Strains and growth media

S. acidocaldarius E4-39 is a spontaneous uracil auxotroph (Figure S1) that was isolated from S. acidocaldarius 639 from DSMZ (Braunschweig, Germany). The strains were cultivated aerobically on ATCC #1723 (YT) medium. The organism was generally grown in liquid medium at 75°C on a rotary water-bath shaker at 200 rpm and at 70°C on solid medium that was composed of YT or T (YT without yeast extract) medium solidified with 0.8% gellan gum (Spectrum and Sigma). The pH of each growth medium was adjusted to 2.5 with 5 M H₂SO₄, and each medium was filter sterilized using 0.2-µm pore size membrane filters. Chemicals used in this study were obtained from Sigma, Spectrum, and Fisher scientific.

Construction of shuttle vectors

Seven shuttle vectors (pRSP1 – pRSP7) harboring four lengths of pRN1, pUC19 and a cassette (pyrE-lacS) containing pyrE and lacS genes from S. solfataricus were assembled by a PCR-based cloning method to verify the minimal sequence of pRN1 that is critical for plasmid replication (Figure 1). Three of the shuttle vectors pRSP2, pRSP4 and pRSP6 were constructed by eliminating lacS from pRSP1, pRSP3 and pRSP5, respectively, by PCR. Each vector contained a 232-bp sequence located 3’ of orf904 that we identified as the putative origin of the plasmid by inspection of un-annotated pRN1 sequence based on A-T abundance. The first set of plasmids (pRSP1, pRSP3, pRSP5 and pRSP7) was constructed by assembling three PCR-amplified DNA fragments: the 1750-bp pUC19 fragment, 2239-bp pyrE-lacS cassette, and a pRN1 fragment (Figure 1 and Table 1). The pUC19 fragment (nt. 800 - 2550) was amplified with primers P/N128 and P/N129 and consists of bla (Amp^R) and the pMB1 origin of replication (Table S1). The pyrE-lacS cassette was amplified from pGlcSTUV_ko (unpublished plasmid, Figure S2) with primers P/N117 and P/N120 and includes the contiguous 80-bp 3’ of the lacS stop codon, which is assumed to contain the native terminator sequence (Figure 1). Translation of lacS was placed under the control of putative RBS of pyrF located at the 3’-end of pyrE and was kept in frame by the addition of an A-T base pair between the stop codon of pyrE and the start of lacS. The pyrE-lacS operon was transcribed by the promoter of the beta subunit (thsB) of S. solfataricus chaperonin (Sso0282). . The four pRN1 fragments were amplified from shuttle vector pC [7] using a forward primer (P/N130) and four reverse primers (P/N131 – 134) described in Table S1.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Figure 1. Construction of pRN1-based shuttle vectors.

(A) Map of pRN1 showing native genes. (B) Map of pRN1-based shuttle vector(s) showing the 232-bp putative pRN1 origin of replication (red double arrow). Each vector harbors the pUC19 Ap^r gene and origin (red box), pyrE with or without lacS and a segment of pRN1 highlighted by black arrows. The lacS gene was deleted from pRSP1, 3, and 5 to generate pRSP2, 4, and 6.

https://doi.org/10.1371/journal.pone.0084664.g001

We assembled pRSP1, pRSP3, pRSP5, and pRSP7 using a modified SLIC cloning strategy [36,37] by combining the pUC19 and pyrE-lacS fragments with one of the four pRN1 fragments in a 30-µl chew-back and annealing (CBA) reaction with all fragments at approximately the same molarity using 50 - 100 ng of the pUC19 fragment as reference. The CBA reaction contained 1X T4 ligase buffer (Fermentas), 0.1 mg/mL BSA, 0.75 µL T4 polymerase (Fermentas) and water. The reaction was carried out in a thermocycler at 37°C for 7 min, then at 75°C for 20 min, and cooled to 60°C at 0.1°C /s. The reaction was held at 60°C for 30 min, and then cooled to 4°C at 0.1°C /s [37]. A fill-in/ligation (PL) reaction was carried out by adding 0.2 µl T4 polymerase, 0.2 µl 10 mM dNTP’s, 0.2 µl 50 mM ATP, and 0.5 µl T4 ligase to 10 µl of the CBA reaction, incubating at 37°C for 30 min, and then inactivating the enzyme at 75°C for 5 min. Samples from each cloning reaction (CBA and PL) were transformed into E. coli DH10B and TOP10 and inoculated onto LB plates containing 100 µg/ml carbenicillin. Constructed vectors were extracted and verified by sequencing the entire plasmid. We subsequently constructed pRSP2, pRSP4 and pRSP6 by removing lacS from pRSP1, pRSP3, and pRSP5, respectively, using P/N135 and P/N136 as forward and reverse primers, respectively (Table S1). A 30-bp adaptor that is complementary to the 3’-end of pyrE was added to the 5’-end of primer P/N135 to ensure amplification of circular plasmids. Details of the primers used in amplifying these fragments can be found in Table S1.

Mapping the putative origin of replication of pRN1

The 232-bp putative origin of replication was functionally mapped by partial or complete deletion of this region from pRSP1 or pRSP3 (Figure 2 and Table 1). The resulting plasmids were methylated, transformed into S. acidocaldarius E4-39, and scored for their ability to generate transformants. First, we divided the 232-bp region into four segments of 52- or 60-bp and deleted each segment from pRSP1 to generate four plasmids (pRSP1a – d) (Table 1 and Figure 2A). We also generated a fifth plasmid (pRSP1e) by deleting a 12-bp sequence (4844 - 4855) that includes a GTG tri-nucleotide motif from pRSP1 (Table 1) to determine its role in plasmid replication. Next, we mapped the 58-bp loop (4755 - 4812) of a putative stem-loop (SL) secondary structure (4734 - 4833) within the 232-bp putative origin in pRN1 to determine if it is critical for plasmid replication. The loop (L) was mapped by deleting 18- or 58-bp segments from this loop in pRSP1, generating four plasmids (pRSP1-L1 – pRSP1-L4) (Table 1 and Figure 2A). We generated a third set of plasmids (pRSP1-G1 – 8) making 3-bp deletions from pRSP1 (Figure 2) to further analyze the 24-bp stretch of DNA (4773 - 4796) within the 58-bp loop that contained a putative GTG motif reported to be important for primase activity of Orf904 [35]. The entire 232-bp putative origin (4681 - 4912) was deleted from pRSP1 to generate pRSP1-NO to test for loss of function; a similar construct (pRSP3-NO) was generated by deleting 207 of the 232-bp (4707 - 4912) from pRSP3. With the exception of pRSP1b and pRSP1d, each plasmid described above was amplified directly from pRSP1 as a linear, blunt-end, PCR fragment lacking the targeted region. Approximately 0.2 - 0.3 pmol of the 5’-terminus of each linear PCR amplicon was phosphorylated at 37°C for 20 min with T4 polymerase kinase (PNK) and ligated overnight at ~22°C using T4 ligase as described by the manufacturer (Fermentas). The plasmids pRSP1b and pRSP1d were amplified as circular plasmids using primers with adaptors with overlapping sequence (Table S1). Each construct was transformed into either E. coli DH10B or TOP10, and deletions in the generated vectors were confirmed by sequencing the surrounding regions. These were subsequently methylated and transformed into S. acidocaldarius and screened for blue colonies expressing LacS. The primers used in the construction of these plasmids are listed in Table S1.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Figure 2. Strategy of mapping pRN1 putative origin of replication (double red arrow in Figure 1).

(A) Various deletions were made within the ‘putative ori’ in pRSP1 to generate plasmids highlighted in each rectangular box to determine the role of the deleted regions in pRN1 replication. We constructed pRSP1-NO and pRSP3-NO by deleting 232-bp and 207-bp, respectively, from the putative origin. (B) Three-nucleotide deletion mutants of pRSP1 were constructed to map the loop of stem-loop structure within pRN1 origin. Each construct was transformed into S. acidocaldarius and scored for production of LacS expressing transformants.

https://doi.org/10.1371/journal.pone.0084664.g002

Methylation and transformation of S. acidocaldarius

Each assembled shuttle vector was transformed into E. coli 1821.pM.EsaBC4I (New England Biolabs) for methylation to protect it against SuaI, a restriction endo-nuclease present in S. acidocaldarius [38]. Approximately 1 - 2 µl of each methylated construct was transformed into 80 µl of electro-competent, uracil-auxotrophic S. acidocaldarius E4-39 cell suspension (OD₆₀₀ = 5 - 10) that was prepared following the “G” procedure described by Kurosawa and Grogan [39]. Prior to the addition of plasmid DNA, the competent cell suspension was diluted with approximately 5 parts of ice-cold 1% sucrose solution. The mixture was transferred into a 1-mm cuvette; electroporated at 1.25 KV, 50 µF, and 750 Ω; and immediately recovered with 80 µl of ice-cold 2X recovery buffer (1% sucrose, 20 mM β-alanine/1.5 mM malate buffer [pH = 4.5] and 10 mM MgSO₄). A negative control reaction (without plasmid DNA) was added to every transformation experiment. The recovered cells were incubated at 70°C for 30 min, cooled on ice for 2 - 5 min, and spread on YT or T plates. The plates were incubated at 70°C in humidified Sterilite™ plastic containers for 5 - 7 days. After incubation, the plates were sprayed with 5 mg/ml X-gal (5-romo-4-chloro-3-indolyl-b-D-galactoside) in dimethyl formamide (DMF) to qualitatively screen for LacS activity. Blue colonies were scored as positive selected for further characterization. Small to medium-sized (white) colonies were selected as positive transformants from the transformation reaction with plasmid lacking lacS (pRSP2, pRSP4 and pRSP6) upon comparing these plates to the negative control plate. Most transformation experiments were repeated at least two times to verify the obtained result. To recover plasmid DNA, each selected transformant was grown in 50 ml of YT medium (OD₆₀₀ ≈ 1.0). The cells were pelleted and re-suspended in 500 μl of P1 buffer supplemented with 15 mg/ml lysozyme to enhance lysis. The re-suspended cells were incubated at 37°C for 10 min, and then treated with 500 μl of buffer P2 and 700 μl of buffer N3 following the protocol described by the manufacturer. The recovered plasmid was re-transformed into E. coli. Plasmids recovered from S. acidocaldarius E4-39 and re-transformed into E. coli were treated with EcoRI to determine the restriction digestion fingerprint.

Plasmid complementation

We constructed pRSP9 and pRSP10 by making in-frame deletions of orf904 and orf56, respectively, from pRSP1 (Table 1). A third plasmid, pRSP10b, was constructed by retaining 75-bp sequence upstream of the orf904 start codon. The plasmids were constructed by PCR amplification of the entire pRSP1 without the gene segment targeted for deletion using primers described in Table S1 and the protocol described above. Each construct was methylated and transformed into S. acidocaldarius E4-39; pRSP9 was paired with pRSP10 or pRSP10b and transformed into the host to test for gene complementation.

Bioinformatics analyses

The plasmids used in this study were designed and analyzed using Invitrogen’s Vector NTI software. DNA sequences were aligned and analyzed using BLAST and ClustalW2 informatics tools which can be accessed via http://www.ncbi.nlm.nih.gov and http://www.ebi.ac.uk/Tools/clustalw2/index.html.

Nucleotide sequence access

The nucleotide sequences of the plasmids used in this study have been deposited in the Joint BioEnergy Institute (JBEI) public part registry and can be accessed via this link: https://public-registry.jbei.org/#page=collections;id=102.

The 232-bp putative ‘origin’ of pRN1 is critical for plasmid function

The origin of replication is critical for the function and replication of a plasmid in a suitable host. We evaluated the impact of a 232-bp DNA sequence on the ability of pRN1 to replicate and yield transformants. Sulfolobus species have been successfully transformed with pRN1-based plasmids but the origin of replication of this plasmid has not been fully characterized. The 232-bp stretch of DNA (5031 - 5262) was identified by visual inspection of the pRN1 sequence and was found to be located at the 3’ of orf904 in agreement with previous report [25]. We evaluated the function of the 232-bp stretch of DNA by constructing shuttle vectors (pRSP1 – pRSP7) bearing various lengths of pRN1 (Figure 1B and Table 1) and transforming the plasmids into S. acidocaldarius E4-39, a uracil auxotroph. All shuttle vectors yielded uracil prototrophic colonies of S. acidocaldarius E4-39, suggesting that each harbored the origin of replication (Table 2 and Figure S3). Colonies from pRSP1, 3, 5, and 7 transformations stained blue on X-gal plates due to β-glycosidase activity encoded by the S. solfataricus lacS present on these plasmids, whereas colonies from pRSP2, 4, and 6 remained white with X-gal exposure due to the absence of lacS (Figure S3). The transformations were repeated to confirm the results. The sequence of each plasmid was confirmed by sequencing the entire plasmid prior to transformation into S. acidocaldarius E4-39. We also verified that each plasmid was self-replicating and not integrated into the host genome by recovering the plasmids from each transformants and re-transforming the recovered plasmid into E. coli. The EcoRI digestion fingerprint of plasmids recovered from S. acidocaldarius E4-39 was identical to those recovered from re-transformed E. coli indicating that the plasmids were self-replicating (Figure 3).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Figure 3. Extraction plasmids from S. acidocaldarius transformants and re-transformation into E. coli.

(A) pRN1-based shuttle vectors were recovered from S. acidocaldarius transformants and re-transformed into E. coli to verify that the plasmids were self-replicating. The identity of the plasmids was confirmed by their EcoRI digestion finger-print. (A) EcoRI digestion finger-print of pRSP1, 5 and 7 recovered from (A) S. acidocaldarius E4-39 transformants and (B) E. coli re-transformed with plasmid DNA from the S. acidocaldarius transformants analyzed in (A).

https://doi.org/10.1371/journal.pone.0084664.g003

The transformation results indicate that pRSP1 and pRSP2, which harbored orf56, orf904, and the 232-bp sequence (Table 2 and Figure 1B), constitute the minimum replication unit of pRN1, in agreement with a previous report [25]. Replication of pRSP1 and pRSP2 suggests that the 232-bp sequence harbors the putative origin of replication. To further confirm the function of the 232-bp sequence, we deleted the feature from pRSP1 and pRSP3. The resulting plasmids (pRSP1-NO and pRSP3-NO) failed to yield transformants (Table 3). The experiment was repeated at least two times, and on every occasion transformants were recovered with pRSP1 and pRSP3, while pRSP1-NO and pRSP3-NO failed to yield transformants. The critical role of the 232-bp putative origin of replication in the propagation of pRN1 was further highlighted by the inability of deletion mutants (pRSP1-NO and pRSP3-NO) lacking this region to yield transformants, while the parent plasmids, pRSP1 and pRSP3, respectively, always yielded transformants (Table 3). The result was highly reproducible and was confirmed on at least three separate occasions. The absence of transformants from pRSP1-NO and pRSP3-NO suggests that orf56 and orf904 or the additional pRN1 sequence in pRSP3 (4913 - 5314) are insufficient for pRN1 replication, thus highlighting the importance of the 232-bp sequence.

The 232-bp region harbors a 100-bp putative replication initiation stem-loop

We mapped the 232-bp sequence by deleting the 52- or 60-bp segments from pRSP1 generating pRSP1a – pRSP1d (Table 2 and Figure 2). Two of the four plasmids (pRSP1b and pRSP1c) did not yield transformants when transformed into S. acidocaldarius E4-39 on at least three different occasions (Table 3 and Figure S4). This result suggests that the regions deleted in pRSP1b and pRSP1c are critical for plasmid function. The other two plasmids (pRSP1a and pRSP1d) yielded transformants with a very tiny growth phenotype on solid medium (Table 3 and Figure S4). Transformants from pRSP1a grew poorly in liquid medium; hence we were not able to recover plasmid DNA, whereas pRSP1d transformants did not grow in liquid medium after several attempts.

Comparative analysis of the 232-bp sequence with pRN2, pHEN7, and pDL10 revealed that the 232-bp sequence is fairly conserved within the pRN family (Figure S5). The alignment also showed that the 232-bp sequence in pRN1 is much closer in homology to that in pHEN7 than to that in pRN2 or pDL10. We identified a 12-bp sequence harboring a GTG motif that was conserved in all four plasmids (Figure S5) and deleted this motif from pRSP1 to generate pRSP1e (Table 1). It was previously reported that the primase activity of orf904 requires a GTG motif [35]. The resulting plasmid (pRSP1e) replicated normally, indicating that this particular GTG motif was not essential for plasmid replication (Table 2).

We determined the secondary structure of the 232-bp feature with mFold using 70°C, Na⁺ = 1.0 M and Mg²⁺ = 0.0 M as input parameters [40]. The analysis identified two stem-loop structures, a 19-bp structure at position 4693 - 4711 and a 100-bp structure at position 4734 - 4833 in pRSP1 (corresponding to 5043 - 5061 and 5084 - 5183 respectively, in pRN1) (Figure 4). A significant portion of the region encompassing the 100-bp stem-loop was deleted in pRSP1b and pRSP1c and could possibly explain the inability to recover transformants from cells transformed with these plasmids (Table 3). The 100-bp stem-loop structure is consistent with the structure of the critical double-strand origin of a rolling circle plasmid [28,29,41]. Lack of transformants from cells transformed with pRSP1b and pRSP1c suggests that part or the entire 100-bp stem-loop structure is critical for replication and likely site of replication initiation.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Figure 4. Stem-loop structures identified within the 232-bp pRN1 putative origin of replication.

(A) The secondary structure of 232-bp putative origin of replication determined using DNA mFold, revealed the presence of 100-bp (4734 - 4833) and 19-bp (4693 - 4711) stem-loop (SL) structures at 70°C. (B) DNA sequence of the 232-bp putative origin of replication showing the 100-bp SL (single underline) and 19-bp SL (double underline) in pRN1 origin. Deletions made in pRSP1a (between ∆), pRSP1b (between∇), pRSP1c (between ▼) and pRSP1d (between ▲) are highlighted by the triangles, while the deletion in pRSP1e is shaded. The ‘GTG’ motif (box) and the loop of the 100-bp SL (red) are also highlighted.

https://doi.org/10.1371/journal.pone.0084664.g004

The 19-bp stem-loop structure was part of the sequence deleted in pRSP1d that produced tiny S. acidocaldarius E4-39 transformants that failed to grow in liquid medium. The role of this 19-bp structure is not clear, but based on its location at the 5’-end of the 100-bp stem-loop structure, we suspect that it might be part of the single-strand origin of replication [42]. It also possible to associate inability of pRSP1d transformants to grow in liquid medium to deletion of the 19-bp structure based on a previous report that deletion of the single-strand origin from a rolling circle plasmid results in plasmid instability [43]. The phenotype displayed by pRSP1d could not be attributed to deletion of the 19-bp stem-loop alone because the deleted sequence also contained two other potential small hairpin structures. Similar small potential hairpin structures are present in the region deleted in pRSP1a. Further work would be required to verify the actual function of the features deleted in pRSP1a and pRSP1d (Figure S6).

Loop of the 100-bp stem-loop structure is critical for replication

In pursuance of our proposal that the 100-bp stem-loop might be the double-strand origin (Figure 4), we deleted the 58-bp loop of the structure from pRSP1 to generate pRSP1-L4 (Figure 2). The loop of the 100-bp stem-loop structure could harbor the putative ‘nick site’ that initiates plasmid replication [28,41]. The 58-bp deletion mutant (pRSP1-L4) failed to yield S. acidocaldarius transformants (Table 3), suggesting that the deleted region was critical for plasmid replication. To further characterize this region, we divided the 58-bp loop into three segments, 18 or 20-bp in length, and deleted each segment from pRSP1 generating pRSP1-L1 – pRSP1-L3. Each plasmid (pRSP1-L1 – pRSP1-L3) reproducibly failed to produce S. acidocaldarius E4-39 transformants after two attempts, compared to pRSP1, which consistently yielded transformants (Table 3). The inability to recover transformants with pRSP1-L1 – pRSP1-L4 supports the importance of the 58-bp loop for plasmid function, further highlighting the potential role of the 100-bp structure as the cis-acting double-strand origin of pRN1 (Figure 4).

We further analyzed a GC-rich region in the middle of the 58-bp loop (4773 - 4786) by making 3-bp deletions from pRSP1 (Figure 2B), generating eight plasmids (pRSP1-G1 – pRSP1-G8). The G-C rich region contains a ‘GTG’ motif that was deleted in pRSP1-G3. Each plasmid was transformed into S. acidocaldarius E4.39 along with the pRSP1 as a positive control. Only pRSP1 produced transformants; the other plasmids (pRSP1-G1 – pRSP1-G8) did not yield transformants after more than three attempts. The results from the 58-bp loop deletion mutants and those from pRSP1b and pRSP1c strongly highlight the critical nature of 100-bp stem-loop structure and its potential function as the putative double-strand origin of replication. We attempted to determine if the 58-bp is sequence specific by making a few single A → G base-pair substitutions 3’ of the ‘GTG’ motif in pRSP1, generating pRSP1-A1 (A4748G), pRSP1-A2 (A4786G), pRSP1-A3 (A4787G) and pRSP1-A7 (A4794G) and these plasmids yielded LacS-expressing transformants comparable to pRSP1.

The trans-acting Orf56 and Orf904 complementation

We attempted to verify the report that orf56 and orf904 are critical for pRN1 replication [7,25] by making in-frame deletions of each gene from pRSP1 generating pRSP10 and pRSP9, respectively (Figure 5). Each of the genes was placed under the control of the putative promoter elements of the co-transcribed genes [22] located at the 5’-end of orf56. We consistently recovered transformants with pRSP1 but not with pRSP9 or pRSP10 (Table 4), confirming the importance of these genes as previously reported [7,25]. We explored the possibility of complementing the genes in-trans by transforming both plasmids into S. acidocaldarius E4.39, but did not recover transformants with the pair of plasmids. We evaluated the possibility that some elements at the 5’ of orf904 start codon might be important for transcription and translation of the gene by generating a third plasmid pRSP10b (Figure 5). The new plasmid (pRSP10b) differs from pRSP10 by the presence of the contiguous 75-bp sequence at the 5’ of orf904 start codon (Figure 5B). As expected, transformation of pRSP10b into S. acidocaldarius E4-39 did not produce transformants (Table 4). However, LacS-expressing transformants were recovered with the pRSP9/pRSP10b pair, albeit with a very low efficiency (Table 4). The pRSP9/pRSP10b transformants grew very poorly in liquid medium, making it challenging to recover plasmid DNA for further analysis. The reason for poor growth of the pRSP9/pRSP10b is not known but the transformation result was consistent after more than three attempts. In addition to the suggestion that some elements within the additional 75-bp sequence in pRSP10b might be important for plasmid function; the result from the pRSP9/pRSP10b transformation highlights the possibility of transforming S. acidocaldarius with more than one plasmid.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Figure 5. In-trans complementation of orf56 and orf904 deletion mutants of pRSP1.

(A) Deletion mutants of pRSP1 generated by deleting orf904 (pRSP9) and orf56 (pRSP10/pRSP10b) from pRSP1; pRSP10 contained a complete orf904 under the control of orf56 transcription and translation elements, while pRSP10b contained an additional 75-bp at the 5’ of orf904 start codon. Each plasmid contains pUC19 and pRN1 origins of replication, the pyrE-lacS cassette transcribed by the S. solfataricus thsB promoter [P(β)] and lacS terminator element [lacS(T)], and AP^r gene (ampicillin resistance). (B) Cis-acting elements at the 5’ of orf56 and orf904 in each plasmid. Transcription of orf904 in pRSP10 and pRSP10b is controlled by of orf56 promoter elements (big bold box) with putative TATA box A (bold single underline) and TATA box B (single underline). The 75-bp element 5’of orf904 in pRSP10b (bold underline) contains a putative RBS (small box). The start of orf904 is bolded, and the stop of orf56 is represented by (***).

https://doi.org/10.1371/journal.pone.0084664.g005