Plasmid Curing is a Promising Approach to Improve Thermophiles for Biotechnological Applications: Perspectives in Archaea Plasmid Curing is a Promising Approach to Improve Thermophiles for Biotechnological Applications: Perspectives in Archaea

Thermophiles are attractive as host cells for microbial processes to produce or degrade various compounds. In these applications, it is often desirable to improve the proper ties of thermophiles, such as their growth rate, cell density, and protein productivity, although this is rarely achieved because of the lack of general approaches. In this chap -ter, we describe the elimination of the pHTA426 plasmid from a moderate thermophile, Geobacillus kaustophilus HTA426, and its effects on the microbial properties. This process, called plasmid curing, was simply achieved using a DNA intercalator and confirmed by phenotypic and genotypic analyses. Of note, pHTA426 curing had beneficial effects on diverse properties, probably because of the reduced energy burden in terms of plasmid replication at high temperatures. The result suggests that plasmid curing is a simple and versatile approach for improving thermophiles. In particular, this approach may be effec - tive for archaeal thermophiles because they grow at much higher temperatures and could have the greater energy burden on plasmid replication. Data mining has also shown that plasmids are distributed in archaeal thermophiles. This chapter provides a new tip for improving archaeal thermophiles, thereby increasing the opportunities for their use in various biotechnological applications. (3.7 ± 0.4 Mb; n = 6), and Halorubrum (3.1 ± 0.8 Mb; n = 13) members. These data suggest that the thermophiles, especially archaeal thermophiles, tend to have smaller genomes than mesophiles.


Introduction
Thermophiles are organisms that are capable of growing at temperatures above 55°C. Archaeal thermophiles generally grow at much higher temperatures and thus comprise most extreme thermophiles and hyperthermophiles. These thermophiles are attractive organisms in biotechnological applications because they produce thermostable enzymes, which can be used as stable industrial catalysts even at high temperatures. Thermophile cells themselves are attractive as hosts for microbial processes at high temperatures.
High-temperature processes have several advantages compared with moderate processes using mesophiles, where an important advantage is that high temperature prevents the growth of animal pathogens, including all viruses, which are killed or at least prevented from proliferating at temperatures above 65°C [1]. High temperature also inhibits growth and/or metabolism by mesophiles, which may hinder processes of interests via involuntary reactions. The advantage is especially important for processes using crude biomass (e.g., sewage, municipal or agricultural waste, and materials from animal farms) because mesophiles and virulent pathogens are common in biomass from natural environments and they may increase during reactions performed under moderate conditions. In addition, high temperature facilitates the removal of volatile products (e.g., ethanol and butanol) while decreasing oxygen solubility; therefore, thermophiles are practical for fermentative production of alcohols [2][3][4][5][6]. Moreover, thermophiles often have remarkable properties useful for bioprocesses. A good example is the hyperthermophilic archaeon Thermococcus kodakarensis KOD1, which can use protons as an electron accepter in catabolism to generate molecular hydrogen and has been studied as a hydrogen production tool [7].
In this chapter, we will demonstrate the salutary effects of plasmid curing on thermophiles using as an example a prokaryotic thermophile that was isolated from deep sea sediments of the Mariana Trench, Geobacillus kaustophilus HTA426 [8,9]. Its growth occurs at temperatures ranging between 42 and 74°C, with optimal growth at 60°C, and is as rapid as Escherichia coli and Bacillus subtilis. Genetic tools are available for this strain [10][11][12][13][14][15][16]. The whole genome sequence has been determined [17], showing that G. kaustophilus HTA426 harbors the circular plasmid pHTA426 (Figure 1). Because the members of the genus Geobacillus include strains that are useful for high-temperature processes, as demonstrated by strains that are capable of degrading hydrocarbons [18], long-chain alkanes [19][20][21], biphenyls [22], paraffin-wax [23], or nylons [24], we have studied biotechnological applications of the genus using G. kaustophilus HTA426 as a model and pilot strain.
Aiming to construct a plasmid-free strain that may be useful for the genetic analysis of pHTA426, this study was originally designed to eliminate this plasmid from an HTA426 derivative, G. kaustophilus MK244. The resultant strain MK633 was then characterized to confirm that plasmid elimination (termed plasmid curing) had no effects on its microbial properties, but the analysis unexpectedly demonstrated that G. kaustophilus MK633 had advantages compared with the parent strain MK244 in terms of several properties. We suggest here that plasmid curing is a promising approach for improving diverse thermophiles.

Plasmid introduction into G. kaustophilus
Plasmids were introduced into G. kaustophilus by conjugative plasmid transfer from E. coli BR408 [12]. Briefly, an E. coli donor (10 ml) and a G. kaustophilus recipient (100 ml) were grown in LB media. The cells were subsequently mixed, centrifuged, and spotted onto LB plates.
After incubation at 37°C for 20 h, the resultant cells were collected and incubated at 60°C on appropriate media to isolate G. kaustophilus transformants. The transformation efficiency, as the number of transformants per 10 6 recipients, was determined as described previously [12]. Data were expressed as the mean ± standard error (n = 3).

pHTA426 curing from G. kaustophilus MK244
The gkp09 gene in pHTA426 was replaced by a bgaB expression cassette using pGKE25-bgaB with pyrF-based counterselection [13]. The resultant clone, strain MK244′, was successively cultured three times in LB media supplemented with 20 µM acridine orange to facilitate pHTA426 curing. In each culture, an aliquot (10 3 cells) was grown on LB plates supplemented with X-gal to identify candidates from which pHTA426 was eliminated along with the bgaB cassette. The candidates were purified using LB plates with X-gal.

Construction of G. kaustophilus MK244 bgaB and MK633 bgaB
A bgaB expression cassette under the control of the P gk704 promoter was integrated at the gk0707 locus in G. kaustophilus using pGAM48-bgaB, as described previously [15]. G. kaustophilus MK244 and MK633 were subjected to this process to generate strains MK244 bgaB and MK633 bgaB , respectively.

BgaB assay
G. kaustophilus MK244 bgaB and MK633 bgaB were cultured for 4 h in MM containing yeast extract but not D-glucose or casamino acids, and then for 20 h in the presence of 10 g l −1 maltose. Cells were subsequently harvested, sonicated in 50 mM sodium phosphate (pH 6.0), and clarified by centrifugation to obtain a lysate. The reaction mixture (100 µl) contained 50 mM sodium phosphate (pH 6.0), 2 mM p-nitrophenyl-β-D-galactopyranoside, and the lysate. The mixture was incubated at 60°C to react and then diluted with ice-cold 2 M sodium carbonate (900 µl) to terminate the reaction. p-Nitrophenol liberated in the mixture was quantified based on the absorbance at 405 nm and using an experimentally derived standard curve. One unit was defined as the amount of enzyme required to generate 1 µmol of p-nitrophenol per min. Proteins were quantified by the Bradford method using a protein assay kit (Nacalai Tesque, Kyoto, Japan). Data were expressed as the mean ± standard error (n = 4-5).

Plasmid stability assay
G. kaustophilus harboring pUCG18T was precultured in LB medium with kanamycin until the optical cell density at 600 nm reached 0.5. An aliquot (200 µl) was then cultured in LB medium (20 ml) without kanamycin until the stationary phase. The resultant cells were incubated on LB plates with or without kanamycin to determine the concentrations of kanamycin-resistant and viable cells, respectively. The plasmid retention rate was defined as the number of kanamycin-resistant cells per viable cells. Data were expressed as the mean ± standard error (n = 3).

Cell density assay
G. kaustophilus cells were cultured in LB medium until the stationary phase. Cells were harvested by centrifugation and analyzed to determine the wet weight. Data were expressed as the mean ± standard error (n = 4-5).

Mutation frequency assay
The frequency of spontaneous mutations was assessed based on the generation of rifampicin-and streptomycin-resistant cells via rpoB and rpsL genes, respectively [26]. G. kaustophilus (10 3 cells) was cultured at 60°C in LB medium until the stationary phase. The resultant cells were then incubated on LB plates with or without efficacious rifampicin or streptomycin (10 mg l −1 ) to determine the concentrations of mutant (rifampicin-or streptomycin-resistant) and viable cells, respectively. The colonies were counted to calculate the ratio of mutant cells relative to the viable cells incubated, which was defined as the mutation frequency. Data were expressed as the mean ± standard error (n = 3).

Nucleotide stability assay
Deoxyribonucleoside triphosphates (1 mM) were incubated for 24 h in 20 mM sodium phosphate (pH 7.0) at 30, 60, 80, and 90°C. The residual nucleotides in samples (5 µl) were analyzed using reversed-phase high-performance liquid chromatography. The chromatography system comprised solvent delivery units (LC-10AT; Shimadzu, Kyoto, Japan), an ultraviolet absorption detector (SPD-10Avp; Shimadzu), a reverse-phase column (Cosmosil 2.5C 18 -MS-II; Nacalai Tesque), and a column bath at 30°C. Solvents A and B comprised 5 mM tetrabutylammonium bromide in 20 mM sodium phosphate (pH 7.0) and 90% (v/v) acetonitrile in water, respectively. After injecting the sample into a column that had been equilibrated with 15% solvent B, the column was isocratically developed at a flow rate of 0.5 ml min −1 for 1 min and then at a linear gradient of 15-60% solvent B over 15 min. The chromatogram was obtained by detection at 260 nm.

Genome data mining
Genome data were collected from the GenBank database

Genetic features of the pHTA426 plasmid
The pHTA426 sequence suggested that the plasmid was a large circular plasmid (47.9 kb) comprising 1.3% of the circular chromosome of G. kaustophilus HTA426 (3.54 Mb) and that it encoded possible proteins for plasmid replication (gkp02) and plasmid partition (gkp36). A partition system has a role in the stable transmission of single-copy plasmids during cell division, so pHTA426 appeared to be present as a single copy in G. kaustophilus HTA426. In addition, the plasmid contained genes for a type II restriction-modification system, which was homologous to the AlwI restriction-modification system (gkp08 encoding methyltransferase and gkp09 encoding endonuclease). A type II restriction-modification system generally comprises an endonuclease and methyltransferase, where the endonuclease cuts exogenous DNA at specific sites, but not endogenous DNA that has been methylated by methyltransferase. In the AlwI system, AlwI methyltransferase is responsible for 5′-GG 6m ATC-3′ and 5′-G 6m ATCC-3′ methylation ( 6m A, N 6 -methyladenin), whereas AlwI endonuclease cuts 5′-GGATC-3′ and 5′-GATCC-3′ sites but not 5′-GG 6m ATC-3′ and 5′-G 6m ATCC-3′ sites. Because plasmids carrying a type II restriction-modification system have greater segregational stability [27], it is likely that the gkp08-gkp09 system contributes to the stable maintenance of pHTA426.  Figure 2A shows the process employed to eliminate pHTA426 from G. kaustophilus MK244.

Construction of G. kaustophilus MK633
To facilitate plasmid curing and readily identify positive clones from which pHTA426 was eliminated, the gkp09 gene was preliminarily replaced by a bgaB expression cassette using pGKE25-bgaB. The resultant strain MK244′ was then cultured successively in the presence of a DNA intercalator, and we screened for positive clones by using the X-gal degradation assay. Fortunately, one positive clone was obtained from the first culture but not from the second culture. From the third culture, 24 positive clones were identified, which suggests that three successive rounds of culture were effective for pHTA426 curing. The positive clone obtained from the first culture was designated as G. kaustophilus MK633.
G. kaustophilus MK244′ degraded X-gal to form blue colonies on LB plates with X-gal, whereas strains MK244 and MK633 did not ( Figure 2B). This suggests that strain MK633 lacked the bgaB gene. The MK633 chromosome is resistant to DpnI (which digests 5′-G 6m ATC-3′ but not 5′-GATC-3′) but sensitive to AlwI (see above), in contrast to the chromosomes from strains MK244 and MK244′ (Figure 2C), so it is likely that strain MK633 lacked the gkp08 gene encoding an AlwI methyltransferase homolog. Southern blot analysis ( Figure 2D) confirmed that strain MK633 lacked the bgaB and gkp30 genes, which are located on opposite sides of pHTA426 (Δgkp09::P sigA -bgaB).
Based on these results, we concluded that G. kaustophilus MK633 lacked pHTA426.

Microbial properties of G. kaustophilus MK244 and MK633
G. kaustophilus MK244 and MK633 were characterized in detail ( Table 3). Both strains grew in LB and MM with comparable doubling times. The difference in their mutation frequencies was also not significant. However, strain MK633 was more transformed efficiently with pUCG18T than strain MK244 and it maintained the plasmid with higher stability. Moreover, strain MK633 grew at higher cell densities than strain MK244. The cell density of G. kaustophilus MK244′ was lower than that of strain MK633 at 60°C (0.30 ± 0.03 g wet), but the pUCG18T retention rate was comparable (43 ± 14%).

Archaeal thermophiles: Smaller genomes
Genomic data were analyzed to compare the genome sizes of thermophiles (capable of growing at > 55°C) and mesophiles (capable of growing at 20-55°C). The genomes of thermophilic bacteria Geobacillus spp.

Distribution of plasmids in archaeal thermophiles
Data mining showed that many thermophiles harbored plasmids, although not the majority.
In Archaea, plasmids were identified frequently in Sulfolobus spp.   Table 3. Microbial properties of G. kaustophilus MK244 and MK633. Analyses were repeated more than three times. Data represent the mean ± standard error. ND, growth was not observed within 2 days. In addition, several cryptic plasmids have been identified in archaeal thermophiles [28].

Discussion
In this study, we analyzed the effects of plasmid curing on thermophiles by characterizing G. kaustophilus MK244 and MK633 ( Table 3 and Figure 3). Both strains exhibited comparable growth at 50-70°C in LB and MM; therefore, pHTA426 had no positive effects on cell growth under standard conditions. The mutation frequencies were largely comparable between strains MK244 and MK633, although G. kaustophilus MK633 lacked the gkp08 gene responsible for dam-like methylation (5′-GG 6m ATC-3′and 5′-G 6m ATCC-3′) and the dam methylation (5′-G 6m ATC-3′) is essential for DNA mismatch repair in E. coli [29]. In B. subtilis 168, DNA mismatch repair only involves mutS and mutL products, which do not depend on DNA methylation [30,31], thereby suggesting that G. kaustophilus may use a mismatch repair system similar to the B. subtilis system rather than the E. coli system. In fact, the G. kaustophilus genome [17] contains mutS and mutL but not mutH, as found in B. subtilis 168.
A restriction-modification system generally protects the host microbe from transformation with exogenous DNA because the system cuts exogenous DNA that is not methylated by methyltransferase. However, a microbe can accept exogenous DNA that imitates the methylation pattern because a restriction-modification system is unable to cut this exogenous DNA [32]. In a previous study [12], we constructed E. coli strains BR397, BR398, and BR408 for conjugative plasmid transfer into G. kaustophilus HTA426. E. coli BR408 produces DNA that imitates the methylation pattern in G. kaustophilus HTA426. E. coli BR398 produces DNA with dam methylation, whereas E. coli BR397 produces methyl-free DNA. Although G. kaustophilus MK244 could not accept the pUCG18T plasmid transferred from the dam − strain E. coli BR397, G. kaustophilus MK633 accepted pUCG18T from E. coli BR397 as well as strains BR398 and BR408. These results can be explained by the elimination of gkp09 from strain MK633 because the gkp09 product digests methyl-free DNA but not DNA with dam methylation (5′-G 6m ATC-3′, which covers gkp08 methylation (5′-GG 6m ATC-3′ and 5′-G 6m ATCC-3′. In addition, G. kaustophilus MK633 maintained pUCG18T with higher stability than strain MK244. This observation is also attributable to the elimination of gkp09 because the gkp09 product can occasionally digest endogenous plasmids that have not undergone gkp08 methylation immediately after plasmid replication. This hypothesis is supported by the fact that G. kaustophilus MK244′ maintained pUCG18T as stably as strain MK633. Thus, G. kaustophilus MK633 acquired advantages compared with strain MK244 in terms of plasmid transformation and plasmid stability due to the elimination of a restriction-modification system along with pHTA426 curing. An intriguing observation was that G. kaustophilus MK633 had a higher cell density in the stationary phase. In contrast to the advantages in terms of plasmid transformation and plasmid stability, this observation cannot be explained by the elimination of gkp09 because G. kaustophilus MK244′ grew less efficiently than strain MK633. It was also interesting that strain MK633 bgaB produced higher amounts of BgaB than MK244 bgaB (Figure 3A). This observation is attributable mainly to the higher cell density of MK633 bgaB (Figure 3B), as observed with strain MK633, and thus the higher protein yields from strain MK633 bgaB (Figure 3C). In addition, this observation can be attributed to the higher BgaB productivity per cell because strain MK633 bgaB had a higher BgaB specific activity ( Figure 3D). Thus, these results suggest that the elimination of pHTA426 improved the cell density per culture and BgaB productivity per cell, thereby remarkably enhancing the production of BgaB.
Sodium dodecyl sulfate-polyacrylamide gel electrophoresis analysis showed that the crude extracts from G. kaustophilus MK244 and MK633 had negligible differences in terms of their protein compositions (data not shown). Therefore, it is unlikely that G. kaustophilus MK633 lost a nonessential protein that was abundantly synthesized from pHTA426, thereby reducing the energy burden to enhance the cell yield and BgaB productivity. Instead, these enhancements may be attributable to the reduced energy burden for plasmid replication.
In fact, evidence suggests that plasmid maintenance places burdens on cell growth and/or metabolism [33][34][35][36]. Given that pHTA426 replicates as a single copy, this plasmid accounts for only 1.3% of total DNA; however, deoxyribonucleoside triphosphates are more unstable at higher temperatures, thereby suggesting that plasmid replication may place a considerable energy burden on thermophiles even though the plasmid is not extremely large or present in high copy numbers. This hypothesis is consistent with our genome data analysis, which showed that thermophiles have relatively smaller genomes than mesophiles, and the negative correlation between genome size and growth temperature reported by Sabath et al [37]. Overall, we consider that the lower energy burden incurred for DNA replication can explain why pHTA426 curing improved the cell density per culture and BgaB productivity per cell.
In conclusion, we demonstrated that pHTA426 curing was effective for improving the performance of a moderate thermophile, G. kaustophilus MK244. The cell density and protein productivity were presumably improved by the reduced amounts of energy required for DNA replication at high temperatures, so plasmid curing may be a simple approach for improving thermophiles in terms of these properties. In particular, this approach may be effective for archaeal thermophiles because they grow at extremely high temperatures and thus could have a greater energy burden on plasmid replication. Moreover, in archaeal thermophiles, plasmids may account for larger fractions of chromosomes than in moderate thermophiles because archaeal thermophile genomes are generally smaller. Therefore, plasmid curing could remarkably reduce the energy burden in archaeal thermophiles. We note that many archaeal thermophiles harbor plasmids, such as S. solfataricus P2 (carrying pSSVi) and T. barophilus MP (carrying pTBMP1), which have been studied as model acidophilic hyperthermophiles [38] and piezophilic hyperthermophiles [39], respectively. Even if these thermophiles harbor a single copy plasmid, our results suggest that plasmid curing can improve their performance in terms of the cell density and protein productivity.

Author details
Tatsuki Mizuno, Takashi Ohshiro and Hirokazu Suzuki* *Address all correspondence to: hirokap@xpost.plala.or.jp Department of Chemistry and Biotechnology, Graduate School of Engineering, Tottori University, Japan