Contents lists available at ScienceDirect

BBA - General Subjects

journal homepage: www.elsevier.com/locate/bbagen

A novel method for cloning of coding sequences of highly toxic proteins

Rafal Krelaa, Elzbieta Porebab, Martyna Weglewskab, Tomasz Skrzypczaka,Krzysztof Lesniewicza,⁎

a Department of Molecular and Cellular Biology, Institute of Molecular Biology and Biotechnology, Adam Mickiewicz University in Poznan, 61-614 Poznan, Umultowska St.89, PolandbDepartment of Molecular Virology, Institute of Experimental Biology, Adam Mickiewicz University in Poznan, 61-614 Poznan, Umultowska St. 89, Poland

A R T I C L E I N F O

Keywords:Toxic proteinProtein overexpressionAcanthamoeba castellaniiNuclease genesFrameshift introduction

A B S T R A C T

Background: During standard gene cloning, the recombinant protein appearing in bacteria as the result of ex-pression leakage very often inhibits cell proliferation leading to blocking of the cloning procedure. Althoughdifferent approaches can reduce transgene basal expression, the recombinant proteins, which even in traceamounts inhibit bacterial growth, can completely prevent the cloning process.Methods: Working to solve the problem of DNase II-like cDNA cloning, we developed a novel cloning approach.The method is based on separate cloning of the 5′ and 3′ fragments of target cDNA into a vector in such a waythat the short Multiple Cloning Site insertion remaining between both fragments changes the reading frame andprevents translation of mRNA arising as a result of promoter leakage. Subsequently, to get the vector with full,uninterrupted Open Reading Frame, the Multiple Cloning Site insertion is removed by in vitro restriction/ligationreactions, utilizing the unique restriction site present in native cDNA.Results: Using this designed method, we cloned a coding sequence of AcDNase II that is extremely toxic forbacteria cells. Then, we demonstrated the usefulness of the construct prepared in this way for overexpression ofAcDNase II in eukaryotic cells.Conclusions: The designed method allows cloning of toxic protein coding sequences that cannot be cloned bystandard methods.General significance: Cloning of cDNAs encoding toxic proteins is still a troublesome problem that hinders theprogress of numerous studies. The method described here is a convenient solution to cloning problems that arecommon in research on toxic proteins.

1. Introduction

For modern molecular biology and biotechnology, cDNA cloning is abasic technique that allows the acquisition of individual coding se-quences as well as entire transcriptomes. Although the majority ofcDNA sequences can be easily cloned using standard methods, thecloning of certain coding sequences still encounters problems. Theseproblems may arise from the specific structure of DNA itself, such as thepresence of numerous tandem or inverted repeats [1], or from the un-favorable properties of proteins encoded by this sequence. The mostserious problems are encountered during the cloning of sequences en-coding proteins that display strong toxic activity toward bacterial cells.The toxic effect of recombinant proteins toward bacteria can be clas-sified at different levels depending on the difficulties that they cause inlaboratory practice. For example, sequences encoding less toxic

proteins can be relatively easy to clone but are laborious to overexpressin bacteria. However, the strong toxic activity of recombinant proteinscan halt the cloning procedure of corresponding cDNA.

Standard cloning procedure involves insert preparation, usually byPCR reaction, and its joining with the vector DNA by in vitro reaction.Subsequently, the reaction products are introduced into bacteria toselect the correct constructs. This step can be critical for cloning whenthe sequence of interest encodes a protein that inhibits the bacterialgrowth on selective plates. Theoretically, at this stage of cloning, thesequences introduced into the vectors should not be expressed sincethey are not under the control of active regulatory sequences. The mostcommonly used bacterial expression vectors possess promoters thatmust be activated by an inducing agent, such as isopropyl β-D-1-thio-galactopyranoside (IPTG). In turn, the eukaryotic expression vectorshave promoters that should not be active in bacterial cells. However,

https://doi.org/10.1016/j.bbagen.2018.12.010Received 23 September 2018; Received in revised form 26 November 2018; Accepted 17 December 2018

⁎ Corresponding author.E-mail addresses: rafal.krela@amu.edu.pl (R. Krela), eporeba@amu.edu.pl (E. Poreba), martyna.weglewska@amu.edu.pl (M. Weglewska),

tomasz.skrzypczak@amu.edu.pl (T. Skrzypczak), lesniew@amu.edu.pl (K. Lesniewicz).

BBA - General Subjects 1863 (2019) 521–527

Available online 19 December 20180304-4165/ © 2018 Elsevier B.V. All rights reserved.

T

many bacterial promoters reveal some degree of expression even in theabsence of the inducer [2]. Similarly, to some extent, the eukaryoticpromoters can be recognized by prokaryotic transcription machinery[3]. Moreover, the read-through effect mediated by some bacterial RNApolymerases may generate the transcript from a cryptic promoter lo-cated upstream of the proper promoter. Usually, this basal expression ofrecombinant sequences does not influence bacterial growth and can beignored. However, even the low expression leakage of genes encodingproteins that interfere with processes that are crucial for bacterialviability can inhibit their growth on selective plates.

Full access to information about proteins toxic for Escherichia coli islimited, since failures in their cloning often mean that they are notpublished. However, available data show that the cDNA of severalgroups of proteins is difficult to clone. One of these comprises somemembrane proteins. Hydrophobic membrane proteins usually revealtheir toxic effect on cells by aggregation in the cytoplasm and causeproblems after induction of their expression. However, the membraneproteins that disrupt cell membrane function, such as some ion chan-nels (for example, CCH1) and transporters [4,5] or virus envelopeproteins [6], can affect bacterial viability even in small amounts andthus may inhibit bacterial growth at cloning stage. Other examples ofproteins that may inhibit bacterial growth are some constituents of ri-bosome, whose toxicity most probably stems from their incompatibilitywith the E. coli translation machinery [7,8]. Moreover, the coding se-quences of different enzyme groups cause difficulties during theircloning. In order to clone some lipases [9], subtilase-like enzymes [10],colicins [11], proteases [12,13] and some enzymes of toxin-antitoxinsystems [14] special cloning procedures are necessary. Numerous pro-teins belonging to colicins, as well as enzymes participating in toxin-antitoxin systems, possess various nuclease activities [15]. Somedeoxyribonucleases, also including certain restriction enzymes [16], aswell as ribonucleases [17,18], are enzymes whose toxicity causes pro-blems during cloning of their coding sequences. The toxicity of theseproteins is due to the fact that even the smallest amount of exogenousnucleases in bacterial cells severely affects their survival by destabili-zation of the genome or by transcript damage.

In addition to studies carried on individual proteins, certaingenomic observations suggest that toxic proteins are widespread indifferent genomes. Sorek et al. [8], who analyzed the horizontal genetransfer from different prokaryotic genomes into E. coli, identified avast collection of “unclonable” genes and concluded that the toxicity ofthe protein products of these genes may be the predominant cause fortransfer failure. Kimelman et al. [19] also arrived at a similar conclu-sion.

In order to enable the cloning of highly toxic genes, different stra-tegies have been developed. Since the main problem arises from ex-pression leakage, efforts have focused on improvements in the controlof gene expression systems. These strategies rely on the application ofphage-derived highly specific RNA polymerases, the use of tightlyregulated promoters, the reduction of basal transcripts by an antisensestrategy, and the blocking of read-through transcription by insertion ofadditional terminators [16,18]. The limitation of basal expression canalso be achieved by using low copy number plasmids [11], the cultureof bacteria in modified media [10], the application of bacteria-freeapproaches [20], and the engineering of new toxin-resistant E. colistrains [21]. All these procedures significantly reduce the backgroundexpression and may be useful for cloning of moderately toxic genes.However, the complete inhibition of active protein formation would beexpected in the case of extremely toxic proteins. This goal can beachieved by insertion of an intron sequence within the target cDNA[22]. Due to the absence of splicing in bacterial cells, the intron remainsstable in mRNA and blocks translation of an active protein if it in-troduces additional stop codons or changes the reading frame. Finally,the problem of cloning genes encoding toxic proteins can be bypassedby the elimination of plasmid propagation in E. coli and instead by theintroduction of an insert into the plasmid by an in vitro reaction, using

homologous recombination [4].We attempted to clone the Acanthamoeba castellanii DNase II-like

coding sequence into a pcDNA3 eukaryotic expression vector whenconducting research on some DNA degradation enzymes. However, wewere not able to clone the correct version of this cDNA, probably due tothe highly toxic activity of its protein product, using the standardcloning methods. To solve this problem, we developed a new methodenabling the cloning of toxic genes. Our purpose was to find a methodthat allows the introduction of the coding sequence of a selected gene toany chosen vector without the necessity of its reconstruction. The de-veloped method is based on separate cloning of 5′ and 3′ fragments oftarget cDNA into a vector Multiple Cloning Site (MCS) in such a waythat a short MCS insertion remains between both parts of the clonedOpen Reading Frame (ORF). Since this MCS insertion changes thereading frame and additionally introduces stop codons inside the ORF,it prevents translation of toxic proteins arising as a result of promoterleakage at the cloning stage. Subsequently, to get the vector with a full,uninterrupted ORF, the MCS insertion was removed by standard, invitro restriction/ligation reactions, utilizing a unique restriction sitepresent in native cDNA. The constructs prepared in this way weresuccessfully used to overexpress DNase II-like proteins in a eukaryoticexpression system.

2. Materials and methods

2.1. Acanthamoeba castellanii culture and mRNA isolation

Acanthamoeba castellanii, strain Neff, was cultured as described byJarmuszkiewicz et al. [23]. Trophozoites of A. castellanii were collectedat the density of 2×105 cell/ml and lysed with Trizol agent (In-vitrogen). The RNA isolation procedure was performed according to themanufacturer's protocol. The SuperScript III First-Strand Synthesis Kit(Invitrogen) was used according to the manufacturers protocol to syn-thesize total cDNA.

2.2. PCR amplification of AcDNase II and EGFP coding sequences

PCR reactions were performed with primers designed for putativeAcanthamoeba castellanii deoxyribonuclease II mRNA (ACA1_041920,mRNA accession number XM_004338820) [24] described here asAcDNase II.

We used the following primers to amplify full-length AcDNase IIcDNA:

AcDNase2F - GCAAGATCTATGAAGAGCAGCTTCGCCCTCCTAcDNase2F - AGTGCTAGCTTAGCACGAGTCCGCGTGGGThe following primers were used to amplify 5′ and 3′ fragments of

AcDNase II:5′FrAcDNase2F AGCAAGCTTATGAAGAGCAGCTTCGCCCTCCT5′FrAcDNase2R AATGGATCCACCGGTTGCTGGACAGGGTG3′FrAcDNase2F CTGGAATTCACCGGTTGGACACCAACCTG3′FrAcDNase2R CGCTCTAGATTAGCACGAGTCCGCGTGGGThe following vector-specific primers were used to identify AcDNase

II sequence ligated with pcDNA3:pcDNA3F CGCAAATGGGCGGTAGGCGTGpcDNA3R GCGATGCAATTTCCTCATTTThe complete coding sequence of EGFP was amplified by PCR using

pSAT6-EGFP-N1 plasmid as a template with following primers:EGFPF - ATAAAGCTTATGGTGAGCAAGGGCGAGGAEGFPR - GGGAAGCTTCTTGTACAGCTCGTCCATGCCGAll PCR reactions were conducted Using Phusion High-Fidelity DNA

Polymerases (Thermo Scientific) and an Eppendorf 5331 MasterCyclerGradient Thermal Cycler. The PCR reaction mixture composition andthermal cycling profiles were optimized according to manufacturer'sprotocol and primers' melting temperature.

R. Krela et al. BBA - General Subjects 1863 (2019) 521–527

522

2.3. The cloning of PCR products

The PCR products were digested with the appropriate FastDigestrestriction endonucleases (Thermo Scientific), purified by PCR Clean-upkit (Qiagen) and ligated using T4 ligase (Thermo Scientific) withpcDNA3 vector digested with corresponding enzymes. The reactionconditions were as described in the manufacturer's instructions.

The competent E. coli cells (DH5α) were transformed with re-combinant vectors using a standard heat shock transformation methodand were selected on LB agar plate containing (50 μg/ml) ampicillin.

The EGFP-AcDNase II fusion protein was prepared by adding PCR-generated EGFP coding sequence lacking stop codons to pcDNA3-AcDNase II construct using HindIII restriction site located upstream ofAcDNase II.

All the resulting constructs were confirmed by sequencing.

2.4. The removal of an insertion before the HeLa cell transfection

The DNA constructs designed for cell transfection were digested byAgeI (Thermo Scientific BshTI) according to manufacturer's instruction.Then, the restriction enzyme was heat inactivated, and plasmid DNAwas purified by PCR Clean-up kit (Qiagen). Finally, plasmid DNA wasligated with T4 ligase (Thermo Scientific) according to standard con-ditions.

2.5. Cell culture and transfection

HeLa cells were grown in Dulbecco minimum essential medium(DMEM) supplemented with 10% fetal bovine serum, glutamine2mmol/l and 1% penicillin-streptomycin. Plates were incubated in a37 °C incubator at 5% CO2. Cells were seeded into 6-well plates at adensity of 100,000 cell per well. Using the Lipofectamine 2000 reagent(Invitrogen) according to the manufacturer's instructions, the cells weretransfected with different DNA constructs after an overnight growth.After transfection, the cells were incubated overnight in standard con-ditions and then were used for the following experiment. Each trans-fection was performed in three biological replicates.

2.6. Fluorescence microscopy analysis

Fluorescence images were collected using Axio Observer Z1 in-verted fluorescent (Carl Zeiss, Germany) supplied with a HBO 100Wmercury lamp and A-Pln 10×/0.25 Ph1 and LD Pln 40×/0.6 Ph2 DIC IIobjectives. Cells growing on 35mm cell imaging dishes with glassbottom were photographed using AxioCam MRm camera andAxioVision 4.8 software. More than one hundred cells were analyzedfor each biological replicate. Approximately one hundred cells werecaptured per biological replicate.

2.7. Western blot

Preparation of protein extracts from transfected cells and Westernblot analysis was performed as reported previously [25,26]. The pri-mary anti-GFP antibody was purchased from Santa Cruz Biotechnology.At least one Western blot test was performed per biological replicate.

2.8. Nuclease activity assay

The in-gel nuclease activity assay was performed as reported pre-viously [27] with the following modifications. Equal amounts of protein(1 μg) were denatured for 5min at 99 °C in sample buffer before loadingonto the gel. The resolving mini gels contained heat denatured calfthymus ssDNA at the concentration of 0.01mg/ml. After electrophor-esis, the resolving gel was soaked twice at room temperature in a buffercontaining 20% 2-propanol for 20min and subsequently was washedfor 15min and incubated overnight in renaturation/reaction buffer

((25mM sodium acetate (pH 5.5); 1% (v/v) Triton X-100, 1mMEDTA)). After incubations, the gel was washed for 15min in ice-coldstaining buffer (10mM Tris–HCl pH 8.0; 1 mM EDTA; 0.01mg/mlethidium bromide) and photographed. One to three In-gel nucleaseassays were performed per biological replicate.

3. Results

3.1. An attempt at cloning DNase II-like cDNA using standard method

As a part of the project conducted in our laboratory, we cloned thecoding sequence of a predicted DNase II-like gene (AcDNase II) fromAcanthamoeba castellanii into eukaryotic expression vector pcDNA3downstream of the CMV promoter sequence. Initial attempts at cloningwere performed by a standard method based on ligation of RT-PCRproducts with vector DNA using restriction sites. The PCR reaction wasperformed with specific primers designed to amplify the full ORF se-quence of AcDNase II. We used primers bearing restriction sites at their5′ ends to facilitate ligation of PCR products with vector DNA. As shownin Fig. 1, the PCR product size is the right size (1158 bp) and its se-quencing confirmed that it comprises the complete reading frame ofAcDNase II. Then, after the standard restriction and ligation reactionwith the pcDNA3 vector and subsequent E. coli transformation, weobtained about 150 colonies on the selective plates. However, none ofthem contained an appropriate insert. All colonies growing on the se-lective plates contained empty vectors or had inserts, which re-presented defectively spliced transcripts or had errors introducedprobably during the RT-PCR reaction. Mutations caused frameshift orintroduced additional, internal stop codons in each of these cases. Sincethe sequencing performed directly on RT-PCR products did not revealthe presence of such forms, we think that they represent a very smallpart of reaction products. The fact that these mutated fragments arecloned preferentially instead of the dominant, correct sequence stronglysuggests that an intact coding sequence has a negative influence onbacterial viability. As shown in Fig. 1, we performed the PCR reactionusing a ligation mixture as the template and primers flanking the MCSregion of the pcDNA3 vector to rule out the possibility that failure incloning could be the result of unknown problems at the ligation step.Since this reaction revealed that target cDNA is ligated with vectorDNA, in our opinion, the only reasonable explanation of cloning failurewas the toxic effect of recombinant AcDNase II derived from CMVpromoter leakage.

3.2. New strategy for cloning of highly toxic protein coding sequences

The cloning of cDNA sequences encoding highly toxic proteins is notproblematic when such sequences contain a mutation causing a fra-meshift or an internal stop codon. However, such mutations prevent itsapplication in protein production in heterologous expression systems,which is the main goal of most experiments in molecular biology.Therefore, utilization of such sequences for protein expression would bepossible only when the inhibitory effect of a frameshift could be re-versed. The main assumption of our approach was to clone the toxic

Fig. 1. PCR products of AcDNase II. RT-PCR product comprising full-lengthORF, amplified with cDNA-specific primers (1) and PCR product amplified frompcDNA3-AcDNase II ligation mixture used as template with vector-specific pri-mers (2).

R. Krela et al. BBA - General Subjects 1863 (2019) 521–527

523

protein coding DNA in two fragments separated by an insertion thatintroduces a frameshift in the given ORF. Then, this insertion is re-moved using the unique restriction site residing within the nativecoding sequence of the given protein to obtain an uninterrupted readingframe prior to target cell transfection (Fig. 2A–E).

Cloning of cDNA with an insertion inside can be easily made byconsecutive ligations of vector restriction sites with two DNA fragmentsthat together represent a full coding sequence. The cDNA division pointwas determined by the position of the native restriction site present inthe target sequence while designing this experiment (Fig. 2A). In thecase of AcDNase II, the first DNA fragment (5′) spans from the startcodon (+1) to the last nucleotide of the native AgeI site (+220). ThisDNA fragment was prepared by a PCR reaction using primers that in-troduced additional HindIII and BamHI restriction sites at 5′ and 3′ endsof the PCR product, respectively. These restriction sites were then uti-lized to ligate the PCR amplified fragment with corresponding restric-tion sites of pcDNA3 MCS. The second fragment of AcDNase II (3′),extending from the first nucleotide of the AgeI site (+215) to the stopcodon (+1065), was also prepared by a PCR reaction, but two otherrestriction sites, Eco RI and Xba I, were added at its 5′ and 3′ ends,respectively (Fig. 2B). These restriction sites were used to ligate thisPCR product downstream of the first DNA fragment (3′) cloned pre-viously in MCS. The resultant plasmid contained a complete cDNA se-quence separated by a short, MSC derived insertion spanning from theBamHI to EcoRI sites and flanked by a duplicated AgeI (Fig. 2C). Sincethis insertion contains double stop codons in a frame with the firstfragment of AcDNase II ORF, it terminates translation at this position.However, the number of nucleotides in the insertion is ten, so that evenin the absence of stop codons, it would change the reading frame andwould prevent the production of correct proteins during bacterialgrowth on selective plates.

In contrast to our previous attempts, when we tried to clone thecomplete cDNA of AcDNase II in a single fragment, the cloning of thecDNA in two fragments, according to the above procedure, was suc-cessful in the first approach. This confirms our assumptions that acoding sequence separated by an insertion does not influence bacterialgrowth. Consequently, we were able to select the correct clone, purifythe required amount of plasmid DNA, and verify the insert by con-ducting a sequencing reaction (Fig. 3).

Designing the cloning strategy for DNA fragments encoding toxicproteins, different determinants should be considered. One of these isthe choice of internal restriction site for cloned cDNA. It obviously mustbe unique, and it should also be located in an appropriate position. Arestriction site located at a short distance from the start codon should beavoided since alternative start codons located downstream of this pointcan initiate translation of toxic proteins. Moreover, since the first (3′)DNA fragment is translated in the correct frame, it should not comprise

Fig. 2. Strategy for cloning of AcDNase II coding sequence. A. Scheme of cDNAsequence of AcDNase II. The numbering below the box (1–1065) refers to thefirst and last nucleotide of ORF. AgeI site is marked as blue box. B. Two frag-ments of AcDNase II (5′ and 3′, respectively) amplified separately by RT-PCRreactions. The restriction sites marked in black were added by primers used toamplify both fragments. C. Both cDNA fragments introduced into pcDNA3vector. The black bold line shows the fragment of vector MCS which introducesframeshift between 5′ and 3′ cloned fragments. The vector built in this way wasable to propagate in E. coli growing on the selective plate. D, E. Preparation ofvector DNA for cell transfection. Restriction reaction with AgeI removes MCSfragment from AcDNase II coding sequence (D) and ligation reaction recoversthe native form of AcDNase II cDNA. This plasmid was used for HeLa celltransfection (E).

Fig. 3. The AcDNase II cDNA sequence cloned into pcDNA3. The sequences of AcDNase II cDNA and vector MCS are shown by blue and black colors, respectively.The first ATG and TAA stop codons of AcDNase II coding sequences are written in bold. The duplicated internal AgeI restriction site (ACCGGT) that determines thesite of division of AcDNase II coding sequence is indicated by open boxes. The MCS restriction sites used to clone both cDNA fragments are underlined, for example:(AAGCTT) HindIII, (GGATCC) BamHI, (GAATTC) EcoRI, (TCTAGA) XbaI. The double stop codons in MCS-derived insertion are indicated by bold, black fonts. (Forinterpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

R. Krela et al. BBA - General Subjects 1863 (2019) 521–527

524

the complete sequence of the domain that is responsible for proteintoxicity. Such a domain must be interrupted by insertion or be locateddownstream of the frameshift sites. In the absence of a natural re-striction site at a suitable position, one can be introduced since somemutations at the second and third position of a codon may generate anew restriction site in an ORF without changing the amino acid se-quence. Indeed, a software tool dedicated to the design of such a mu-tagenesis is available on-line (i.e., http://www.insilicase.com/Web/Mutagenesis.aspx). Such a restriction site can be easily introduced intocloned DNA during amplification of both fragments applying a modifiedreverse primer used to amplify the 3′ cDNA fragment and a forwardprimer applied to the generate 5′ fragment. The small number of re-striction sites in a given MCS is also not a limitation with this method.We could use four different restriction sites available in MCS whileworking with a pcDNA3 vector, but in fact, even a single restriction siteis sufficient for this procedure. In such a case, the additional restrictionsites can be introduced at the 3′ end of the first PCR product with thereverse primer used for its amplification.

Application of cDNA containing an insertion for protein productionis possible after treatment that removes the insertion, and thereby re-stores the correct reading frame. This goal can be achieved by in vitrorestriction/ligation reactions in the described method. Since the inser-tion is flanked by a duplicated AgeI restriction site as an element of theAcDNase II coding sequence, the digestion performed with this enzymeremoves the insertion (Fig. 2D), and the subsequent ligation reactionrestores the complete ORF to a form that is identical to the wild-typesequence (Fig. 2E). Most transfection experiments are performed withplasmid DNA purified directly from bacteria. However, numerous ex-amples show that DNA fragments obtained from enzymatic reactions,such as PCR [20,28,29] or homologous recombination [4], can also beused for efficient cell transfection.

Plasmid DNA obtained from in vitro reactions can be applied for celltransfection only when it adopts the correct, active form, and additionalreaction mixture components do not interfere with the cell growth.Application of restriction and ligation reactions allows the removal ofan insertion with high efficiency and in a relatively short time. In ourmethod, since the restriction reaction generates a linear plasmid withsticky or cohesive ends, its recirculation is a favored process.Theoretically, the insertion that has been cut out from plasmid by re-striction enzyme could be removed from the reaction mixture by the gelelution method. However, in our experience, it would be more favor-able to use commercially available kits designed to purify DNA fromenzymatic reactions since they usually allow purification of plasmidDNA from short DNA fragments as well as from other reaction com-ponents. Nonetheless, our further transfection experiments did not re-veal any influence of DNA purification method on the effectiveness ofprotein overexpression.

By performing a PCR reaction, we tested the efficiency of insertionremoval. The plasmid DNA with the removed insertion was used as thetemplate for PCR reactions with primers flanking the site of insertion.The plasmid DNA possessing an unremoved insertion was used as acontrol. Since both PCR products should differ in size and the presenceof MCS-derived restriction sites, they were digested by BamHI andanalyzed electrophoretically. As shown in Fig. 4, the PCR productamplified on a plasmid containing an insertion was cleaved by BamHIdue to the presence of a BamHI restriction site in the insertion. A similarreaction performed with a PCR product obtained from a plasmid treatedto remove the insertion is resistant to the above digestion, confirmingthat the applied procedure removes insertions from cloned cDNA.

We used the pcDNA3-AcDNase II construct with deleted insertionfor bacterial transformation to confirm the toxic effect of AcDNase IIproteins on E. coli growth. However, as expected, this plasmid was notable to grow on selective plates, unlike pcDNA3-AcDNase II containinginsertion in its coding sequence (data not shown).

3.3. Application of cloned cDNA for protein overexpression

Cloned coding sequences are mainly used to generate recombinantproteins in heterologous expression systems. Therefore, we decided tocheck the usefulness of our cloning method for the expression ofAcDNase II proteins in eukaryotic cell lines. For this purpose, we usedtwo different constructs, that is, pcDNA3-AcDNase II and pcDNA3-EGFP-AcDNase II, that allowed us to express the native and EGFP-tagged recombinant nucleases, respectively. Both constructs, treated toremove insertions, were used to transfect the HeLa cells.

We made an overexpression of AcDNase II fused to EGFP to analyzethe usefulness of our approach in experiments involving microscopicobservations and Western blot analysis. EGFP fluorescence was de-tected in HeLa cells transfected with pcDNA3-EGFP-AcDNase II, as il-lustrated in Fig. 5A. The EGFP-AcDNase II signal distribution is dif-ferent than in control cells transfected with an empty vector (pcDNA3-EGFP) that was typically diffused throughout the cytoplasm. This showsthat the cellular location of EGFP is conditioned by fused AcDNase II.However, because the accurate establishment of the subcellular loca-tion of AcDNase II should be analyzed in the natural host of this proteinand since this problem is beyond the scope of this paper, it will not bedescribed here.

To verify the efficiency of protein synthesis, the protein extractsfrom transfected cells were also subjected to Western blot analysis usingan antibody against EGFP. Two bands of slightly different mobility wereseen in the protein extract obtained from cells transfected by pcDNA3-EGFP-AcDNase II, as shown in Fig. 5B. The estimated molecular weightof these bands was approximately 65 kDa, roughly the same as thatpredicted from the deduced amino acid sequence of the EGFP-AcDNaseII fusion protein. We assume that the presence of two bands on theWestern blot was most likely due to additional post-translationalmodifications of AcDNase II. Moreover, two additional bands, probablyproteolytic cleavage products, are visible below.

The third test which we performed to analyze the usefulness of ourcloning method aimed to detect the enzymatic activity of overexpressedproteins. Since a cloned DNA sequence encodes the protein showinghigh similarity to known DNases II, we assumed that an overexpressedAcDNase II-like protein also possesses nucleolytic activity. The well-characterized mammalian DNase II hydrolyzes DNA non-specificallyunder acidic conditions without the requirement for divalent cations.To analyze potential nuclease activity of recombinant amoebalAcDNase II, we used the In-gel DNase assay, applying similar catalyticconditions to those reported for known mammalian DNases II. Asshown in Fig. 5C, the protein extract obtained from HeLa cells trans-fected with pcDNA3-AcDNase II revealed strong nuclease activity cor-responding to a protein of 35 kDa in contrast to the control samples,

Fig. 4. PCR reaction demonstrating effectiveness of insertion removal. (1) PCRproducts amplified from pcDNA3-AcDNase II template possessing unremovedinsertion inside AcDNase II coding sequence. Since this insertion comprisesBamHI restriction site, the cleavage performed by this enzyme (+) generatestwo DNA fragments of 893 bp and 230 bp. (2) The PCR product obtained frompcDNA3-AcDNase II treated to remove insertion is resistant to BamHI digestion(+).

R. Krela et al. BBA - General Subjects 1863 (2019) 521–527

525

which were protein extracts from cells transfected with an empty vectorand a vector containing an AcDNase II coding sequence containing anundeleted insertion. This test again confirmed that the described pro-cedure for cDNA cloning can be applied for effective toxic proteinoverexpression.

4. Discussion

The cloning of sequences that encode proteins toxic for bacteria isstill a problem that negatively affects the achievement of numerousscientific objectives in modern molecular biology and biotechnology.This problem is often raised in research papers and is a common topic ofdiscussion on molecular biology and biotechnology Internet forums.The importance of this problem is also well illustrated by some genomichigh-throughput experiments. Kimelman et al. [19] have found thatclone-based genome sequencing—unlike cloning-independent meth-ods—often results in cloning gaps. They analyzed 393 microbial gen-omes and found more than 15,000 “unclonable” genes that encodeproteins whose activity most probably is responsible for cloning fail-ures. It is significant that the same uncharacterized gene orthologousderived from different genomes cannot usually be cloned into E. coli,suggesting that cloning failure is usually associated with specific fea-tures of a particular protein family. Similar conclusions can also bedrawn from the study of horizontal gene transfer between differentbacterial species [8]. The analysis of gene movements from 79 pro-karyotic genomes allowed the identification of the set of genes that areresistant to such transfer. The exploration of 287,884 genes from 79genomes resulted in the identification of 642 genes that are un-transferable to E. coli in the wild. Experimental results performed with40 selected genes confirmed that the vast majority of them inhibitedbacterial growth, which suggests that gene-transfer barriers may beassociated with the toxicity of the transferred gene to the host [8].

These and other results show that genes whose protein products

prevent the cloning of their coding sequence are very common and maybe counted in thousands. Their examination is very difficult becausethey are often hidden in the genome gaps or, even when available indatabases, they are hard to manipulate using standard cloning techni-ques. However, it is important to emphasize here that these genes havegreat biotechnological potential. Their ability to inhibit bacterialgrowth is especially significant for medicine since they can be used forantimicrobial intervention. It is also worth noting that known collec-tions of proteins that reveal antibacterial activities comprise multipleproteins whose molecular activities are as yet unknown [19]. As anincreasing demand for new antimicrobial compounds has been ob-served in recent years, the identification of novel mechanisms re-sponsible for bacterial growth inhibition opens new, promising ther-apeutic opportunities.

The majority of techniques developed to enable the cloning of se-quences encoding toxic proteins have focused on prevention of ex-pression leakage. This approach can solve the problem of proteins thatreveal moderate toxic activity, but it does not block the unwantedprotein expression completely. However, in the case of highly toxicproteins, even a very low basal expression can prevent bacterial growth,thus blocking the cloning process. Moreover, it usually requires mod-ification of expression vectors and time-consuming elaboration ofbacterial growth conditions. Other approaches, relying on the use of in-vitro-prepared templates, such as PCR or recombination reaction pro-ducts, allow the bypassing of the propagation of DNA in bacteria, buthinder efficient genetic manipulation such as site-directed mutagenesis.

In our study, we decided to develop a method that does not reducethe expression leakage but completely blocks protein expression at thetranslation level. This approach relies on cloning of cDNA sequencesharboring frameshift that can subsequently be removed by simple invitro reactions that restore the reading frame to a form identical withnative cDNA. The described method resembles, to some extent, an ap-proach based on the use of intron insertion that also inhibits proteintranslation in bacteria. However, although the application of an introneffectively blocks protein expression, it complicates the cloning proce-dure and narrows the range of potential hosts for heterologous ex-pression. In our method, to introduce the frameshift in a given cDNA,the fragment of vector multicloning site is used, and this can be easilyremoved by simple restriction/ligation reactions. Therefore, a widerange of vectors can be used without the necessity of their modification.Moreover, unlike the method based on the use of an intron, the choiceof hosts is not limited by cell-specific splicing. In summary, the sim-plicity and effectiveness of the method presented here, in our opinion,allows the circumvention of problems stemming from high toxicity ofproteins encoded by cloned DNA.

5. Conclusions

We have developed a new method enabling the cloning of cDNAsequences encoding proteins whose toxic properties prevent the use ofstandard cloning procedures. This method is based on separate cloningof 5′ and 3′ fragments of target cDNA into a vector MCS in such a waythat a short MCS insertion remains between both parts of the clonedORF. Since this MCS insertion introduces a frameshift, it preventstranslation of toxic proteins arising as a result of promoter leakage atthe cloning stage. Subsequently, to generate a vector with a full, un-interrupted ORF, the MCS insertion was removed by standard, restric-tion/ligation reactions, utilizing the unique restriction site present innative cDNA. This very simple method allows the fast and effectivecloning of sequences encoding even extremely toxic proteins. The se-quences cloned in this way can be used for overexpression of toxicproteins in eukaryotic expression systems.

Author contribution

Rałal Krela: Investigation. Elzbieta Poreba: Validation, Writing –

Fig. 5. Application of a novel approach of cDNA cloning for protein over-expression in HeLa cells. HeLa cells were transfected with pcDNA3-EGFP-AcDNase II or pcDNA3-AcDNase II and empty pcDNA3 vector as a control. A.fluorescence microscopy of HeLa cells transfected with pcDNA3-EGFP vectoralone as a control (left panel) and pcDNA3-EGFP-AcDNase II (right panel). B.Western blot analysis of protein extract from HeLa cells transfected withpcDNA3-EGFP vector alone as a control (1) and pcDNA3-EGFP-AcDNase II (2),with anti-GFP antibody. C. In-gel nuclease activity assay showing nucleolyticactivity of AcDNase II in protein extracts obtained from HeLa cells transfectedwith empty vector as a control (1) and pcDNA3-AcDNase II (2). Transfectionwith pcDNA3-AcDNase II containing undeleted insertion (3) shows the in-hibitory effect of insertion on AcDNase II expression.

R. Krela et al. BBA - General Subjects 1863 (2019) 521–527

526

Review & Editing. Martyna Weglewska: Investigation. TomaszSkrzypczak: Visualization. Krzysztof Lesniewicz: Conceptualization,Methodology, Supervision, Writing – Original Draft.

Acknowledgements

This study was supported by the National Science Center [grants2014/15/NZ3/00863 to K.L., NN303 813140 to EP and 2013/09/N/NZ3/00532 to T.S].

References

[1] R. Godiska, D. Mead, V. Dhodda, C. Wu, R. Hochstein, A. Karsi, K. Usdin,A. Entezam, N. Ravin, Linear plasmid vector for cloning of repetitive or unstablesequences in Escherichia coli, Nucleic Acids Res. 38 (2010) e88.

[2] L. Hashemzadeh-Bonehi, F. Mehraein-Ghomi, C. Mitsopoulos, J.P. Jacob,E.S. Hennessey, J.K. Broome-Smith, Importance of using lac rather than ara pro-moter vectors for modulating the levels of toxic gene products in Escherichia coli,Mol. Microbiol. 30 (1998) 676–678.

[3] A. Lewin, M. Mayer, J. Chusainow, D. Jacob, B. Appel, Viral promoters can initiateexpression of toxin genes introduced into Escherichia coli, BMC Biotechnol. 5(2005) 19.

[4] K. Vu, J. Bautos, M.P. Hong, A. Gelli, The functional expression of toxic genes:lessons learned from molecular cloning of CCH1, a high-affinity Ca2+ channel,Anal. Biochem. 393 (2009) 234–241.

[5] E. Cascales, S.K. Buchanan, D. Duché, C. Kleanthous, R. Lloubès, K. Postle, M. Riley,S. Slatin, D. Cavard, Colicin biology, Microbiol. Mol. Biol. Rev. 71 (2007) 158–229.

[6] C. Montigny, F. Penin, C. Lethias, P. Falson, Overcoming the toxicity of membranepeptide expression in bacteria by upstream insertion of Asp-Pro sequence, Biochim.Biophys. Acta 1660 (2004) 53–65.

[7] X. Cheng, G. Liu, G. Ye, H. Wang, X. Shen, K. Wu, J. Xie, I. Altosaar, Screening andcloning of antimicrobial DNA sequences using a vital staining method, Gene 430(2009) 132–139.

[8] R. Sorek, Y. Zhu, C.J. Creevey, M.P. Francino, P. Bork, E.M. Rubin, Genome-wideexperimental determination of barriers to horizontal gene transfer, Science 318(2007) 1449–1452.

[9] N.B. Mabizela-Mokoena, S.W. Limani, I. Ncube, L.A. Piater, D. Litthauer,M.B. Nthangeni, Genetic determinant of Bacillus pumilus lipase lethality and itsapplication as positive selection cloning vector in Escherichia coli, Protein Expr.Purif. 137 (2017) 43–51.

[10] Q.Y. Sun, L.W. Ding, L.L. He, Y.B. Sun, J.L. Shao, M. Luo, Z.F. Xu, Culture ofEscherichia coli in SOC medium improves the cloning efficiency of toxic proteingenes, Anal. Biochem. 394 (2009) 144–146.

[11] L.C. Anthony, H. Suzuki, M. Filutowicz, Tightly regulated vectors for the cloningand expression of toxic genes, J. Microbiol. Methods 58 (2004) 243–250.

[12] K. Kwon, J. Hasseman, S. Latham, C. Grose, Y. Do, R.D. Fleischmann, R. Pieper,S.N. Peterson, Recombinant expression and functional analysis of proteases fromStreptococcus pneumoniae, Bacillus anthracis, and Yersinia pestis, BMC Biochem.12 (2011) 17.

[13] A. Azarnezhad, Z. Sharifi, R. Seyedabadi, A. Hosseini, B. Johari, M. Sobhani Fard,Cloning and expression of soluble recombinant HIV-1 CRF35 protease-HP thior-edoxin fusion protein, Avicenna J. Med. Biotechnol. 8 (2016) 175–181.

[14] H. Sberro, A. Leavitt, R. Kiro, E. Koh, Y. Peleg, U. Qimron, R. Sorek, Discovery offunctional toxin/antitoxin systems in bacteria by shotgun cloning, Mol. Cell 50(2013) 136–148.

[15] A. Jamet, X. Nassif, New players in the toxin field: polymorphic toxin systems inbacteria, MBio 6 (2015) (e00285-15).

[16] H. Li, Ch. Hao, D. Xu, Development of a novel vector for cloning and expressingextremely toxic genes in Escherichia coli, Electron. J. Biotechnol. 30 (2017) 88–94.

[17] A. Jamet, X. Nassif, Characterization of the Maf family of polymorphic toxins inpathogenic Neisseria species, Microb. Cell 2 (2015) 88–90.

[18] F. Saïda, M. Uzan, B. Odaert, F. Bontems, Expression of highly toxic genes in E. coli:special strategies and genetic tools, Curr. Protein Pept. Sci. 7 (2006) 47–56.

[19] A. Kimelman, A. Levy, H. Sberro, S. Kidron, A. Leavitt, G. Amitai, D.R. Yoder-Himes,O. Wurtzel, Y. Zhu, E.M. Rubin, R. Sorek, A vast collection of microbial genes thatare toxic to bacteria, Genome Res. 22 (2012) 802–809.

[20] J. Edmonds, E. van Grinsven, N. Prow, A. Bosco-Lauth, A.C. Brault, R.A. Bowen,R.A. Hall, A.A. Khromykh, A novel bacterium-free method for generation of flavi-virus infectious DNA by circular polymerase extension reaction allows accuraterecapitulation of viral heterogeneity, J. Virol. 87 (2013) 2367–2372.

[21] S. Schlegel, M. Klepsch, D. Gialama, D. Wickström, D.J. Slotboom, J.W. de Gier,Revolutionizing membrane protein overexpression in bacteria, Microb. Biotechnol.3 (2010) 403–411.

[22] I.E. Johansen, O.S. Lund, Insertion of introns: a strategy to facilitate assembly ofinfectious full length clones, Methods Mol. Biol. 451 (2008) 535–544.

[23] W. Jarmuszkiewicz, A.M. Wagner, M.J. Wagner, L. Hryniewiecka, Immunologicalidentification of the alternative oxidase of Acanthamoeba castellanii mitochondria,FEBS Lett. 411 (1997) 110–114.

[24] M. Clarke, A.J. Lohan, B. Liu, I. Lagkouvardos, S. Roy, N. Zafar, C. Bertelli,C. Schilde, A. Kianianmomeni, T.R. Bürglin, C. Frech, B. Turcotte, K.O. Kopec,J.M. Synnott, C. Choo, I. Paponov, A. Finkler, C.S. Heng Tan, A.P. Hutchins,T. Weinmeier, T. Rattei, J.S. Chu, G. Gimenez, M. Irimia, D.J. Rigden,D.A. Fitzpatrick, J. Lorenzo-Morales, A. Bateman, C.H. Chiu, P. Tang, P. Hegemann,H. Fromm, D. Raoult, G. Greub, D. Miranda-Saavedra, N. Chen, P. Nash,M.L. Ginger, M. Horn, P. Schaap, L. Caler, B.J. Loftus, Genome of Acanthamoebacastellanii highlights extensive lateral gene transfer and early evolution of tyrosinekinase signaling, Genome Biol. 14 (2013) R11.

[25] A. Bieluszewska, M. Weglewska, T. Bieluszewski, K. Lesniewicz, E. Poreba, PKA-binding domain of AKAP8 is essential for direct interaction with DPY30protein,FEBS J. 285 (2018) 947–964.

[26] K. Lesniewicz, E. Poreba, M. Smolarkiewicz, N. Wolff, S. Stanisławski, P. Wojtaszek,Plant plasma membrane-bound staphylococcal-like DNases as a novel class of eu-karyotic nucleases, BMC Plant Biol. 12 (2012) 195.

[27] K. Leśniewicz, J. Pieńkowska, E. Poreba, Characterization of nucleases involved inseedling development of cauliflower, J. Plant Physiol. 167 (2010) 1093–1100.

[28] L. Yang, X. Wang, W. Deng, W. Mo, J. Gao, Q. Liu, C. Zhang, Q. Wang, C. Lin,Z. Zuo, Using HEK293T expression system to study photoactive plant crypto-chromes, Front. Plant Sci. 7 (2016) 940.

[29] M.I. Santori, C. Gonzalez, L. Serrano, M. Isalan, Localized transfection with mag-netic beads coated with PCR products and other nucleic acids, Nat. Protoc. 1 (2006)526–531.

R. Krela et al. BBA - General Subjects 1863 (2019) 521–527

527