Development of the Multiple Gene Knockout System with One-Step PCR in Thermoacidophilic Crenarchaeon Sulfolobus acidocaldarius

Multiple gene knockout systems developed in the thermoacidophilic crenarchaeon Sulfolobus acidocaldarius are powerful genetic tools. However, plasmid construction typically requires several steps. Alternatively, PCR tailing for high-throughput gene disruption was also developed in S. acidocaldarius, but repeated gene knockout based on PCR tailing has been limited due to lack of a genetic marker system. In this study, we demonstrated efficient homologous recombination frequency (2.8 × 104 ± 6.9 × 103 colonies/μg DNA) by optimizing the transformation conditions. This optimized protocol allowed to develop reliable gene knockout via double crossover using short homologous arms and to establish the multiple gene knockout system with one-step PCR (MONSTER). In the MONSTER, a multiple gene knockout cassette was simply and rapidly constructed by one-step PCR without plasmid construction, and the PCR product can be immediately used for target gene deletion. As an example of the applications of this strategy, we successfully made a DNA photolyase- (phr-) and arginine decarboxylase- (argD-) deficient strain of S. acidocaldarius. In addition, an agmatine selection system consisting of an agmatine-auxotrophic strain and argD marker was also established. The MONSTER provides an alternative strategy that enables the very simple construction of multiple gene knockout cassettes for genetic studies in S. acidocaldarius.


Introduction
High-throughput PCR tailing for gene disruption has been developed in the thermoacidophilic crenarchaeon Sulfolobus acidocaldarius [1]. We attempted to improve this technique and develop an efficient multiple gene knockout strategy with a PCR tailing (one-step PCR) method.
Gene knockout via homologous recombination is a powerful tool for the generation of specific mutants and subsequent functional analysis of the gene. Three unmarked gene deletion methodologies, that is, plasmid integration and segregation (PIS), marker replacement and looping out (MRL), and marker insertion and unmarked target gene deletion (MID), have been employed in S. acidocaldarius and S. islandicus [2][3][4]. These pop-out recombination-based approaches are effective for multiple gene knockout [5][6][7], but plasmid construction is required. In contrast, one-step PCR followed by a marker replacement system using the pyrE selection marker flanked by 40-50 bp of homologous regions, for example, 5′ and 3′ flanking regions of the target gene, has been developed in S. acidocaldarius [1]. This PCR-tailing method allows for effective, high-throughput gene functional analysis without plasmid construction [1]. However, this method was not sufficient for repeated gene disruptions because only the uracil selection system (pyrimidine-auxotrophic strain and selectable marker [pyrE] gene) was available in S. acidocaldarius. A pop-out recombination system using one-step PCR for multiple gene knockout has not been reported in hyperthermophilic archaea. Furthermore, the homologous recombination efficiency using the PCR-tailing technique has not been reported [1].
We recently constructed the restriction endonuclease SuaI-deficient S. acidocaldarius strain SK-1 (ΔpyrE ΔsuaI), which has the potential for efficient and flexible direct modification of the genome using synthetic oligonucleotides or PCR products without any methylation procedures [8]. In our current study, we estimated the effects of transformation conditions (plating methods, DNA topology, CaCl 2 treatment, recovery buffer, growth phase of cells, DNA volume, and flanking region length) on homologous recombination efficiency and optimized the transformation protocol for PCR tailing. If a combination system consisting of one-step PCR and pop-out excision is developed, alternative multiple gene knockout systems become accessible. To this end, pyrE-lacS dual marker genes were utilized for positive, negative, and blue selection. This effective approach (multiple gene knockout system with one-step PCR) was validated by unmarked gene knockout of the DNA photolyaseand arginine decarboxylase-encoding genes (phr and argD, resp.) in the SuaI-deficient S. acidocaldarius strain SK-1 (ΔpyrE ΔsuaI).

Strains and Growth
Conditions. The strains used in this study are listed in Table 1. The S. acidocaldarius pyrimidine-auxotrophic and restriction endonuclease SuaIdeficient strain SK-1 (ΔpyrE ΔsuaI) was used as basic host strain [8]. This strain and its derivatives were cultivated in xyrose and tryptone (XT) medium (pH 3) [9] containing 1× basal salts (3 g K 2 SO 4 , 2 g NaH 2 PO 4  To solidify plates, identical components of 1× basal salts containing 2.9 g MgSO 4 ·7 H 2 O and 0.5 g CaCl 2 ·2 H 2 O were used. For growth of the uracilauxotrophic strain, 0.02 g/L uracil was added to XT medium (XTU). XTU medium supplemented with 50 μg/mL 5-FOA was used for counterselection with the pop-out recombination method. For cultivation of the argD mutant, 1 mg/mL agmatine (agmatine sulfate [Tokyo Chemical Industry]) was added to the XTU medium. Escherichia coli strain DH5α, used for general manipulation, was routinely cultivated at 37°C in Luria-Bertani medium supplemented with ampicillin (100 μg/mL).

2.2.
General DNA Manipulation. The reagents used in these experiments were prepared as previously described [8]. PCR products and plasmid DNA were purified using NucleoSpin Gel and PCR Clean-up and NucleoSpin Quick-Pure kits (Macherey-Nagel), respectively.  of the suaI (Saci_1976) locus at both ends of the pyrE-lacS marker. The lacS gene, together with its putative promoter and terminator regions, was amplified from the S. solfataricus P2 genomic DNA using primers SSOlacS-F/R (containing PstI/BamHI restriction sites). The PCR product was digested with PstI/BamHI, then purified and inserted into pSuaIPOP [8] at the corresponding restriction sites. Linear DNA of the pyrE-lacS dual marker cassette containing various lengths (800, 50, 40, 30, 20, and 10 bp) of the 5 ′ and 3 ′ homologous arms was amplified from placSpyrE as a template using the corresponding primers (E800-20-F/R and E10-2F/2R) and Emerald Amp MAX PCR Master mix (Takara Bio). The PCR products were purified in 5 mM Tris-HCl (pH 8.5) and transformed into SK-1 to estimate the homologous recombination efficiency via double crossover ( Figure 1).

Construction of phr and argD Knockout PCR Products.
A MID strategy [3] and PCR-tailing technique [1] were combined to develop our multiple gene knockout system with one-step PCR (MONSTER). The MONSTER was utilized for phr (Saci_1227) and argD (Saci_1363) knockout cassette construction. In brief, the phr knockout PCR product (MONSTER-phr) was amplified from placSpyrE as a template using primers phr-pop-F/R (containing the 48 bp and 30 bp 5 ′ and 3 ′ flanking regions of phr and a 48 bp region of phr as the Tg-arm) and Premix Taq (Ex Taq Version 2.0; Takara Bio) under the following conditions: 94°C for 3 min; 30 cycles of 94°C for 30 s, 56°C for 30 s, and 72°C for 3 min, and a final extension for 3 min. Similarly, the argD knockout PCR product (MONSTER-argD) was amplified from plac-SpyrE as a template using primers argD-pop-F/R (containing a 48 bp region of argD as the Tg-arm and the 30 bp and 48 bp 5 ′ and 3 ′ flanking regions of argD) and the LA-Taq DNA polymerase (Takara Bio) under same PCR conditions. The purified PCR products were used in subsequent experiments.

Construction of an argD-Based Shuttle
Vector. The S. solfataricus argD gene with approximately 100 bp of the 5 ′ and 3′ flanking regions was amplified by PCR using the primers SsoargD-KpnI-F/PstI-R, which contain the KpnI and PstI restriction sites, respectively, and Premix Taq (Ex Taq Version 2.0; Takara Bio). The SsopyrEF marker genes in pSAV2 [8] were replaced by the SsoargD marker gene at the KpnI and PstI sites, thus generating the argD-based shuttle vector pSAV2-argD.

Transformation
Procedure. Preparation of electrocompetent cells and transformation were completed as previously Restriction sites are underlined and sequences of MONSTER primers that anneal with the pyrE-lacS marker genes are in bold.
described [8] with the following modifications. Cells were incubated in a 1 L DURAN bottle (Schott) containing 200 mL of medium with shaking using a Bio shaker (TAITEC). S. acidocaldarius (strain SK-1 [ΔpyrE ΔsuaI]) electrocompetent cells for transformation with a shuttle vector and via homologous recombination were prepared from a late log to stationary phase culture (OD 600 ≧ 0.7) and an early to midlog phase culture (OD 600 = 0.1-0.4) incubated in XTU medium, respectively. Cells were harvested by centrifugation (10160 ×g for 15 min at 25°C) using a Kubota 6500 and were washed once in 0.3 volumes of the original culture volume of 20 mM sucrose at room temperature. The final optical density at 600 nm (OD 600 ) of cells was adjusted to 5.9 (2 × 10 9 cells/mL) by calculation, and aliquots were frozen at −84°C in an ultralow freezer (Sanyo). All transformation procedures (including preparation of competent cells) were carried out at room temperature. Two hundred microliters of competent cells (4 × 10 8 cells) were thawed by hand and mixed with 1-10 μL of DNA in 5 mM Tris-HCl (pH 8.5  [11]. Cells were then incubated at 77°C-78°C for 30 min without shaking in a hot block (TOHO). After the regenerated samples were centrifuged (11000 ×g for 1 min at 25°C), 800 μL of supernatant was removed and the pellet was suspended in 200 μL followed by spreading on plates. Two plating methods, that is, direct plating and overlay cultivation, were performed. For direct plating, the transformed cells were immediately spread onto XT plates and incubated at 75°C for 6-7 days in sealed plastic cases. For overlay cultivation, transformed cells (~1 mL) were mixed with 10 mL of prewarmed top gel solution (5 mL of XT medium, 5 mL of 0.4% gellan gum, 50 μL of 0.5 M CaCl 2 , and 50 μL of 2 M MgSO 4 ) at 75°C, then poured onto XT plates and cultivated at 75°C for 6-7 days in sealed plastic cases.
2.5. X-Gal Assay. β-Glycosidase activity encoded by the lacS gene was detected in transformant colonies by spraying a 10 mg/mL X-gal (Wako or Carbosynth) solution on the plates and incubating at 75°C for 1 day. Transformants (lacS + ) convert the chemical into a strong blue substance, whereas nontransformants (wild-type S. acidocaldarius) do not [12].
2.6. Estimation of Transformation Efficiency. When pyrE or argD selectable marker was used for positive selection, colonies appearing on the plate were scored except for tiny colonies that might have been background.

Characterization of Mutant Strains.
To characterize the phenotypes of the DNA photolyase-deficient strain DP-1 (ΔpyrE ΔsuaI Δphr) and argD deletion mutant SK-5 (ΔpyrE ΔsuaI ΔargD), UV sensitivity and agmatine auxotrophy were examined, respectively. To assess photoreactivation [1] in the strain DP-1, the growth properties under light and dark conditions after UV irradiation were examined. One milliliter of each overnight culture (late log to stationary phase) was poured in 90 × 15 mm plastic petri dishes (IWAKI) and irradiated with a UV lamp (UVM-57) (304 nm, 6 W) (Tech-jam) positioned 6.5 cm from the top of the dish at room temperature for 60 s (1200 J/m 2 ). UV-irradiated cultures were immediately inoculated into 6 mL of XTU liquid medium to yield an initial  OD 600 of 0.005. Cells were then cultivated with shaking. For mock-treated control cultures, the same procedure was followed without UV irradiation. For dark conditions, test tubes and Bio shakers (TAITEC) were covered in foil. For light conditions, cells were cultivated under a white LED using ODS-LS16-W (Ohm Electric). Cell growth was monitored thereafter.
To compare the growth properties of strain SK-5 in the presence or absence of agmatine, overnight cultures (late log to stationary phase) were inoculated into 6 mL of XTU liquid medium supplemented with 100-1000 μg/mL agmatine to yield an initial OD 600 of 0.005. Cells were cultivated with shaking and cell growth was monitored thereafter.
Transformant genotypes were analyzed by sequencing the target region following PCR amplification using primer sets that anneal outside the flanking target gene locus.

Effects of Transformation Conditions on Homologous
Recombination Efficiency. The PCR-tailing technique for gene disruption was developed in the thermoacidophilic crenarchaeon S. acidocaldarius [1]; however, transformation efficiency has not been reported. Homologous region length can significantly impact transformation efficiency [13]. The efficiency of homologous recombination via double crossover using very short homologous arms (50~10 bp) is likely very low. The widely used transformation procedure for Sulfolobus has been reported; however, the effects of transformation conditions on transformation efficiency have not been characterized in detail as compared to those of other model systems [14][15][16]. To develop a reliable multiple gene knockout system using PCR tailing, we first optimized the following transformation conditions: plating methods, DNA topology, CaCl 2 treatment, recovery buffer, growth phase of cells, DNA volume, and length of flanking regions ( Figure 1).
We examined the effects of two plating methods on transformation efficiency (Figure 1, vii). We used 200 ng of linear DNA from pyrElacS800 (800 bp homologous arms Table 1), and competent cells were harvested at midlog phase culture (OD 600 = 0.391). After electroporation (12.5 kV/cm, 20 ms), the samples were immediately plated using two plating methods: direct plating or overlay cultivation. The transformation efficiency for direct plating was 7.5 × 10 2 ± 2.2 × 10 2 colonies/μg DNA, while that of overlay cultivation was 2.7 × 10 2 ± 4.0 × 10 colonies/μg DNA. Thus, the transformation efficiency for direct plating was 2.7-fold higher than that of overlay cultivation. The experiments were repeated in triplicate.
To analyze the effect of DNA topology on homologous recombination via double-crossover events, circular and linear marker cassettes containing 800 bp homologous regions were tested (Figure 1, i). The experimental conditions were identical to those described in the previous paragraph, except that 300 ng of circular DNA pyrElacS800 (placSpyrE) and another previously reported electric parameter (15 kV/cm, 9 ms) [17] were utilized. When DNA was electroporated, the transformation efficiency using linear DNA was 24-fold higher than that of the circular DNA: 1.5 × 10 3 ± 4.2 × 10 2 colonies/μg DNA and 6.2 × 10 ± 7.0 colonies/μg DNA, respectively. The experiments were repeated in triplicate.
Electroporation in the presence of Ca 2+ enhanced the transformation efficiency of E. coli [14]; however, this effect has not been reported in the hyperthermophilic genus Sulfolobus. For validation of the effect of CaCl 2 treatment on homologous recombination efficiency in S. acidocaldarius, electrotransformation was performed in the presence and absence of CaCl 2 (Figure 1, ii). Competent cells were collected at midlog phase (OD 600 = 0.420). Concentrations of 0.1, 0.2, and 0.4 mM CaCl 2 were selected because these concentrations did not cause arching during electroporation. However, CaCl 2 treatment did not improve transformation efficiency when compared with control experiments (data not shown). The experiments were repeated in triplicate. We speculated that DNA volume is important for improving transformation efficiency with CaCl 2 treatment (Figure 1, iv). However, DNA volume (1000 ng) did not improve the transformation efficiency with CaCl 2 treatment (data not shown).
Subsequently, to study the transformation efficiency using linear DNA for homologous recombination in S. acidocaldarius with double-crossover events, marker cassettes containing 50-10 bp 5 ′ and 3 ′ homologous regions of the target locus at both ends of the pyrE-lacS marker were constructed (Figure 1, v). Competent cells harvested at midlog phase (OD 600 = 0.420) were transformed with 1 μg of marker cassettes. The transformation efficiency increased with the length of the homologous arms ( Figure 3). When DNA with 10-20 bp of flanking regions was used, no transformants grew. Transformation efficiencies slightly improved by regeneration with MBS buffer. A few colonies transformed with DNA attached to 20 bp flanking arms were detected after regeneration. Thus, efficient marker replacement was possible with as few as 30-50 bp of flanking homology of the target region.
The following set of conditions was established as the optimized transformation protocol: DNA was introduced into competent cells collected from the early log phase by electroporation. The pulse duration was 9 ms and the field strength was 15 kV/cm. After electroporation, cells were regenerated in MBS recovery buffer and the pellet was spread on plates.
3.2. Establishment of the MONSTER. The multiple gene knockout system with one-step PCR (MONSTER) was developed by combining a MID strategy [3] and one-step inactivation using a linear PCR product [1] (Figures 4 and 5). Two 48 bp homologous arms were used for double-crossover events (marker integration), followed by pop-out recombination at 30 bp duplicated arms for the excision of a marker cassette (unmarked gene deletion). Thus, two MONSTER primers need to be designed for incorporation of 5′, 3′, and Tg (target gene) arms into PCR products as 5′ extensions of primers ( Table 2). Sequences of forward and reverse MON-STER primers that anneal with pyrE-lacS marker genes are identical although the attached flanking regions of target genes (5′, 3′, and Tg) are different. Next, the MONSTER cassette was amplified by one-step PCR using MONSTER primers. Then, we designed different constructs of MON-STER cassettes (MONSTER-phr and MONSTER-argD) for confirming the reliability (Figures 4(a) and 5(a)). The dual marker (pyrE-lacS) was utilized for effective selection of correct transformants (Figures 4 and 5).

Construction of a DNA Photolyase-Deficient Strain via the MONSTER.
To validate the MONSTER, we constructed a mutant with an in-frame deletion of DNA photolyase. DNA photolyase-encoding gene (Saci_1227) (named phr in this study) has been identified as a functional gene of photoreactivation (repair of UV-damaged DNA under light conditions) [1,18]. To disrupt phr, MONSTER-phr was constructed by one-step PCR (Figure 4(a)). When 1.6 μg of MONSTER-phr was electroporated into SK-1 using the optimized transformation protocol with competent cells harvested at midlog phase (OD 600 = 0.420), approximately 60 colonies grew on XT plates (Figure 4(b)). Next, three blue colonies were selected after applying 1 μL of X-gal solution (10 mg/mL) onto the plates for 1 h at 75°C (Figure 4(b)). Two blue colonies were purified by single isolation and analyzed by PCR screening using primers phr-out-F/R (Figure 4(a)). As shown in Figure 4(d), the two colonies were positive intermediate transformants (named DP-1 Int-1 and Int-2). A total of 2.3 × 10 8 DP-1 Int cells were spread on XTU plates containing 5-FOA for pop-out recombination. X-gal visualization of the plates indicated that blue and white colonies formed with a ratio of 100 : 13 (Figure 4(c), 65 ± 35 white colonies grew). Ten 5-FOA r white colonies were randomly selected for PCR analysis. The genotypes of 9 out of ten colonies were expected with an approximate 1.3 kb deletion in the phr locus (Figure 4(d)). One correct Δphr in-frame mutant confirmed by sequencing was designated as S. acidocaldarius strain DP-1 and used for phenotypic analysis.

Characterization of the DNA Photolyase-Deficient
Strain DP-1. To characterize the DNA photolyase-deficient strain, the growth properties of wild-type (SK-1) and Δphr (DP-1) under light or dark conditions after UV irradiation were investigated ( Figure 6). When both strains were not irradiated with UV light, their growth properties were identical under light and dark conditions. In addition, the growth of UV-treated DP-1 under dark conditions was similar to that of the host strain. In contrast, the UV-exposed DNA  photolyase-deficient strain DP-1 grew slower when compared with the SK-1 strain under light conditions, indicating that deletion of the phr locus eliminated photoreactivation.

Construction of the argD-Deficient S. acidocaldarius
Strain SK-5 via the MONSTER. We disrupted the argD gene using the MONSTER to establish a robust unmarked gene disruption system, and a positive selectable marker in S. acidocaldarius ( Figure 5). argD (Saci_1363) encodes arginine decarboxylase, which catalyzes L -arginine to produce agmatine [19], and is a homolog to SSO0536 in S. solfataricus P2 and SisM164_1585 in S. islandicus M.16.4, sharing 73% and 74% identity by Blastp analysis, respectively. For construction of the argD in-frame deletion mutant (Figure 5(a)), 2 μg of one-step constructed MONSTER-argD was introduced into SK-1 cells harvested at the late-log phase (OD 600 = 0.558; electroporation conditions: 12.5 kV/cm and 20 ms) and then cultivated on XT plates containing 200 μg/ mL agmatine at 75°C for 6 days. As shown in Figure 5(b), five colonies grew. X-gal selection revealed three blue colonies. Two of these blue colonies were purified on XT plates and analyzed by PCR screening using primers argD-F-F/R ( Figure 5(a)). As shown in Figure 5(d), both clones contained 2.5 kb (pyrE-lacS marker and 30 bp 5′ regions) inserted bands, indicating that two blue colonies were positive intermediate transformants (named SK-5 Int-1 and Int-2). These transformants grew in XT liquid culture, suggesting that insertion of the marker between the stop codon and the 3′ region of the argD locus did not affect arginine decarboxylase activity (data not shown). A total of 3.4 × 10 8 SK-5 Int cells were spread on XTU plates containing 5-FOA and 1 mg/mL agmatine for pop-out recombination. X-gal visualization demonstrated that blue and white colonies formed with a ratio of 167 : 16 ( Figure 5(c), 16 ± 6 white colonies grew). Twelve 5-FOA r white colonies were randomly selected for PCR analysis using outer primers. The genotypes of 10 out of twelve colonies showed the expected approximately 0.4 kb deletion in the argD locus ( Figure 5(d)). One correct ΔargD in-frame deletion mutant confirmed by sequencing, designated S. acidocaldarius strain SK-5, was characterized for phenotypic analysis.
3.6. Characterization of the argD Deletion Mutant SK-5. The growth of the argD-deficient strain SK-5 (ΔpyrE ΔsuaI ΔargD) was studied using XTU liquid culture in the presence or absence of agmatine (Figure 7). When SK-5 was cultivated in the presence of 1 mg/mL agmatine, growth was slightly retarded when compared with that of the host strain in the absence of agmatine. Particularly, the slowed growth of SK-5 became more striking at lower concentrations of agmatine. In contrast, SK-5 was not grown with less than 100 μg/mL agmatine.

Construction of a Stringent-Positive Selection Marker
System Based on Agmatine Selection in S. acidocaldarius. The agmatine selection system has been reported as a stringent-positive selection marker system in the hyperthermophilic archaea Pyrococcus furiosus and S. islandicus [10,20]; however, this system has not been developed in S. acidocaldarius. To establish a selection marker system based on complementation of the argD gene, a S. acidocaldarius-E. coli shuttle vector pSAV2-argD was constructed by replacing the S. solfataricus pyrEF marker genes of pSAV2 with the S. solfataricus argD gene (SSO0536) (Figure 8(a)).

Discussion
The goal of the present study was to establish a multiple gene knockout system with PCR tailing in the thermoacidophilic crenarchaeon S. acidocaldarius. For this, we first optimized the transformation protocol by characterizing the effects of transformation conditions on transformation efficiency. Next, we successfully developed a multiple gene knockout system with one-step PCR (MONSTER) by combining marker recycling with PCR tailing. This technique allows for the simple one-step construction of an unmarked gene knockout cassette and isolation of targeted gene deletion mutants. Unmarked gene deletion methodologies have been troublesome for genetic studies of other recombinogenic hyperthermophilic archaea. Although the development of PCR-tailing methods is possible for hyperthermophilic archaea, the potential for multiple gene knockout systems is limited due to the limited selectable marker systems. Thus, the MONSTER may be a speedy and powerful genetic tool for other recombinogenic hyperthermophilic archaea. In addition, we also constructed a stringent selectable marker system using agmatine, which provides the basis for further genetic manipulation in S. acidocaldarius.
Our results indicated that the main factors affecting transformation (homologous recombination) efficiency via double-crossover events were DNA topology, recovery conditions after electroporation, and flanking region length. In addition, the plating methods and the growth phase of competent cells were also important for optimizing transformation. In contrast, CaCl 2 treatment and DNA volume did not affect transformation efficiency in this study.
The effects of DNA form on homologous recombination were reported in Sulfolobus species [2,13]. Our results support a previous report that the transformation efficiency using linear DNA was higher than that of circular DNA [2,13].
To develop a gene manipulation system based on PCR tailing, we focused on the possibility of sufficient homologous recombination via double-crossover events with very short homologous regions. The effects of flanking region length on homologous recombination efficiency in S. acidocaldarius were previously reported by Kurosawa and Grogan [13], and our data support their findings (Figure 3). The PCR-tailing technique was also previously established [1]. In contrast, our study is the first to report that sufficient transformation  efficiency for gene manipulation was demonstrated even with very short (30-50 bp) flanking homologous arms. When 40 bp homologous arms were attached, the transformation efficiency using our protocol (20 ± 7 colonies/μg) was slightly higher than that of the recombinogenic P. furiosus strain COM1 (6 colonies/μg) reported by Farkas et al. [21] (Figure 3). To our knowledge, no similar observation has been reported in the literature.
Effective multiple gene knockout techniques have been developed in Sulfolobus [2][3][4]. However, the cloning steps of two to four fragments for construction of knockout vectors are required for these genetic tools. In addition, the screening of positive clones that contain the correct construct must be randomly selected during subcloning because X-gal selection cannot be utilized. In contrast, PCR tailing is a highthroughput gene knockout technique [27]. However, the possibility of using this method for multiple gene knockout is limited in S. acidocaldarius because marker genes are lacking [1]. We developed the MONSTER by combining the multiple gene knockout technique with PCR tailing in S. acidocaldarius. The main advantage of the MONSTER compared with published unmarked gene deletion methodologies [2][3][4] is the very simple construction of multiple gene knockout cassettes without any plasmid construction. The usefulness of this technique was proven by unmarked gene knockout of the phr and argD genes. Another advantage of the MON-STER is that multiple unmarked gene knockout cassettes can also be simultaneously amplified under the same PCR conditions because the sequences of MONSTER primers that anneal with the dual (pyrE-lacS) marker genes are identical, even though the attached flanking regions of the target genes are different. Therefore, MONSTER is a high-throughput method compared with the widely used methods in Sulfolobus [2][3][4]. Notably, the purification of intermediate transformants (Int strain) was very important for pop-out selection (Figures 4(c) and 5(c)). Thus, this study provides an alternative and versatile strategy for the genetic manipulation of S. acidocaldarius with several advantages.
To establish the MONSTER in other hyperthermophilic archaea, dual marker genes are required for counterselection and screening (Figures 4(c) and 5(c)). In addition, recombinogenic host strains that allow for homologous recombination using very short flanking homologous regions is likely required.
A uracil-based selection system (pyrE-, pyrF-, or pyrEF-deficient strains and marker genes) cannot efficiently estimate transformation efficiency in hyperthermophilic archaea due to the interference caused by background growth of the pyrEF-deficient strain on solid medium [10]. In contrast, an agmatine selection system is a powerful genetic marker due to the lack of background colony growth on plates (Figure 8(b)) [10]. Therefore, the genetic marker system developed in this study will allow versatile genetic manipulation in S. acidocaldarius. Notably, a higher concentration of agmatine was required for cultivation of the S. acidocaldarius argD-deficient strain when compared with other hyperthermophiles [10,20,23,28].
We previously reported that no background colonies appeared in the host-vector system (especially the SK-1 strain) using uracil selection for a 7-day cultivation [8]. This advantage was confirmed with our stringent positive marker system based on agmatine selection (Figure 8(b)).
Additionally, we constructed the DNA photolyasedeficient strain DP-1 as a genetic host strain that does not require dark conditions for the functional genetic analysis of candidate genes involved in the UV response [29][30][31].

Conclusion
We combined marker recycling (pop-out recombination) with PCR tailing to develop a multiple gene knockout system with one-step PCR. In addition to the widely used multiple gene knockout techniques in S. acidocaldarius, this study describes an alternative strategy that enables the very simple construction of multiple gene knockout cassettes. Indeed, we believe our techniques will contribute to the genetic study of this archaeon.

Conflicts of Interest
The authors declare that they have no conflicts of interest.