Restriction EnzymeInduction in Escherichia coli.

Kenji Higashi,　 GraduateSchool of Media and Governance

Abstract

　ThePaeR7I endonuclease encoding gene was clonedinto an inducible expression vector, resulting in an inducible genomedestruction system.　 Escherichia colihost genome degradation triggered by the inductionof this system was observed by time series.　Host cells that did not cease growth after induction of endonucleasewere found at a frequency of approximately one in 50,000 cells.

Introduction

　Theaim of this work is to create a system in which the Escherichia coli genome can be reliably destructed through the induction of arestriction enzyme.　 It may be possibleto introduce genes artificially accumulated onto a plasmid or a BAC intoresulting genomeless cells and to make them function as artificial genomes.　 A thoroughly characterized genomeless cellcould become a convenient simple target for whole cell simulation.

　Endonuclease induction systems have been created by variousresearchers in the past [3,4,5,7,8,9].　These authors created these systems with purposes such as to demonstratethe stringency accomplished by inducible expression systems that they created,or to analyze DNA repair that occurs in response to endonuclease scission.　 These systems were not designed to create ahomogenous population of genomeless cells.　

　Asa preliminary attempt to construct such a reliable system, an endonucleaseinduction system was created in this study by cloning the endonuclease of the PaeR7I restriction modification system of Pseudomonas aeruginosa [1] into pZS*24-MCS1. pZS*24-MCS1[3] is a stringently repressibleexpression vector that utilizes the artificial promoter P_lac/ara-1,which combines Lac Operators and AraC binding sites to accomplish tightregulation.　 pZS*24-MCS1 possesses akanamycin resistance gene and has a modified pSC101 origin which is known toprovide 3-4 copies per host cell[2].　　 P_lac/ara1is inducible by adding IPTG (isopropyl-beta-D-thiogalactopyranoside) and/orL(+)Arabinose to the medium.　

Figure 1.　 Overview of thegenome destruction strategy taken in this work.

Materials &Methods

Bacterial strains and plasmids

　E. coli K-12 strains used in this workwere DH5α, DH5αZ1 (DH5α lacI^q Sp^R), andYF001 (DH5αharboring plasmid pPAORM3.8).　Expression vector plasmid pZS*24-MCS1 was obtained from professor HermanBujard of the university of Heidelberg under a material transferagreement.　 Plasmid pPAORM3.8 wasmini-prep purified from YF001 which was obtained from Dr. Yoko Fukuda of theuniversity of Tokyo.

Enzymes, chemicals & Oligonucleotides

　Restrictionenzymes ClaI, KpnIand MluI were obtained from Takara Bio Inc.Ligation High from Toyobo was used for ligation reactions. Reagent chemicalswere purchased from Wako and Oligonucleotides were from Hokkaido SystemScience.　 Kanamycin was used in culturemedium at a final concentration of 40µg/ml, and spectinomycin at50µg/ml.　 L(+)Arabinose (Arab.)was applied at a final concentration of 0.5%.　

Cloningof PaeR7I endonuclease

　The vector plasmid pZS*24-MCS1 wastransformed by electroporation into DH5αhost cells and midi-prep purified after overnight amplification in500ml LB/Km medium.

　 Theendonuclease encoding gene PaeR7I.R of the PaeR7I restriction modification system was PCR amplified from templateplasmid pPAORM3.8, which was mini-prep purified after its endonuclease activitywas confirmed by a lambda phage plaque forming unit assay.　 Appropriate restriction enzyme recognitionsequences were introduced into the ends of the DNA fragment through 5’ ends ofthe PCR primers.　 The coding region ofthe endonuclease gene was incorporated into the multi-cloning site ofpZS*24-MCS1 in two slightly different positions as illustrated in figure2.　　 The first design in which PaeR7I.R is inserted via KpnI-MluI allows the endonuclease gene to utilize a ribosomal binding site(RBS) that originates from the vector.　In the second design the endonuclease gene is inserted via ClaI-MluI, resulting in the deletion ofthe RBS.　 The ClaI-MluI design is intended to have anexpression rate substantially lower than that of the KpnI-MluI design, providing even lessbackground expression under repressed conditions. 　　The two clones were constructed accordingly and named pZS*24-PaeR and pZS*24-PaeRΔRBS respectively.　 Constructs were transformed into DH5αZ1 competent cells.

Figure2.

Genomic DNA degradation assay

　Thetime course of Genomic DNA degradation in endonuclease induced cells wasobserved by submarine electrophoresis and pulse-field electrophoresis of DNAextracted from cultures of gradient induction time lengths.　 DNA samples for submarine electrophoresiswere prepared as follows: 1ml of overnight pre-culture of DH5αZ1(pZS*24-PaeR) was inoculated into 125ml LB/km in a 500ml flask and incubated at37℃.　 After 6 hours, a 10ml aliquot was transferedfrom the flask into a 15ml falcon tube and immediately placed on ice.　 Without delay, L(+)Arabinose and IPTG wereadded to the culture at　 finalconcentrations of 0.5% and 10mM respectively, inducing expression of PaeR7I.R.　 Additional 10mlaliquots were taken from the culture at 5, 10, 30 and 60 minutes afterinduction.　 Each aliquot was pelletedand DNA samples for electrophoresis were exracted by mini-prep.　 Pulse-field electrophoresis samples wereprepared using the same procedures except that aliquots were taken at 0, 6, 9,12,15, 30 and 70 minutes from induction and that they were used to creategel-blocks after pelleting.　 Pulse fieldelectrophoresis was conducted for 21hours at 90V with 12second intervals.

Results &Discussion

Endonuclease Activity Screening

　pZS*24-PaeR and pZS*24-PaeRΔRBStransformants of DH5αZ1 were selected on LB/km plates, and 8 colonies each were furtherscreened for endonuclease activity by patch inoculation into 1ml of LB/km(repressed) and 1ml of LB/km/IPTG(1mM)/Arab (induced) liquid medium.　 After 10 hours of shaking at 37℃, culture turbidity wasvisually observed to screen isolates that fail to proliferate in induced mediawhile growing when inoculated into repressed medium. 　4 out of the 8 pZS*24-PaeR carryingcells that were tested met this screening criterion.　 All 8 of the pZS*24-PaeRΔRBS carrying cells thatwere tested grew to saturation in both repressed and induced mediums. Thissuggests that pZS*24-PaeR is capable ofrepression/expression of PaeR7I.R, and that theRBS lacking construct is incapable of expressing the endonuclease at a levelsufficient to arrest host growth.　Growth arrest of pZS*24-PaeR harboringcells was observed in LB/km/IPTG(1mM)/Arab medium, but not in eitherLB/km/IPTG(1mM) or LB/km/Arab.　

Endonuclease Induction Survivors

　Endonuclease activity screening was replicated with 24 differentpZS*24-PaeR transformant colonies, and successful growth arrest under inducedconditions was observed in 22 of these 24 isolates after 10 hours incubation at37℃.　 When these IPTG/Arab induced DH5αZ1(pZS*24-PaeR) cultures were continuously shaken at 37℃ for an additional 14hours, all 24 cultures propagated to saturation.　 This indicates that a fraction of the inoculated cells survive PaeR7I.R induction, and fail to lose proliferation potential.　

　Abundance of such survivors is problematic since it is importantto obtain a homogenous population of cells with destructed genomes in order toanalyze the properties of genome less cells, or to use the cells as platformsfor expression of artificial genomes.　The rate of cells that survive endonuclease induction was assayed.　 Results are shown in table 1.

Table 1.	LB/km/IPTG/Arab	LB/km	Rate of survivors
Sample A -- Dilution1	63	1.6E+6	3.94E-5
Sample A -- Dilution2	20	6E+5	3.33E-5
Sample B -- Dilution1	20	7E+5	2.86E-5
Sample B -- Dilution2	10	8E+5	1.25E-5

Early log phasecultures A and B of DH5αZ1(pZS*24-PaeR) were each diluted in two series andspread onto LB/km and LB/km/IPTG(10mM)/Arab plates.　　

　Onepossible cause of these abundant survivors is the frequent occurrence ofmutations in the Restriction enzyme gene. Another possibility is the frequentloss of the endonuclease coding plasmid following a mutation causing kanamycinresistance.　 The high concentration ofkanamycin used in this experiment (40µg/ml) is a factor that is likely toincrease the chance of such a mutation to occur.

　Loss of the PaeR7I.R gene due tohomologous recombination among sibling copies of the plasmid is unlikely to beoccurring since a recA1 host is being used.

Time course of GenomicDegradation

　Results of the time course observation of genomic degradation areshown in figures 3,4 and 5.

　Genomic DNA is not visible in the submarine electrophoresissamples in figure 3 at ±0 minutes induction time. This is because the genome is not digestedyet at these times, and intact genomic DNA cannot elute from a mini-prepcolumn.　 Bands of digested genomic DNAbecome visible at 10 minutes after induction, and at 30 and 60 minutes frominduction these bands seem to be degraded into smears.　 Gradual degradation of the plasmid pZS*24-PaeR can be observed as well in figure 3.　 pZS*24-PaeR contains one PaeR7I recognition site.　 All ofthe results noted here apply to both samples A & B of a duplicateexperiment.

　Genomic DNA digestion and gradual degradation was once againconfirmed by the results of pulse-field electrophoresis (results shown infigures 4 and 5.)

Figure 3 Result of submarine electrophoresis of mini-prep purifiedinduction time gradient samples.

The white arrowindicates the position of the plasmid pZS*24-PaeR. Lanes are: (left to right)φX174-HaeIII digest,

SampleA (-0, +0,5min, 10min, 30min, 60min induction), 100bp ladder DNA,φX174-HaeIII digest,

SampleB (-0, +0,5min, 10min, 30min, 60min induction), 100bp ladder DNA

Figure 5

Figure 4

　

Figure 4 Lanesare: (left to right) λHindIII digest,λconcatemer,λDNAladder,　 SampleA (0min, 6min, 9min,12min, 15min, 30min, 70min, λconcatemer,λHindIII digest　　　　　

Figure 5　 Lanes are: (left to right) λconcatemer, λDNAladder,　 SampleB (0min, 6min, 9min,12min, 15min, 30min, 70min, λconcatemer,λHindIII digest

Time course of viable kmresistant cells

　Two late logphase cultures (A and B) of DH5αZ1(pZS*24-PaeR) were each dilutedinto two series and spread onto LB/km plates (time 0).　 Inducers were then added (10mM IPTG,0.5%Arab), and aliquots were taken from the culture, diluted and spread onLB/km plates at 6, 12, 20, 40, 80, 150 and 300 minutes after induction.Colonies formed on these plates were counted after 20 hours of incubation at 37℃. Results (figure 6) showthat the number of viable kanamycin resistant cells decreased greatly in thefirst 6 minutes after endonuclease induction, and continued to decrease between80 and 150 minutes after induction.　This implies that delayed genome destruction occurs in a fraction of thecells.　 The cause of this delay isunknown, but an insufficient endonuclease induction strength as a consequencetraded off for the tight repression in the non-induced state could beinvolved.　 Another possible cause is theall-or-none expression found in araC controlled genes due to positive feedbackof araE permease expression [6].

Figure 6. 　　　　　　　 A1:culture A, dilution series #1

　　　　　　　　　　　　　　　 A2:culture A, dilution series #2

　　　　　　　　　　　　　　　 B1:culture B, dilution series #1

　　　　　　　　　　　　　　　 B2:culture B, dilution series #2

Conclusion

　Ithas been confirmed in this work that the induction of an endonuclease gene in E.coli results in the degradation of genomic DNA andan arrest of cell growth. Furthermore, it has been found that in the case ofpZS*24-PaeR, the rate of cells that surviveendonuclease induction is less than 5e-5.　

　Cloningof an additional restriction enzyme under stringent repression into either thesame plasmid or into a compatible plasmid should be effective to greatly reducethe rate of cells that survive.　

Acknowledgements

　This work was performed at the Institute for Advanced Biosciencesunder the instruction of assistant prof. Yoichi Nakayama.　 Prof. Mitsuhiro Itaya, assistant prof.Tomoya Baba, Azusa Kuroki, Dongru Qiu and Dr. Yoshiaki Ohashi of HumanMetabolome Technologies, Inc. provided useful discussion and essentialmentorship.　 Prof. Herman Bujard at theuniversity of Heidelberg provided the plasmid vector pZS*24-MCS1 and thebacterial strain DH5αZ1.　 Dr. Yoko Fukuda at TokyoUniversity provided the bacterial strain YF001.　 Ryo Hattori and Hiromi Komai are members of this researchproject.　 Prof. Masaru Tomita providedcomprehensive support.

References

1.　　　　　　 Gingeras,T. R. & Brooks, J. E. (1983) Proc Natl Acad Sci U S A 80, 402-6.

2.　　　　　　 Manen,Det. al. (1994) Molecular Microbiology 11(5),875-884

3. Luts, R & Bujard, H. (1997) N.A.R. 25(6)1203-1210

4. O’Connor, C.D.& Humphreys, G.O. (1982) Gene, 20219-229

5. O’Connor, C.D.& Timmis, K.N. (1987) J. Bacteriol, 169(10),4457-4462

6. Siegel, D.A &Hu, J.C. (1997) Proc Natl Acad Sci U S A 94,8168-8172.

7. Heitman, J et. al.　　(1989) Proc Natl Acad Sci U S A 86,2281-2285

8. Kuhn I et. Al. (1986) Gene 44 253-263

9. Meddows, T.R. et. al. (2004) Molecular Microbiology 52 119-32