Creating Bacterial Strains from Genomes That Have Been Cloned and Engineered in Yeast

See allHide authors and affiliations

Science  25 Sep 2009:
Vol. 325, Issue 5948, pp. 1693-1696
DOI: 10.1126/science.1173759

Character Transplant

When engineering bacteria, it can be advantageous to propagate the genomes in yeast. However, to be truly useful, one must be able to transplant the bacterial chromosome from yeast back into a recipient bacterial cell. But because yeast does not contain restriction-modification systems, such transplantation poses problems not encountered in transplantation from one bacterial cell to another. Bacterial genomes isolated after growth in yeast are likely to be susceptible to the restriction-modification system(s) of the recipient cell, as well as their own. Lartigue et al. (p. 1693, published online 20 August) describe multiple steps, including in vitro DNA methylation, developed to overcome such barriers. A Mycoplasma mycoides large-colony genome was propagated in yeast as a centromeric plasmid, engineered via yeast genetic systems, and, after specific methylation, transplanted into M. capricolum to produce a bacterial cell with the genotype and phenotype of the altered M. mycoides large-colony genome.


We recently reported the chemical synthesis, assembly, and cloning of a bacterial genome in yeast. To produce a synthetic cell, the genome must be transferred from yeast to a receptive cytoplasm. Here we describe methods to accomplish this. We cloned a Mycoplasma mycoides genome as a yeast centromeric plasmid and then transplanted it into Mycoplasma capricolum to produce a viable M. mycoides cell. While in yeast, the genome was altered by using yeast genetic systems and then transplanted to produce a new strain of M. mycoides. These methods allow the construction of strains that could not be produced with genetic tools available for this bacterium.

We have described the transplantation of the genome of Mycoplasma mycoides subspecies capri (13) from its native cellular environment into a related species, Mycoplasma capricolum subspecies capricolum (4). We have also described the complete chemical synthesis of the 580-kb Mycoplasma genitalium genome (5, 6). Initial stages of the synthesis were carried out by in vitro assembly reactions, and pieces up to a quarter of a genome in size were cloned in Escherichia coli. We overcame difficulties in cloning larger segments of DNA in E. coli by using homologous recombination in the yeast Saccharomyces cerevisiae to assemble the subgenomic synthetic DNA segments into a complete M. genitalium genome. To complete our construction of a living microbe, we must isolate our synthetic genome from yeast and transfer it into a cellular environment that will accept and execute the genetic instructions sufficient to produce a replicating organism. In this paper, we describe methods for transplanting natural 1.1-Mb M. mycoides genomes cloned in yeast into M. capricolum recipient cells. These species are more convenient experimental organisms than M. genitalium because of their faster growth rate.

M. mycoides was transformed (7) with a vector containing a selectable tetracycline-resistance marker and a β-galactosidase gene for screening. The vector also contained a yeast auxotrophic marker, a yeast centromere, and a yeast autonomously replicating sequence, for selection and propagation in yeast as a yeast centromeric plasmid (YCp). Direct genomic sequencing (8) of one clone (YCpMmyc1.1) showed that the entire vector integrated into the genome. This clone grew robustly and transplanted efficiently into M. capricolum (9), so it was chosen for cloning into yeast. The genome of this clone will be called YCpMmyc1.1 throughout this paper, regardless of the cellular source. YCpMmyc1.1 can refer to: (i) the original M. mycoides strain (the “native” M. mycoides YCpMmyc1.1 genome), (ii) the same genome cloned in yeast, (iii) the genome transplanted from M. mycoides or from yeast, or (iv) this genome as free DNA from any of these sources.

YCpMmyc1.1 genomes were isolated from M. mycoides (9) and transformed into yeast spheroplasts (10) of strains VL6-48N (11) and W303a. Clones were analyzed for completeness and size by multiplex polymerase chain reaction (PCR) and clamped homogenous electric fields (CHEF) gel electrophoresis. To test whether deletions occur during routine propagation in yeast, we screened 40 individual colonies derived from a single intact clone of YCpMmyc1.1 in W303a. All appeared to contain complete genomes (fig. S1), which indicates that this bacterial genome is stable in yeast. Sequences between the amplicons were not interrogated in this experiment; however, sizable deletions would have been detected by this approach, and none were observed. Sequencing of a complete genome transplanted from yeast (see below) provided a definitive demonstration of stability of the M. mycoides genome in yeast.

We engineered YCpMmyc1.1 in yeast by creating a seamless deletion in a nonessential Type III restriction endonuclease gene (Fig. 1). This modification cannot be made with the genetic tools available for this bacterium. We first transformed a YCpMmyc1.1 yeast clone with a cassette containing a URA3 marker and the SCEI endonuclease gene under the control of the GAL1 promoter. We selected for insertion of the cassette into the Type III gene. Four of five clones contained intact genomes, and one contained a genome with a large deletion (YCpMmyc1.1-Δ500kb) (figs. S2 and S3). The URA3 cassette was removed by cleavage at an I-Sce I recognition site near one end of the cassette (Fig. 1). Counter selection with 5-fluoroorotic acid (12) produced clones that had lost the URA3 cassette. Thus, we obtained two M. mycoides YCp genomes, one that contained the URA3 cassette and the other that contained a seamless deletion of the Type III restriction enzyme gene (Fig. 1A). The changes to the genome were verified by PCR (Fig. 1B).

Fig. 1

Generation of Type III restriction enzyme deletions. (A) To make an M. mycoides Type III restriction enzyme gene (typeIIIres) deletion in yeast (iii), we constructed a linear DNA fragment, knockout cassette, by fusing two PCR products, CORE and tandem repeat sequence (TR) (i). This cassette was then transformed into a yeast W303a strain harboring the YCpMmyc1.1 M. mycoides genome (ii). Growth on (–)His (–)Ura medium selected for replacement of the Type III restriction enzyme open reading frame (ORF) by the cassette via the 50–base pair (bp) sequences homologous to the target sites (ΔtypeIIIres::URA3). Galactose induction results in the expression of I-Sce I endonuclease, which cleaves the 18-bp I-Sce I site (asterisk) to create a double-strand break that promotes homologous recombination between two tandem repeat sequences (TR) (red line). Recombination between the TRs creates a seamless deletion of the typeIIIres gene (ΔtypeIIIres), which was isolated following 5-flouroorotic acid (5-FOA) counterselection against the URA3 gene. IGR, intergenic region. (B) The arrows above the DNA in (A) represent PCR primers (P299 and P302) used to verify the presence or absence of the knockout cassette. PCRs of representative transplant clones with (ii) and without (iii) the knockout cassette are shown. PCRs of the YCpMmyc1.1 clone in yeast (i) are shown for comparison. The expected sizes are obtained for each amplicon.

We isolated YCpMmyc1.1 from yeast and attempted transplantation into wild-type M. capricolum cells. However, we did not recover any transplants (Table 1). We reasoned that the principal obstacle was a restriction endonuclease in the recipient M. capricolum that degraded the unmethylated YCpMmyc1.1 donor DNA isolated from yeast (fig. S4).

Table 1

Transplantation of M. mycoides YCp genomes from yeast into wild-type and RE(–) M. capricolum recipient cells. The number of tetracycline-resistant, blue colonies obtained after the transplantation of M. mycoides YCp genomes from yeast into M. capricolum recipients was counted. Wild-type M. capricolum and M. capricolum RE(–) transplantation was performed using methods described in fig. S1. For untreated samples, yeast plugs were digested with β-agarase (melting step) and transplanted into both recipient cells. The treated samples were methylated and treated with proteinase K before the melting step. The mock-methylated sample was treated the same as the methylated samples, except that no extract or purified methyltransferases were added. VL6-48N yeast agarose plugs used in this experiment carried YCpMmyc1.1. W303a yeast agarose plugs carried YCpMmyc1.1, YCpMmyc1.1 that was engineered in yeast (YCpMmyc1.1-ΔtypeIIIres::URA3 or YCpMmyc1.1-ΔtypeIIIres), or YCpMmyc1.1-Δ500kb. The number of transplants is the average of at least three experiments. The error reported is the absolute mean deviation.

View this table:

Two methods were used to overcome the M. capricolum restriction barrier. First, we inactivated the single restriction enzyme in M. capricolum by integration of a puromycin-resistance marker into the coding region of the gene. No detectable restriction enzyme activity was seen in extracts of this altered strain [M. capricolum RE(–)] (fig. S5). Removal of M. capricolum restriction activity should allow donor M. mycoides YCp genomes isolated from yeast to survive initial contact with the M. capricolum cytoplasm. The second method was to protect the donor DNA isolated from yeast by in vitro methylation, using M. capricolum extracts. An extract of M. mycoides also protected the incoming donor DNA, because M. mycoides contains an ortholog of the system found in M. capricolum (fig. S6). The additional restriction-modification systems present in the M. mycoides donor genome did not affect transplantation.

We isolated YCpMmyc1.1 from M. mycoides and transplanted it into wild-type M. capricolum and M. capricolum RE(–) recipient cells (fig. S7). Results were scored by selecting for growth of blue colonies on SP4 medium containing tetracycline at 37°C. Successful transplantations were obtained using YCpMmyc1.1 from yeast with both recipient cells (Table 1). Colonies were obtained using M. capricolum RE(–) as recipient cells when the donor genomic DNA was untreated, mock-methylated, treated with M. capricolum or M. mycoides extracts, or treated with purified M. mycoides methyltransferases. However, transplantation using wild-type M. capricolum recipient cells occurred only when the donor YCp genome from yeast was methylated with M. capricolum extract, M. mycoides extract, or purified M. mycoides methyltransferases. No colonies were obtained when mock-treated or untreated YCpMmyc1.1 was transplanted into wild-type M. capricolum recipient cells. Thus, avoidance of the M. capricolum recipient restriction system is vital for successful transplantation of M. mycoides YCp genomes from yeast.

YCpMmyc1.1, as well as the engineered YCp genomes (YCpMmyc1.1-ΔtypeIIIres::URA3 and YCpMmyc1.1-ΔtypeIIIres), were also isolated from yeast strain W303a. Transplantation of all three YCp genomes into M. capricolum recipient cells resulted in similar numbers of tetracycline-resistant blue colonies (Table 1). The large deletion clone (YCpMmyc1.1-Δ500kb) discussed above served as an appropriate control because it lacks many presumed essential genes yet retains the YCp element and tetM. As expected, no colonies were recovered when this genome was transplanted into M. capricolum recipient cells.

Recovery of colonies in all these transplantation experiments was dependent on the presence of both M. capricolum recipient cells and an M. mycoides genome. The experiments described here used donor YCp genome DNA that included yeast genomic DNA. However, purifying the donor YCp genome DNA away from yeast genomic DNA did not substantially alter transplantation results, which suggests that the recipient M. capricolum cells are able to tolerate the presence of nonspecific or carrier DNA (Table 1). Positive transplantation results were obtained with donor YCp genome DNA isolated from four independent transformant cultures of strain VL6-48N and four of strain W303a. Thus, bacterial genomes can be stably cloned in both yeast strains.

We verified that the recovered colonies were M. mycoides by Southern blot analysis using an M. mycoides–specific IS1296 element as probe (Fig. 2A). We showed that the Type III restriction gene was deleted in the engineered bacterium by PCR (Fig. 1B), by Southern blot analysis using the Type III restriction gene sequence as probe (Fig. 2B), and by sequencing the locus (Fig. 2C).

Fig. 2

Southern blot analysis of M. mycoides transplants. (A) Hind III–restricted genomic DNA from representative YCpMmyc1.1, YCpMmyc1.1-ΔtypeIIIres::URA3 (next-to-last columns), or YCpMmyc1.1-ΔtypeIIIres (last columns) M. mycoides genome transplants from yeast was probed with IS1296 sequences. The genomic DNAs of native M. mycoides YCpMmyc1.1 cells and M. capricolum recipient cells were also probed for comparison. All of the selected transplant colonies contain the same IS1296 pattern as native M. mycoides YCpMmyc1.1. (B) The engineered M. mycoides transplants were tested for the absence of the Type III restriction enzyme gene by Southern blot analysis (representative clones shown). Eco RV–restricted genomic DNA from YCpMmyc1.1-ΔtypeIIIres::URA3, and YCpMmyc1.1-ΔtypeIIIres M. mycoides cells derived by genome transplantation was probed with the typeIIIres gene sequence. The genomic DNAs of native M. mycoides YCpMmyc1.1 cells and M. capricolum recipient cells were also probed for comparison. The typeIIIres gene is absent in the engineered genomes but present in the native M. mycoides YCpMmyc1.1 genome. (C) Sequencing of the seamless deletion region of the YCpMmyc1.1-ΔtypeIIIres M. mycoides genome transplant verified that the Type III restriction gene was removed as designed. The sequence text colors are the same as the gene region colors in the genetic maps in Fig. 1. A small portion of the typeIIIres gene remained after the deletion because of the overlap between the typeIIImod and typeIIIres genes. The start and stop codons of the typeIIIres gene are boxed in red; the stop codon of the typeIIImod gene is boxed in black.

To confirm that our transplants were entirely M. mycoides and not chimeras containing yeast sequences or M. capricolum recipient cell sequences, we sequenced the genome of one transplant from which the Type III restriction gene had been deleted (GenBank accession no. CP001668) (9). All of the assembled genome matched M. mycoides, except for those regions that matched the YCp vector. In addition, except for alterations we made in the genome, our transplant YCpMmyc1.1-ΔtypeIIIres genome sequence was identical to M. mycoides YCpMmyc1.1, which was used to generate the original M. mycoides yeast clone. Thus, there was no recombination of either yeast or recipient cell genomes with the M. mycoides YCp donor genome, and these bacterial genome sequences are stable as YCps.

Here we describe methods to transfer a genome between branches of life, from a bacterium to a eukaryote, where it can be genetically altered, and then back again into a bacterium (Fig. 3). Our method of yeast vector insertion uses bacterial selection to ensure a viable integration point. This avoids a potentially lethal genome disruption, which would prevent transplantation. Full-length clones of mycoplasma genomes have proven stable in yeast during routine propagation and genome transplantation. As described, 40 individual colonies derived from a complete YCp clone of the M. mycoides genome all contained full-length genomes. Furthermore, eight independent full-length yeast clones of the M. mycoides genome yielded viable bacteria when the genome was transplanted. Finally, the complete genome sequence of M. mycoides was unchanged during cloning into yeast and transplantation back into a bacterial cell. We have never seen deletions in our YCp clones except after selection following DNA transformation.

Fig. 3

Moving a bacterial genome into yeast, engineering it, and installing it back into a bacterium by genome transplantation. A yeast vector is inserted into a bacterial genome by transformation. That genome is cloned into yeast. After cloning, the repertoire of yeast genetic methods is used to create insertions, deletions, rearrangements, or any combination of modifications in the bacterial genome. This engineered genome is then isolated and transplanted into a recipient cell to generate an engineered bacterium. Before transplantation it may be necessary to methylate the donor DNA in order to protect it from the recipient cell’s restriction system(s). This cycle can be repeated starting from the newly engineered genome (dashed arrow).

We previously reported transplantation of naked genomic DNA purified from M. mycoides cells (4). The transplant events were rare, and there remained the possibility that they resulted from damaged cells that could be somehow repaired in the presence of recipient cells, or from genomes that were in complex with some M. mycoides component other than genomic DNA. Transplantation from yeast of the nonmethylated M. mycoides genome into the M. capricolum RE(–) recipient cells eliminates the possibility that components of the M. mycoides cells are required for transplantation (fig. S7).

Our original transplant experiments used genomes that were resistant to the restriction enzyme of the recipient cells because the donor cells contain the same restriction modification system as the recipient cell (4). M. mycoides DNA sequences from yeast lack the specific methylations imparted by M. mycoides restriction modification systems. The natural DNA sequences encoding the methylases cannot be expressed because they contain UGA tryptophan codons, which function as stop codons in yeast. Transplantation from yeast was achieved either by methylation of the donor genome in vitro or by inactivation of the restriction enzyme in the recipient cell. It was unnecessary to protect the M. mycoides genome from its own restriction systems, because either inactivation of the recipient cell’s endonuclease or methylation with the recipient cell’s methylase was sufficient to allow transplantation. When transplanting other bacterial genomes from yeast, it may be necessary to methylate the donor genome in vitro to protect it from its own restriction enzymes.

Genetic manipulation of M. mycoides, and mycoplasmas in general, is limited. Integration of plasmid DNA by single crossover events allows the targeted addition or disruption of genes in M. mycoides (13). However, because there are only a few selection markers, the number of genetic alterations that can be performed in a single M. mycoides cell is limited. The maintenance of the M. mycoides genome in yeast allowed us to access the powerful repertoire of yeast genetic methods and to produce an M. mycoides strain that had not previously existed. Thus, we report engineering of a bacterial cell by altering its genome outside of its native cellular environment.

It is now possible to readily generate M. mycoides strains with multiple targeted gene deletions, insertions, and rearrangements. It would also be possible to engineer bacterial genomes in yeast by using random mutagenesis methods so that transplantation would yield populations of altered bacteria. After screening for a desired trait, these methods could be reapplied in a cyclical manner to introduce new traits (Fig. 3). This transplantation system potentially allows M. mycoides and closely related species to be model systems for exploring the pathogenicity and biology of mycoplasmas. The mycoides group of mycoplasmas causes major diseases of ruminants, and there is an urgent need for vaccines (14, 15). This technology could accelerate the construction of live vaccine strains.

Many medically or industrially important microbes are difficult to manipulate genetically. This has severely limited our understanding of pathogenesis and our ability to exploit the knowledge of microbial biology on a practical level. We hope that the cycle presented here can be applied to other species, to help solve these problems.

Supporting Online Material

Materials and Methods

Figs. S1 to S7

Table S1


  • * Present address: Biotechnology Industry Organization (BIO), 1201 Maryland Avenue SW, Washington, DC 20024, USA.

References and Notes

  1. Manso-Silvan et al. (2) report the very recent renaming of Mycoplasma mycoides subspecies mycoides Large Colony to Mycoplasma mycoides subspecies capri.
  2. Materials and methods are available as supporting material on Science Online.
  3. We thank R. Roberts for helpful information and the Synthetic Biology team at the J. Craig Venter Institute (JCVI) for critical discussions about the manuscript. This work was supported by Synthetic Genomics, Inc. (SGI). J.C.V. is Chief Executive Officer and Co-Chief Scientific Officer of SGI. H.O.S. is Co-Chief Scientific Officer and on the Board of Directors of SGI. C.A.H. is Chairman of the SGI Scientific Advisory Board. All three of these authors and JCVI hold SGI stock. JCVI has filed patent applications on some of the techniques described in this paper.
View Abstract

Stay Connected to Science

Navigate This Article