffusion of the two layers takes place, whichfollows the principle of a continuous reaction.51) Asshown in Fig. 5B (right), fluorescence of synthesizedGFP was already observed after 5min, and thereaction continued for 15 hours through slow mixingof the two layers throughout the incubation. Theyield of GFP from this experiment was estimatedto be around 1mg/ml of the bottom layer after 15hours. In Fig. 5C, parallel production and purifica-tion of synthesized proteins in the bilayer system isshown. Five mRNAs were transcribed in parallelfrom PCR-generated DNA templates, in which asequence encoding the histidine-tag was added at the3B-end of the ORF. Analysis of each sample by SDS-PAGE clearly showed discrete Coomassie brilliantblue (CBB)-stained bands of the expected molecularweights among endogenous proteins from the wheatgerm extract. Purified products obtained from asimple affinity tag purification step are shown(Fig. 5C, lanes 3, 5, 7, 9, and 11). Based on this

bilayer mode reaction using a 96-well plate, dozensof proteins can be easily prepared at once. Moreover,co-expression of multiple mRNA species was accom-plished,52),53) and several membrane protein com-plexes were prepared in an active form in vitro.54),55)

Very recently, a simple protocol to reconstitutechromatin using this bilayer wheat system wasreported, in which the mRNAs encoding the fourcore histones and chromatin assembly factors wereco-expressed in the presence of a closed circularDNA.56)

A flow diagram of our cell-free system, combin-ing the elementary technologies described above, isshown in Fig. 6. A comparison of the performance ofour system and other conventional cell-free methodsis listed in Table 1. It is worth mentioning here that,because the dormant embryos contain few amino-acid-metabolizing enzymes except for two trans-aminases,59),60) no scrambling of labeled atoms amongamino acids takes place during the translationreaction. This made it easy to prepare a stableisotope-labeled protein,49),61),62) and this method hasbeen applied in the recent advent of cellularquantitative proteomic studies.63)–66) These methodsmay provide a procedure for the systematic studyof protein structure–function relationships and aseries of applications delineating the scope of modernproteomics, including high-throughput enzymatictesting of a large number of genomic expressionproducts, high-throughput crystallization of proteins,and identification of their three-dimensional struc-ture through NMR, cryo-electron microscopy, or X-ray diffraction, rapid evolutionary design of proteins,and construction of technologies for studying pro-tein–protein interactions.

(3) Automated cell-free protein synthesizers.Combining all elementary technologies described sofar, we have developed three types of fully and semi-automated systems (Fig. 7). The Gendecorder has acapacity to synthesize 384 proteins overnight with asmall translation reaction size of 150 µl by feeding thetemplate DNAs for the transcription and translationreactions (Fig. 7A). With this technology, a humanprotein library has been constructed.67),68) TheProtemist DTII is equipped with functions fortranscription, translation, and affinity purificationof up to 6 samples in a medium reaction size of 6ml(Fig. 7B). Finally, the Protemist XE is for a large-scale protein production (Fig. 7C), equipped withtwo functions, a repeated substrate buffer exchange(intermittent exchange batch-mode reaction basedon the same principle as the CECF) and a continuous


supply of mRNA during the translation reaction.Productivity may reach 1 g per 100ml reaction, andthus is especially useful for downstream applicationsthat require high quality and quantity proteinsamples, for example for NMR spectroscopy, X-raycrystallography, animal studies, and preparation ofantigens.

3. From basic research toeducational applications

Since first published in 2002, our wheat germcell-free technologies have been used in various up-to-

date research applications such as protein structureand folding, including rather challenging membranetransporters, protein MS, protein–protein interac-tions, and others, as listed on the website of CellFreeSciences Co., Ltd. (https://www.cfsciences.com/eg/resources/reference-list). CellFree Sciences was estab-lished as a venture company of Ehime University tomake the wheat germ systems accessible to theresearch community and to support applied sciencesin industry.

Out of the many different applications of thesystem, our present research initiatives to develop a

Table 1. Comparing the performance of cell-free expression systems49)

Materials E. coli Rabbit reticulocyte Wheat embryos

Productivity (per ml) 6mg µg order 9.7mg

N-terminala N-formylmethionine Mature Mature

Folding Post-translation Co-translation Co-translation

Quality Low High High

Codon preference Tight Tight Loose

Amino acid-specific isotopic labeling Difficult — Yes

Post-translational modification No Yes Yes

Membrane protein Yes Yes Yes

Disulfide-bond formation Yes — YesaN-terminal methionine is processed by the endogenous N-terminal methionine peptidase.57),58)

1

Fig. 6. A flow diagram of protein production using the wheat cell-free system. Genes of interest are selected from DNA databases in-silico, and transcription templates can be constructed using cDNAs or chemically synthesized ORFs using the split-PCR method. Atypical DNA construct is comprised of an SP6 promoter, E01 enhancer, ORF, reporter, and/or a purification tag (SP6-Tag1-ORF-Tag2). After transcription, the solution is directly used as an input mRNA in the bilayer translation system using multi titer wells.This parallel production system not only allows higher productivity but can also be used in the optimization of the translationconditions (e.g., fine-tuning of ion concentrations, supplementation of a prosthetic group, co-translation with a partner protein(s),selection of detergents for membrane proteins, and incubation temperature) for production of high quality proteins. When one canfind a proper protein, move on to the next massive production protocol, from cloning of the construct of interest into pEU andamplifying it to produce a sufficient amount of mRNA, and the translation can be run in CECF mode under a dialysis membrane orintermittent-exchange reaction.


https://www.cfsciences.com/support/reference-list





malaria vaccine are introduced below as an example.Additionally, ongoing outreach to the communitybased on our wheat cell-free technologies is described.

(1) Development of a malaria vaccine.Malaria is a life-threatening infectious disease witha very long history of human infection. Following theemergence of anti-malaria drug-resistant parasites,anti-insecticide-resistant mosquitoes, and the re-cently emerged multidrug-resistance, malaria vaccinedevelopment is a high priority for supporting peoplein areas where malaria is endemic, and a vaccine mayalso be of great benefit for travelers from malaria-freecountries. Modern malaria vaccine developmentbegan with seminal studies in mice using irradiatedparasites.69) Decades of endeavor have taught us thatachieving this goal will be challenging. One of theproblems of malaria vaccine development lies in thepreparation of quality proteins from malaria speciesinfecting humans. It is not realistic to preparemalaria proteins from cultured parasites becausetheir hosts are only mosquitos and humans. The firstwhole genome sequence of Plasmodium falciparummalaria was released in 2002, and attempts weremade with the use of a cDNA library to produceactive proteins using recombinant DNA technology.However, recombinant DNA technology was notsuccessful for template generation, because 1) thecodon usage in the genome of the parasite is unique,hence it is essential to adjust codons to match thoseof host cells such as E. coli cells prior to production,and 2) most of the protein products were recoveredonly in insoluble inclusion bodies. In contrast, thewheat cell-free system was expected to suit the high-

throughput preparation of malaria proteins whilemaintaining high quality. Results from preliminaryexperiments exceeded expectations, including codonoptimization being unnecessary for most of themalaria genes, and the proteins were obtained mostlyin a soluble form.70) Currently, a set of malariaproteins has been prepared using the Protemist DTIIautomated system for the discovery of vaccinecandidates with regard to the three life cycle stages,prevention of infection (merozoites in the liver),prevention development of disease (trophozoite inerythrocytes), and prevention of transmission (game-tocytes in mosquitos). Very recently, marker proteinsfor the detection of infection with Plasmodium vivaxmalaria were reported that will hopefully allow thedetection of patients with dormant parasites in theliver,71) which until now has been challenging.Furthermore, several vaccine candidates to preventdisease development were discovered.72) A trans-mission-blocking vaccine is targeting the sexual stageof the parasite in mosquitos and thus blocking the lifecycle and preventing the spread of malaria throughthe community. Tsuboi and colleagues have beenworking on this dream medicine, and the project isat present at the preclinical stage and soon to becontinued with the first in human phase 1 clinicaltrial.

(2) Development of teaching materials forlocal schools. Besides various contributions to lifesciences, I have thought about how to educatestudents and the general audience on the questionof “What is life? ”. The basic idea, developed on thebasis of what I have learned through my career, isshown below.

“Where Do We Come From? What Are We?Where Are We Going? ” These questions asked byPaul Gauguin in the title of his masterpiece havetroubled philosophers for thousands of years, andas long as we live, we struggle with them. To thefundamental questions about our existence, RenéDescartes, who is commonly regarded as the father ofmodern philosophy, ascertained that “I think, there-fore I am”. This principle, which is believed to haveformed the foundation of modern science started inthe 17th century and appears simple enough torequire no further explanation. Nevertheless, I do notthink it has provided us with the ultimate answers wehave been looking for. Neither am I convinced thatphilosophical reasoning would eventually lead to ananswer to this most difficult, most intimate, yet mostremote questions that challenge human beings. Itherefore propose that we incorporate life science in

Fig. 7. Automated protein synthesizers. (A) Gendecorder. (B)Protemist DTII. (C) Protemist XE.


the foundations of education in Japan and let youngpeople search for answers to these questions throughlife science approaches. Living things are all bornfrom parent(s), inheriting their DNA, which encodesgenetic information to synthesize proteins. Wehuman beings live and conduct social activitiesthanks to a vast number of protein molecules – asmany as a million different kinds. Life is dictated bythe central dogma (general principle) of molecularbiology regarding the replication of genetic informa-tion and gene expression, which occur through threereactions in the cell, (1) DNA synthesis, (2) RNAsynthesis, and (3) protein synthesis. While reproduc-tion of reactions (1) and (2) became possible in atest tube several decades ago, the third reaction (3),namely, making protein synthesis in a test tube, wasnot possible for a long time. During that time, livingcells were the only means for preparing proteinsamples to study the mechanisms of life carried outby proteins. Then, in the year 2000 at EhimeUniversity, a system of protein synthesis extractedfrom wheat embryos was successfully adapted tocreate a practical protein synthesis method. With theadvent of this method, protein synthesis in a testtube at will became a reality for the first time. Thisnew technology has been applied for life scienceresearch, drug discovery, and medical technologydevelopment and has also been used as an R & D toolin other new biotechnology industries. My presentproposal is to expand its application to education.The central dogma proposed by Crick about 60 yearsago has taught us that life’s activities are carried outthrough chemical reactions. Although this is widelyknown and accepted in life sciences, there has been noeducational tool available to teach it to laypeople.Such tools can be readily provided by the new wheatcell-free technology. The new technology permits usto demonstrate the life processes described by thecentral dogma as they are reproduced in a test tube.These educational tools will then help young peopleto understand logically that all our activities includ-ing lofty mental functions are conducted throughchemical reactions. Our brain functions do not seemto have evolved to find answers to it through merethinking; such a capability does not seem to haveproved a favorable trait through evolution. Humansneed to try things to learn from experience. If so, whyshould we not enrich education with what life sciencehas accomplished? Visit the website (https://pim-sympo.jp/public_seminar/, in Japanese) for moredetails.

Conclusions and perspectives

We have developed a powerful cell-free proteinsynthesis system by fully using the vital power ofwheat embryos. The system has overcome variouslimitations underling the classical protein productionmethodologies, in both the cell-free and cell-basedapproaches. The wheat system converting geneticinformation into functional proteins in a test tube,solved other serious problems in protein production,namely, biohazardous and ethical issues, becausewheat seeds are food products causing no harm tohumans. These features have helped the wheatsystem to make important contributions to drive lifescience research and technical developments. Sincelast year, while preparing this review article, westarted facing the global COVID-19 pandemic, whereonce again, our technology showed its value. Our cell-free technology has enabled serological tests fordetecting COVID-19 patients.73)–75) These examplesdemonstrated for the first time that our system canbe scaled-up to find a suitable antigen and is alsoapplicable to the production of a diagnostic test in anindustrial setting.

Although the described cell-free system herecan materialize any given mRNA information, theproduction of active membrane proteins and glyco-proteins is sometimes exceptional. As was statedearlier, particular membrane proteins can be pre-pared in cell-free systems from wheat embryos andE. coli cells in the presence of artificial lipid-bilayer.76) However, it is necessary to optimize eachprotein expression protocol through trial and error,and the success rate is generally low in part due tothe highly hydrophobic and amphipathic nature ofmembrane proteins. Regarding glycoprotein produc-tion, N-glycosylation in the form of dolichol phos-phate,76),77) our current protocol has limitations dueto the lack of appropriate glycosylation apparatusin the system. Our current protocol is able to addthe core sugar when the reaction is coupled with amicrosome fraction from canine pancreas, althoughmature sugar chains may not be formed as readily asin other cell-free systems. Recently, Jaroentomeechaiand colleagues reported a more advanced method-ology using E. coli cell-free translation and glyco-sylation components.78)–80) At this moment, weshould see how these methods will turn out in theproduction of glycoproteins. Ideally, efficient systemsfor the production of these proteins may be realizedby adding fractions of endoplasmic reticulum andGolgi apparatus to the cell-free translation system


https://pim-sympo.jp/public_seminar/



when protocols for the preparation of high qualitysubcellular components are established.

Important future contributions are envisioned ofthe wheat system until maybe an entirely differentprotein synthesis system will replace the ones weknow today. Recently, a new concept for proteinsynthesis using automated flow chemistry has beenreported,81) yet it is still challenging to producefolded proteins in a functional form, includingmembrane proteins and glycosylated proteins, insuch a system.

At last, let me draw some perspectives based onour technologies with my personal wishes. Most ofthe microbes present on the surface of the earth arenot cultivable, and it is almost impossible to conductbasic research on such microbes even though theymight be highly useful in the biotechnology andindustrial sectors. For example, various microorgan-isms are found as extremophiles in deep earth anddeep sea environments, yet most of them are notcultivable in the laboratory. I believe it is possible tomaterialize genetic information collected by massivemeta-genome sequencing datasets from those ex-treme environments and use protein products madein wheat cell-free systems for studying their biologyand biochemistry. This approach may also extend toouter space, where life is currently unknown, but mydreams still reach out to the heavens.

Acknowledgements

I would like to thank all the past members in mylaboratory, including post-docs, graduate and under-graduate students, and staff. In particular, I offermy deep appreciation to Dr. Tatsuya Sawasaki,Mr. Tomio Ogasawara, Dr. Ryo Morishita, and Dr.Takafumi Tsuboi, who have participated in theaccomplishments described here over 20 years ofdevelopment of our wheat germ cell-free synthesissystem. I also wish to thank Dr. Matthias Harbersand Dr. Anton Gluck for help in preparing this paper.Lastly, I sincerely thank a special friend of mypersonal and academic life, Dr. Harry F. Noller.

References

1) Nirenberg, M.W. and Matthaei, J.H. (1961) Thedependence of cell-free protein synthesis in E. coliupon naturally occurring or synthetic polyribonu-cleotides. Proc. Natl. Acad. Sci. U.S.A. 47, 1588–1602.

2) Higo, K., Held, W., Kahan, L. and Nomura, M.(1973) Functional correspondence between 30Sribosomal proteins of Escherichia coli and Bacillus

stearothermophilus. Proc. Natl. Acad. Sci. U.S.A.70, 944–948.

3) Wittmann, H.G. (1982) Components of bacterialribosomes. Annu. Rev. Biochem. 51, 155–183.

4) Endo, Y. and Wool, I.G. (1982) The site of actionof ,-sarcin on eukaryotic ribosomes. The sequenceat the ,-sarcin cleavage site in 28S ribosomalribonucleic acids. J. Biol. Chem. 257, 9054–9060.

5) Endo, Y., Huber, P.W. and Wool, I.G. (1983) Theribonuclease activity of the cytotoxin ,-sarcin. Thecharacteristics of the enzymatic activity of ,-sarcinwith ribosomes and ribonucleic acids as substrates.J. Biol. Chem. 258, 2662–2667.

6) Endo, Y., Gluck, A., Chan, Y.L., Tsurugi, K. andWool, I.G. (1990) RNA-protein interaction. Ananalysis with RNA oligonucleotides of the recog-nition by ,-sarcin of a ribosomal domain criticalfor function. J. Biol. Chem. 265, 2216–2222.

7) Endo, Y., Mitsui, K., Motizuki, M. and Tsurugi, K.(1987) The mechanism of action of ricin andrelated toxic lectins on eukaryotic ribosomes. Thesite and the characteristics of the modification in28 S ribosomal RNA caused by the toxins. J. Biol.Chem. 262, 5908–5912.

8) Endo, Y. and Tsurugi, K. (1987) RNA N-glycosidaseactivity of ricin A-chain. Mechanism of action ofthe toxic lectin ricin on eukaryotic ribosomes.J. Biol. Chem. 262, 8128–8130.

9) Endo, Y. and Tsurugi, K. (1988) The RNA N-glycosidase activity of ricin A-chain. The charac-teristics of the enzymatic activity of ricin A-chainwith ribosomes and with rRNA. J. Biol. Chem.263, 8735–8739.

10) Endo, Y., Chan, Y.L., Lin, A., Tsurugi, K. and Wool,I.G. (1988) The cytotoxins alpha-sarcin and ricinretain their specificity when tested on a syntheticoligoribonucleotide (35-mer) that mimics a regionof 28 S ribosomal ribonucleic acid. J. Biol. Chem.263, 7917–7920.

11) Gluck, A., Endo, Y. and Wool, I.G. (1992) Riboso-mal RNA identity elements for ricin A-chainrecognition and catalysis. Analysis with tetraloopmutants. J. Mol. Biol. 226, 411–424.

12) Endo, Y., Tsurugi, K., Yutsudo, T., Takeda, Y.,Ogasawara, T. and Igarashi, K. (1988) Site ofaction of a Vero toxin (VT2) from Escherichia coliO157:H7 and of Shiga toxin on eukaryotic ribo-somes, RNA N-glycosidase activity of the toxins.Eur. J. Biochem. 171, 45–50.

13) Furutani, M., Kashiwagi, K., Ito, K., Endo, Y. andIgarashi, K. (1992) Comparison of the modes ofaction of a Vero toxin (a Shiga-like toxin) fromEscherichia coli, of ricin, and of ,-sarcin. Arch.Biochem. Biophys. 293, 140–146.

14) Wool, I.G., Gluck, A. and Endo, Y. (1992) Ribotoxinrecognition of ribosomal RNA and a proposal forthe mechanism of translocation. Trends Biochem.Sci. 17, 266–269.

15) Orita, M., Nishikawa, F., Kohno, T., Senda, T.,Mitsui, Y., Endo, Y. et al. (1996) High-resolutionNMR study of a GdAGA tetranucleotide loop thatis an improved substrate for ricin, a cytotoxic plant


protein. Nucleic Acids Res. 24, 611–618.16) Seggerson, K. and Moore, P.B. (1998) Structure and

stability of variants of the sarcin-ricin loop of 28SrRNA: NMR studies of the prokaryotic SRL and afunctional mutant. RNA 4, 1203–1215.

17) Correll, C.C., Munishkin, A., Chan, Y.-L., Ren, Z.,Wool, I.G. and Steitz, T.A. (1998) Crystalstructure of the ribosomal RNA domain essentialfor binding elongation factors. Proc. Natl. Acad.Sci. U.S.A. 95, 13436–13441.

18) Noller, H.F., Hoffarth, V. and Zimniak, L. (1992)Unusual resistance of peptidyl transferase toprotein extraction procedures. Science 256, 1416–1419.

19) Lancaster, L., Lambert, N.L., Maklan, E.J., Horan,L.H. and Noller, H.F. (2008) The sarcin-ricin loopof 23S rRNA is essential for assembly of thefunctional core of the 50S ribosomal subunit. RNA14, 1999–2012.

20) Zhou, J., Lancaster, L., Donohue, J.P. and Noller,H.F. (2014) How the ribosome hands the A-sitetRNA to the P site during EF-G–catalyzed trans-location. Science 345, 1188–1191.

21) Cech, T.R. (2000) The ribosome is a ribozyme.Science 289, 878–879.

22) Endo, Y., Tsurugi, K. and Lambert, J.M. (1988) Thesite of action of six different ribosome-inactivatingproteins from plants on eukaryotic ribosomes: TheRNA N-glycosidase activity of the proteins. Bio-chem. Biophys. Res. Commun. 150, 1032–1036.

23) Endo, Y., Tsurugi, K. and Franz, H. (1988) The siteof action of the A-chain of mistletoe lectin I oneukaryotic ribosomes. The RNA N-glycosidaseactivity of the protein. FEBS Lett. 231, 378–380.

24) Endo, Y., Tsurugi, K. and Ebert, R.F. (1988) Themechanism of action of barley toxin: A type 1ribosome-inactivating protein with RNA N-glyco-sidase activity. Biochim. Biophys. Acta 954, 224–226.

25) Endo, Y., Oka, T., Tsurugi, K. and Franz, H. (1989)The mechanism of action of the cytotoxic lectinfrom Phoradendron californicum: The RNA N-glycosidase activity of the protein. FEBS Lett.248, 115–118.

26) Ready, M.P., Brown, D.T. and Robertus, J.D. (1986)Extracellular localization of pokeweed antiviralprotein. Proc. Natl. Acad. Sci. U.S.A. 83, 5053–5056.

27) Yoshinari, S., Yokota, S., Sawamoto, H., Koresawa,S., Tamura, M. and Endo, Y. (1996) Purification,characterization and subcellular localization of atype-1 ribosome-inactivating protein from thesarcocarp of Cucurbita pepo. Eur. J. Biochem.242, 585–591.

28) Yoshinari, S., Koresawa, S., Yokota, S., Sawamoto,H., Tamura, M. and Endo, Y. (1997) Gypsophilin,a new type 1 ribosome-inactivating protein fromGypsophila elegans: Purification, enzymatic char-acterization, and subcellular localization. Biosci.Biotechnol. Biochem. 61, 324–331.

29) Ogasawara, T., Sawasaki, T., Morishita, R., Ozawa,A., Madin, K. and Endo, Y. (1999) A new class of

enzyme acting on damaged ribosomes: RibosomalRNA apurinic site specific lyase found in wheatgerm. EMBO J. 18, 6522–6531.

30) Samuel, C.E. (2001) Antiviral actions of interferons.Clin. Microbiol. Rev. 14, 778–809.

31) Roberts, B.E. and Paterson, B.M. (1973) Efficienttranslation of tobacco mosaic virus RNA andRabbit Globin 9S RNA in a cell-free system fromcommercial wheat germ. Proc. Natl. Acad. Sci.U.S.A. 70, 2330–2334.

32) Massiah, A.J. and Hartley, M.R. (1995) Wheatribosome-inactivating proteins: Seed and leaf formswith different specificities and cofactor require-ments. Planta 197, 633–640.

33) Stewart, T.S., Hruby, E., Sharma, O.K. and Robert,W.K. (1977) An ATP-dependent inhibition ofprotein synthesis in ascites cell extracts by wheatgerm protein. Biochim. Biophys. Acta 479, 31–38.

34) Taylor, B.E. and Irvin, D. (1990) Depurination ofplant ribosomes by pokeweed antiviral protein.FEBS Lett. 273, 144–146.

35) Spirin, A.S., Baranov, V.I., Ryabova, A., Ovodov,S.Y. and Alakhov, Y.B. (1988) A continuous cell-free translation system capable of producingpolypeptides in high yield. Science 242, 1162–1164.

36) Endo, Y., Otsuzuki, S., Ito, K. and Miura, K. (1992)Production of an enzymatic active protein using acontinuous flow cell-free translation system. J.Biotechnol. 25, 221–230.

37) Netzer, W.J. and Hartl, F.U. (1997) Recombinationof protein domains facilitated by co-translationalfolding in eukaryotes. Nature 388, 343–349.

38) Hartl, F.U. and Hayer-Hartl, M. (2002) Molecularchaperones in the cytosol: From nascent chain tofolded protein. Science 295, 1852–1858.

39) Hirao, I., Madin, K., Endo, Y., Yokoyama, S. andEllington, A.D. (2000) RNA aptamers that bind toand inhibit the ribosome-inactivating protein,pepocin. J. Biol. Chem. 275, 4943–4948.

40) Madin, K., Sawasaki, T., Ogasawara, T. and Endo,Y. (2000) A highly efficient and robust cell-freeprotein synthesis system prepared from wheatembryos: Plants apparently contain a suicidesystem directed at ribosomes. Proc. Natl. Acad.Sci. U.S.A. 97, 559–564.

41) Madin, K., Sawasaki, T., Kamura, N., Takai, K.,Ogasawara, T., Yazaki, K. et al. (2004) Formationof circular polyribosomes in wheat germ cell-freeprotein synthesis system. FEBS Lett. 562, 155–159.

42) Bohlmann, H. (1994) The role of thionins in plantprotection. Crit. Rev. Plant Sci. 13, 1–16.

43) Colilla, J., Rocher, A. and Mendez, E. (1990) .-purothionins: Amino acid sequence of two polypep-tides of a new family of thionins from wheatendosperm. FEBS Lett. 270, 191–194.

44) Brummer, J., Thole, H. and Kloppstech, K. (1994)Hordothionins inhibit protein synthesis at the levelof initiation in the wheat-germ system. Eur. J.Biochem. 219, 425–433.

45) Kawata, H., Katayama, K., Akimoto, M. and


Nishimoto, K. (1969) Studies on nucieases insprouts of plant seeds. J. Ferment. Technol. 47,25–31.

46) Gallie, D.R., Gallie, D.R. and Walbot, V. (1992)Identification of the motifs within the tobaccomosaic virus 5B-leader responsible for enhancingtranslation. Nucleic Acids Res. 20, 4631–4638.

47) Kamura, N., Sawasaki, T., Kasahara, Y., Takai, K.and Endo, Y. (2005) Selection of 5B-untranslatedsequences that enhance initiation of translation ina cell-free protein synthesis system from wheatembryos. Bioorg. Med. Chem. Lett. 15, 5402–5406.

48) Jacobs, J.S., Anderson, A.R. and Parker, R.P. (1998)The 3B to 5B degradation of yeast mRNAs is ageneral mechanism for mRNA turnover thatrequires the SKI2 DEVH box protein and 3B to 5Bexonucleases of the exosome complex. EMBO J.17, 1497–1506.

49) Sawasaki, T., Ogasawara, T., Morishita, R. andEndo, Y. (2002) A cell-free protein synthesissystem for high-throughput proteomics. Proc.Natl. Acad. Sci. U.S.A. 99, 14652–14657.

50) Takai, K., Sawasaki, T. and Endo, Y. (2010)Practical cell-free protein synthesis system usingpurified wheat embryos. Nat. Protocol. 5, 227–238.

51) Sawasaki, T., Hasegawa, Y., Tsuchimochi, M.,Kamura, N., Ogasawara, T., Kuroita, T. et al.(2002) A bilayer cell-free protein synthesis systemfor high-throughput screening of gene products.FEBS Lett. 514, 102–105.

52) Matsumoto, K., Tomikawa, C., Toyooka, T., Ochi,A., Takano, Y., Takayanagi, N. et al. (2007) RNArecognition mechanism of eukaryote tRNA(m7G46) methyltransferase (Trm8-Trm82 com-plex). FEBS Lett. 581, 1599–1604.

53) Takagi, K., Kim, S., Yukii, H., Ueno, M., Morishita,R., Endo, Y. et al. (2012) Structural basis forspecific recognition of Rpt1p, an ATPase subunitof 26 S proteasome, by proteasome-dedicatedchaperone Hsm3p. J. Biol. Chem. 287, 12172–12182.

54) Endo, Y. and Sawasaki, T. (2006) Cell-free expres-sion systems for eukaryotic protein production.Curr. Opin. Biotechnol. 17, 1–8.

55) Takeda, H., Ogasawara, T., Ozawa, T., Muraguchi,A., Jih, P.-J., Morishita, R. et al. (2015) Produc-tion of monoclonal antibodies against GPCR usingcell-free synthesized GPCR antigen and biotiny-lated liposome-based interaction assay. Sci. Rep. 5,11333.

56) Okimune, K., Nagy, S.K., Hataya, S., Endo, Y. andTakasuka, T.E. (2020) Reconstitution of drosophi-la and human chromatins by wheat germ cell-freeco-expression system. BMC Biotechnol. 20, 1–12.

57) Boissel, J.-P., Kasper, T.J. and Franklin, B. (1988)Cotranslational Amino-terminal Processing of Cy-tosolic Proteins, cell-free expression of site-directedmutations of human hemoglobin. J. Biol. Chem.263, 8443–8449.

58) Kanno, T., Kitano, M., Kato, R., Omori, A., Endo,Y. and Tozawa, Y. (2007) Sequence specificity andefficiency of protein N-terminal methionine elimi-

nation in wheat-embryo cell-free system. ProteinExpr. Purif. 52, 59–65.

59) Morita, E.H., Sawasaki, T., Tanaka, R., Endo, Y.and Kohno, T. (2003) A wheat germ cell-freesystem is a novel way to screen protein folding andfunction. Protein Sci. 12, 1216–1221.

60) Kohno, T. and Endo, Y. (2007) Production ofprotein for nuclear magnetic resonance study usingthe wheat germ cell-free system. Methods Mol.Biol. 375, 257–272.

61) Tonelli, M., Singarapu, K.K., Makino, S., Sahu, S.C.,Matsubara, Y., Endo, Y. et al. (2011) Hydrogenexchange during cell-free incorporation of deuter-ated amino acids and an approach to its inhibition.J. Biomol. NMR 51, 467–476.

62) Markely, J.L., Aceti, D.J., Bingman, C.A., Fox,B.G., Frederic, R.O., Makino, S. et al. (2009) TheCenter for Eukaryotic Structural Genomics. J.Struct. Funct. Genomics 10, 165–179.

63) Takemori, N., Takemori, A., Matsuoka, K.,Morishita, R., Matsushita, N., Aoshima, M. et al.(2015) High-throughput synthesis of stable iso-tope-labeled transmembrane proteins for targetedtransmembrane proteomics using a wheat germcell-free protein synthesis system. Mol. Biosyst. 11,361–365.

64) Takemori, N., Takemori, A., Tanaka, Y., Ishizaki, J.,Hasegawa, H., Shiraishi, A. et al. (2016) High-throughput production of a stable isotope-labeledpeptide library for targeted proteomics using awheat germ cell-free synthesis system. Mol. Bio-syst. 12, 2389–2393.

65) Takemori, N., Takemori, A., Tanaka, Y., Endo, Y.,Hurst, J.L., Gómez-Baena, G. et al. (2017) Multi-plexed Efficient Cell Free Expression of Recombi-nant QconCATs For Large Scale Absolute Pro-teome Quantification. Mol. Cell. Proteomics 16,2169–2183.

66) Valgepea, K., de Souza Pinto Lemgruber, R.,Abdalla, T., Binos, S., Takemori, N., Takemori,A. et al. (2018) H2 drives metabolic rearrange-ments in gas-fermenting Clostridium autoethano-genum. Biotechnol. Biofuels 11, 55.

67) Goshima, N., Kawamura, Y., Fukumoto, A., Miura,A., Honma, R., Satoh, R. et al. (2008) Humanprotein factory for converting the transcriptomeinto an in vitro-expressed proteome. Nat. Methods5, 1011–1017.

68) Morishita, R., Sugiyama, S., Denda, M., Tokunaga,S., Kido, K., Shioya, S. et al. (2019) CF-PA2Vtech:A cell-free human protein array technology forantibody validation against human proteins. Sci.Rep. 9, 19349.

69) Nussennzweig, R.S., Vanderberg, J., Most, H. andOrton, C. (1967) Protective immunity produced bythe injection of X-irradiated sporozoites of plasmo-dium berghei. Nature 216, 160–162.

70) Jiang, G., Shi, M., Conteh, S., Richie, N., Banania,G., Geneshan, H. et al. (2009) Sterile protectionagainst Plasmodium knowlesi in rhesus monkeysfrom a malaria vaccine: Comparison of heterolo-gous prime boost strategies. PLoS One 4, e6559.


71) Longley, R.J., White, M.T., Takashima, E.J.,Morita, M., Harbers, M., Obadia, T. et al. (2020)Development and validation of serological markersfor detecting recent Plasmodium vivax infection.Nat. Med. 26, 741–749.

72) Nagaoka, H., Kanoi, B., Ntege, E.H., Aoki, M.,Fukushima, A., Tsuboi, T. et al. (2020) Antibodiesagainst a short region of PfRipr inhibit Plasmo-dium falciparum merozoite invasion and PfRiprinteraction with Rh5 and SEMA7A. Sci. Rep. 10,6573.

73) Yamada, Y., Jeremiah, S.S., Miyakawa, K., Saji, R.,Nishii, M., Takeuchi, I. et al. (2020) Wholenucleocapsid protein of SARS-CoV-2 may causefalse-positive results in serological assays. Clin.Infect. Dis. ciaa637.

74) Matsuba, I., Hatori, N., Koide, N., Watanabe, Y.,Ebara, F., Matsuzawa, Y. et al. (2020) Survey ofthe current status of subclinical coronavirusdisease 2019 (COVID-19). J. Infect. Chemother.26, 1294–1300.

75) Kubo, S., Ohtake, N., Miyakawa, K., Jeremiah, S.S.,Yamada, Y., Murohasji, K. et al. (2020) Develop-ment of an automated chemiluminescence assaysystem for quantitative measurement of multipleanti-SARS-CoV-2 antibodies. medRxiv2020.11.04.20225805.

76) Pandy, A., Shin, K., Patterson, R.E., Liu, X.-Q. andRainey, K. (2016) Current strategies for protein

production and purification enabling membraneprotein structural biology. Biochem. Cell Biol. 94,507–527.

77) Zao, W. and Takeda, H. (2020) Cell-free productionof proteoliposomes for functional analysis andantibody development targeting membrane pro-teins. J. Vis. Exp. 163, e61871.

78) Lingappa, V.R., Lingappa, J.R., Prasad, R., Ebner,K.E. and Blobel, G. (1978) Coupled cell-freesynthesis, segregation, and core glycosylation of asecretory protein. Proc. Natl. Acad. Sci. U.S.A. 75,2338–2342.

79) Jaroentomeechai, T., Stark, J.C., Natarajan, A.,Glasscock, C.J., Yates, L.E., Hsu, K.J. et al. (2018)Single-pot glycoprotein biosynthesis using a cell-free transcription-translation system enriched withglycosylation machinery. Nat. Commun. 9, 2686.

80) Jaroentomeechai, T., Taw, M.N., Li, M., Aquino, A.,Agashe, N., Chung, S. et al. (2020) Cell-freesynthetic glycobiology: Designing and engineeringglycomolecules outside of living cells. Front Chem.8, 645.

81) Hartrampf, N., Saebi, A., Poskus, M., Gates, Z.P.,Callahan, A.J., Cowfer, A.E. et al. (2020) Synthe-sis of proteins by automated flow chemistry.Science 368, 980–987.

(Received Jan. 12, 2021; accepted Feb. 17, 2021)

Profile

Yaeta Endo was born in Tokushima Prefecture in 1946 and graduated fromTokushima University, School of Medicine in 1969. He received his Ph.D. degree fromTokushima University in 1975, He started his research career as an Assistant Professor inthe Department of Nutritional Biochemistry, Tokushima University Graduate School in1975, and was appointed as a Lecturer in the department in 1979. He studied in theU.S.A. as a postdoctoral fellow from 1980 to 1982 with Dr. Ira G. Wool, Cummings LifeScience Center, University of Chicago. He was appointed Associate Professor in thedepartment of Biochemistry, Yamanashi Medical College in 1984. He was promoted toProfessor in the Department of Applied Chemistry School of Engineering, EhimeUniversity in 1992, and then Director of the Venture Business Laboratory, EhimeUniversity, 1999, an Executive Director of Ehime University, and a Special University Professor Emeritus of EhimeUniversity from 2011 to the present. He has worked for the University of California, Santa Cruz as a VisitingProfessor (2011–2015). He is currently working as a Visiting Professor in Ehime Prefectural University of HealthSciences. He won the 3rd Yamazaki-Teiichi Prize in 2003, Kei Arima Memorial Japan Bioindustry AssociationAward in 2006, and The Commendation for Science and Technology by the Minister of Education, Culture, Sports,Science and Technology in 2008.


development of a cell-free protein synthesis system for

Documents