One-step assembly in yeast of 25 overlapping
DNA fragments to form a complete synthetic
Mycoplasma genitalium genome
Daniel G. Gibsona,1, Gwynedd A. Bendersb, Kevin C. Axelroda, Jayshree Zaveria, Mikkel A. Algirea, Monzia Moodiea,
Michael G. Montaguea, J. Craig Ventera, Hamilton O. Smithb, and Clyde A. Hutchison IIIb,1
aThe J. Craig Venter Institute, Synthetic Biology Group, Rockville, MD 20850 andbThe J. Craig Venter Institute, Synthetic Biology Group,
San Diego, CA 92121
Contributed by Clyde A. Hutchison III, October 30, 2008 (sent for review September 11, 2008)
We previously reported assembly and cloning of the synthetic
Mycoplasma genitalium JCVI-1.0 genome in the yeast Saccharo-
myces cerevisiae by recombination of six overlapping DNA frag-
ments to produce a 592-kb circle. Here we extend this approach by
demonstrating assembly of the synthetic genome from 25 over-
lapping fragments in a single step. The use of yeast recombination
greatly simplifies the assembly of large DNA molecules from both
synthetic and natural fragments.
in vivo DNA assembly ? genome synthesis ? combinatorial assembly ?
yeast transformation ? Mycoplasma genitalium ? synthetic biology
ments. More than 30 years ago, Hinnen et al. (1) reported the
restoration of leucine biosynthesis in Saccharomyces cerevisiae by
spheroplasts, CaCl2, and PEG. Soon after, Orr-Weaver et al. (2)
reported mechanistic studies demonstrating that DNA molecules
taken up during yeast transformation can integrate into yeast
chromosomes through homologous recombination, and that the
ends of the linear-transforming DNA are highly recombinogenic
and react directly with homologous chromosomal sequences,
whereas circular plasmids carrying yeast sequences integrate by a
single crossover and only at low frequency. Subsequently, yeast
transformation has become an indispensable tool in yeast genetics.
Yeast recombination has since been applied to the construction
of plasmids and yeast artificial chromosomes (YACs). In 1987, Ma
et al. (3) constructed plasmids from two cotransformed DNA
fragments containing homologous regions. In another process,
called linker-mediated assembly, any DNA sequence can be joined
to a vector DNA using short synthetic linkers that bridge the ends
(4, 5). Similarly, four or five overlapping DNA pieces can be
assembled and joined to vector DNA (4, 6, 7). This work demon-
assemble the pieces into a single recombinant molecule.
The limitations of assembly methods in yeast remain unknown.
We recently assembled an entire synthetic M. genitalium genome
using a combination of in vitro enzymatic recombination in early
stages and in vivo yeast recombination in the final stage to produce
the complete genome (8). In the first stage, overlapping ?6-kb
DNA cassettes were joined four at a time to form 25 ?24-kb
A-series assemblies. Three A-series assemblies were then joined to
make 1/8 genome ?72-kb B-series assemblies, and then two B-
series assemblies were assembled to make each of the ?145-kb
quarter-genome C-series assemblies. We accomplished the final
assembly in yeast using three quarter-genome fragments and a
fourth quarter fragment that had been cleaved by a restriction
enzyme to provide a site for insertion of the vector DNA. Thus,
up the overlapping genome and vector DNA fragments (a total of
six pieces) and recombine them to produce the full 592-kb circular
east has long been considered a genetically tractable organism
because of its ability to take up and recombine DNA frag-
JCVI-1.0 genome. At the time, we wondered whether it would be
possible to assemble DNA fragments from earlier stages into a
complete genome in a single step in yeast.
The assembly intermediates from our construction of the syn-
of yeast uptake and assembly. Here we report the successful
assembly of the entire synthetic genome from 25 A-series assem-
blies in a single step.
Assembly of the M. genitalium Genome in Yeast from 25 Overlapping
DNA Fragments. Our 25 A-series assemblies comprising the entire
M. genitalium genome were all synthesized and cloned as bacterial
artificial chromosomes (BACs) in Escherichia coli as described
previously (8) (Table 1). For this study, two variants of assembly
JCVI-1.1 and JCVI-1.9. Each variant has a vector inserted at the
same site within a nonessential gene. Each vector contains a
histidine auxotrophic marker, a centromere, and an origin of
replication for selection and maintenance in yeast. The only dif-
ference between JCVI-1.1 and JCVI-1.9 is in the additional vector
features. All 25 of these A-series molecules were purified after
propagation in E. coli. The assemblies were released from their
respective BACs by cleavage with restriction enzymes (Table 1).
The presence of yeast propagation elements within one of our
ability to assemble all of these 25 fragments into the M. genitalium
cell take up all 25 of these pieces, translocate them to the nucleus,
and then recombine them into the complete genome?
After digestion, each of the 25 assemblies contained at least 80
(Table 1). Equal amounts (?100 ng or ?4 ? 109genome equiva-
lents) of each digested piece were pooled together without gel
purification to assemble M. genitalium JCVI-1.1. Then ?108yeast
spheroplasts were added to this DNA pool, and yeast transforma-
equivalents, or ?1000 DNA molecules (40 ? 25) per yeast sphero-
plast. After incubation for 3 days at 30 °C on selective medium,
?800 colonies were obtained.
Author contributions: D.G.G., G.A.B., J.C.V., H.O.S., and C.A.H. designed research; M.A.A.
and M.G.M. contributed new reagents/analytic tools; D.G.G., G.A.B., K.C.A., J.Z., and M.M.
performed research; D.G.G., G.A.B., H.O.S., and C.A.H. analyzed data; and D.G.G., G.A.B.,
H.O.S., and C.A.H. wrote the paper.
Genomics, Inc (SGI). The research reported in this paper is in the field in which SGI is
developing products. SGI has funded the research reported in the paper.
Freely available online through the PNAS open access option.
1To whom correspondence may be addressed. E-mail: email@example.com or chutchison@
© 2008 by The National Academy of Sciences of the USA
December 23, 2008 ?
vol. 105 ?
no. 51 www.pnas.org?cgi?doi?10.1073?pnas.0811011106
Analysis of 10 Clones by Multiplex PCR. We first screened for yeast
cells that took up all 25 pieces. Forty primer pairs, producing
amplicons ranging in size from 100 bp to 1075 bp, were designed
multiplex PCR reactions (Fig. 2A). Multiplex primer sets 1 and 2
were described previously (8). The 40 amplicons are positioned
around the M. genitalium genome approximately every 15 kb (Fig.
by PCR, thus giving a very good indication that all 25 pieces were
incorporated into a yeast cell.
Ten individual colonies were transferred to a single selective
plate as small patches. After incubation for 2 days at 30 °C, ?107
cells from each of these 10 patches were scraped into 1 ml of water.
Using multiplex primer set 1 and DNA extracted from these 10
clones, multiplex PCR was performed, and the products were
analyzed by gel electrophoresis. Clones 1 and 4 produced all 10
amplicons (Fig. 2C). Amplicon 1-h was observable when the
exposure time was increased. This PCR analysis confirmed the
Table 1. Summary of the 25 A-series assemblies used
in constructing the M. genitalium synthetic genomes
SalI and NotI
25 overlapping DNA fragments
A1-4, A5-8, etc. (17-35 kb)
genome in yeast
were transformed with 25 different overlapping A-series DNA segments (blue
arrows; ?17 kb to ?35 kb each) composing the M. genitalium genome. To
genome, called JCVI-1.1, is 590,011 bp, including the vector sequence (red trian-
gle) shown internal to A86–89. The yeast propagation elements contained
depicted here, some yeast cells may take up fewer than 25 different pieces and
produce subassemblies of the genome by a mechanism such as NHEJ (see text).
Others may take up more than 25 fragments and produce more than one
assembled molecule per cell (not illustrated).
Construction of a synthetic M. genitalium genome in yeast. Yeast cells
c10 c9c8 c7c6c5c4 c3c2 c1L
A1-4 A5-8A9-12 .....etc.
Set 1 Set 2 Set 3 Set 4
Amplicon size (bp)
Primer setPrimer set
segments. (A) Forty amplicons were designed such that 10 products could be
could be easily separated by electrophoresis on 2% agarose gels (C and D). (B)
The 40 sets of primers amplified a small portion of the M. genitalium genome
approximately once every 15 kb. Each of the 25 fragments (gray arrows)
provided primer-binding sites for at least 1 of the 40 amplicons (red, green,
1, was performed. Clones 1 and 4 (c1 and c4) efficiently generated 9 of the 10
predicted amplicons. Clone 4 was selected for further analysis. (D) Multiplex
PCR was performed on clone 4 and JCVI-1.0 as a positive control using all four
primer sets. With the exception of amplicon 1-h, all 40 amplicons were
efficiently generated from these clones. In the lanes labeled ‘‘L,’’ the 100-bp
ladder (New England Biolabs) was loaded; sizes are indicated.
Multiplex PCR analysis to screen for yeast cells that took up all 25
Gibson et al.
December 23, 2008 ?
vol. 105 ?
no. 51 ?
presence of 10 of the 25 assemblies in clones 1 and 4 (compare with
Detailed Analysis of Clone 4. Individual colonies, obtained by re-
streaking the patches onto selective medium, were analyzed (see
Materials and Methods). Using multiplex primer sets 1–4, multiplex
PCR was performed, and the products were visualized using gel
electrophoresis (Fig. 2D). With the exception of amplicon 1-h, all
40 PCR products were efficiently generated from DNA extracted
from clone 4. This confirms the presence of all 25 pieces (compare
Fig. 2B). A70–73 was represented by amplicons 1-h and 3-h.
amplicons, amplicon 3-h was produced efficiently, providing strong
evidence for the presence of A70–73 in clone 4. Furthermore,
amplicon 1-h was not efficiently generated even in JCVI-1.0, the
sequence-confirmed positive control (Fig. 2D) (8).
To determine whether all 25 pieces were assembled into the
complete genome in clone 4, DNA was prepared from these yeast
cells in agarose plugs for restriction analyses. As a negative control,
DNA also was extracted from the untransformed host strain. To
enrich for the circular genome in the plugs, most of the linear yeast
sis. Digestion was done with EagI, BssHII, and AatII (Fig. 3A and
B). Five restriction fragments predicted by these digests for a
complete genome assembly were observed when the digested plugs
were subjected to field-inversion gel electrophoresis (FIGE) (Fig.
3C). The host yeast control showed none of these bands, and only
a smear of DNA smaller than ?150 kb was observed. This smear
obscured the other predicted fragments of the assembled genome.
Together with the multiplex PCR analyses (Fig. 2), these results
indicate that clone 4 had an assembled JCVI-1.1 synthetic genome,
and thus yeast cells have the capacity to take up at least 25
overlapping segments of DNA and recombine them in vivo.
Clone 1 exhibited a pattern similar to that of clone 4 when
multiplex PCR products, amplified using primer set 1, were ana-
lyzed (Fig. 2C). Therefore, we analyzed clone 1 further. Using the
four multiplex PCR primer sets shown in Fig. 2, all 40 amplicons
3), the predicted 590-kb band was present (data not shown). Thus,
2 out of 10 yeast clones were confirmed to contain a complete
JCVI-1.1 genome after this single transformation experiment.
Accurate Assembly of the 25 Overlapping Segments. Many organ-
isms, including S. cerevisiae, perform nonhomologous end joining
(NHEJ) as a way to maintain the integrity of the genome in the
presence of double-strand breaks (10). In contrast to homologous
recombination, this mechanism for joining DNA does not depend
joining one or more of the A-series molecules. Moreover, because
most of the fragments contain NotI cohesive ends, some joining
could have occurred by cohesive-end ligation. Either mechanism
could result in a genome containing more than or fewer than one
each of the 25 segments; for example, one segment may not
participate in assembly and be replaced by a duplicate A-series
all 25 pieces could then be explained by multiple subassemblies
within the same cell. To address this possibility, the linearized
JCVI-1.1 (C4) genome fragment was gel-purified along with JCVI-
used as template for multiplex PCR using the four sets of primers
described in Fig. 2. All 40 amplicons were detected from the
gel-purified JCVI-1.0 and JCVI-1.1 genomes (Fig. 4B), confirming
the presence of all 25 segments within the linearized JCVI-1.1
genome. This analysis does not verify that all 25 pieces were
assembled in the correct order, however. To validate a correctly
assembled genome, PCR primers were designed to produce an
were produced after PCR, and the sizes exactly matched those
predicted (Fig. 4D). This convincingly demonstrates that the JCVI-
1.1 genome was assembled in vivo by homologous recombination.
We cannot rule out the possibility that one or more additional
molecules were assembled within the same cell.
Reproducibility of the Single-Step, 25-Piece Assembly. We investi-
of the 25 A-series fragments into yeast yielded 150 colonies. DNA
was extracted from 10 of these colonies (c11–c20), and then
multiplex PCR, using primer set 3, was performed. DNA from
clone 11 produced all 10 amplicons predicted for a complete
assembly (Fig. 5A), so this clone was further analyzed using all four
multiplex primer sets (Fig. 5B). All 40 amplicons were efficiently
molecules within these cells. To determine whether this genome,
termed JCVI-1.9, was completely assembled, DNA from this clone
was prepared in agarose plugs as shown in Fig. 3 and then analyzed
by contour-clamped homogeneous electric field (CHEF) gel elec-
trophoresis alongside a host-only negative control (strain VL6–
48N). A fragment of the correct size for JCVI-1.9 (?587 kb) is
evident (Fig. 5C). To determine whether this genome was assem-
(A) Diagram of the EagI (red), BssHII (blue), and AatII (green) restriction
fragments expected for a complete and proper assembly of the synthetic
(B) The sizes of the five restriction fragments indicated in (A). (C) Total DNA
from clone 4 (Y ? Mg) and its host strain alone (Y) were isolated from yeast
cells embedded in agarose. Most of the linear DNA was electrophoresed out
and analyzed by FIGE on 1% agarose gels. Restriction fragments correspond-
ing to the correct sizes are indicated by the fragment numbers shown in (A)
loaded; sizes are indicated.
Validation of an intact M. genitalium genome by restriction analysis.
www.pnas.org?cgi?doi?10.1073?pnas.0811011106 Gibson et al.
bled in the correct order by homologous recombination, this
at each junction, as shown in Fig. 4. All predicted amplicons were
produced for JCVI-1.9, and the PCR pattern exactly matched that
for the positive control, JCVI-1.0 (Fig. 5D; compare with Fig. 4D),
demonstrating construction of a second accurately assembled syn-
thetic genome by an independent transformation event of 25
Statistics for Yeast DNA Uptake. An individual yeast cell must take
genome. If a cell randomly takes up exactly 25 pieces, then the
probability that it takes up one of each is 25!/2525? 1.7 ? 10?10—a
very rare occurrence. Thus, the average yeast cell likely took up
substantially more than 25 pieces, given the results we observed. If
a yeast cell randomly takes up n ? 25 equally represented pieces,
what is the probability of getting all 25 different pieces? This is an
example of the ‘‘collector’s problem’’ (11). An approximate calcu-
lation of the probabilities for different n can be readily obtained by
then it has about a 50% chance of taking up all 25 different pieces.
In a typical experiment, 108yeast cells are exposed to DNA.
Therefore, even if the average cell were to take up only 40 pieces,
about 1 in 1000 would get a complete set of pieces. The efficiency
need to take up a complete set of overlapping pieces. If the
J01J02J03J04J05J06J07J09 J11 J12J13J14 J15J17J18 J19J20J21J22J23J24J25J08J10J16
two genome fragments of ?233 kb and ?350 kb as well as a released ?9-kb vector. The linearized JCVI-1.0 and JCVI-1.1 genome fragments were cut out of the
gel, as indicated by red rectangles. (B) DNA from the gel slices in (A) was extracted and purified, then used as template for multiplex PCR, as shown in Fig. 2. (C)
Twenty-five amplicons (J1-J25, black lines) are produced from 25 sets of primers that span each of the 25 junctions at which joining of the A-series assemblies
(gray arrows) occurs. The predicted amplicon sizes for correctly assembled molecules are indicated. It was necessary to design larger amplicons for J08, J10, and
J16 to ensure unique primer binding sites due to MgPa repeats at these junctions. (D) Purified JCVI-1.0 and JCVI-1.1 DNA extracted from the gel slices in (A)
produced all 25 expected fragments after PCR. Products were analyzed on 2% agarose gels except for J08, J10, and J16, which were analyzed on 0.8% agarose
gels. In the lanes labeled ‘‘M,’’ the 1-kb ladder (New England Biolabs) was loaded; sizes are indicated.
Gibson et al.
December 23, 2008 ?
vol. 105 ?
no. 51 ?
efficiency is low, then relatively more yeast cells would require a
complete set. Independent measurements of the quantity of DNA
taken up per cell would help answer some of these questions.
Exploring the Capacity of DNA Assembly in Yeast. We still do not
know the limits of DNA assembly in yeast. Once we assembled the
synthetic M. genitalium genome from six pieces, we wondered
whether we could have assembled it from 25. Now we wonder
whether this genome could have been assembled from cotransfor-
mation of all 101 DNA cassettes (8). We also do not yet know the
Mb have been cloned (12). Presumably, any DNA construct can be
assembled inside a yeast cell as long as the overlapping subfrag-
ments can get into the nucleus, and the constructed DNA segment
presence of the highly conserved and essential autonomously
replicating sequence (ARS) consensus sequence, 5?-A
every ?262 kb in an assembled construct when all four bases are
equally represented; however, these elements should be present
more frequently in the M. genitalium genome given its high A?T
composition. Indeed, 20 instances of the ARS consensus sequence
T-3? (13). Statistically, this sequence should be present once
This is an important consideration, because a 170-kb YAC lacking
efficient origins of replication can invoke a cell-cycle checkpoint
response in yeast (14).
Once individual clones have been identified as containing a
complete genome, instability within yeast has not been observed
(8). This is a common concern, however; deletions ranging in size
growth of the transformants have been reported (15). These
deletions can result from recombination between repetitive se-
quences within the genome, which can be suppressed by using a
recombination-deficient yeast strain (16).
Combinatorial Assembly. It should be possible to use the method
described here to create many combinations of DNA molecules in
yeast. Genetic pathways can be constructed in a combinatorial
fashion such that each member in the combinatorial library has a
different combination of gene variants. For example, a combina-
torial library of 65variants can be constructed from 30 individual
DNA elements (6 variants of each of 5 overlapping elements in the
pathway). The elements can be genes, transcriptional promoters
and terminators, and appropriate signals for initiation of protein
synthesis. Through screening and selection methods, yeast cells
bearing the best pathways with the highest yields can be obtained.
The yeast then can become factories for the manufacture of the
products specified by the designed pathways. Alternatively, if yeast
a pool of individual yeast clones, each containing a variant of the
pathway, and transformed into the desired host, assuming that the
required vector elements are present.
Rapid Construction of Large DNA Molecules. We have made use of
yeast’s remarkable capacity to take up and recombine numerous
overlapping DNA molecules to assemble a complete genome.
Assembling large, genome-size molecules in host organisms by in
vivo recombination is not a new concept. Holt et al. (17) proposed
using the lambda Red recombination system to assemble an
et al. (18) and Yonemura et al. (19) developed methods for
set 3, was performed. Clone 11 generated all 10 predicted amplicons and thus was selected for further analysis. (B) Multiplex PCR was performed on clone 11
using all four primer sets. All 40 amplicons were efficiently generated from this clone. (C) DNA was prepared from clone 11 in agarose plugs, then digested with
EagI, as shown in Fig. 3. The linearized genome (?587 kb) was separated on a 1% agarose gel in 0.5X Tris-borate-EDTA buffer at 14 °C on a Bio-Rad Mapper XA
CHEF system. The run time was 16 h at 6 V/cm with a 50–90 s switch time ramp at an included angle of 120o. The linearized JCVI-1.9 genome fragment was cut
out of the gel as indicated by the red rectangle. (D) DNA from the gel slice in (C) was extracted and purified, then used as template for PCR at the 25 junctions
shown in Fig. 4. All 25 predicted amplicons were observed after analysis on 2% and 0.8% agarose gels.
Pieces taken up by a yeast cell
Probability of success
Data are from a computer simulation. N draws were made from a bag of 25
objects (pieces of DNA), replacing the object after each draw. The trial was a
success if all 25 objects were drawn at least once. A million trials were
conducted for each N to obtain a good approximation to the true probability
Plot of the probability that an entire set of 25 DNA pieces is taken up
www.pnas.org?cgi?doi?10.1073?pnas.0811011106Gibson et al.
of subfragments. Genome assembly in yeast, as we have described
it, is accomplished not by the addition of overlapping segments one
at a time, but rather by cotransformation of 25 different pieces at
once. Thus, large DNA molecules can be assembled much more
rapidly from synthetic or naturally occurring subfragments than
with any other system described previously. Our methods should
accelerate research projects, particularly in the emerging field of
Materials and Methods
Production of A-Series Assemblies. Eachofthe25A-seriesassemblies(Table1)is
(Epicentre). This cloning system permits induction to 10 or more copies of these
150 ml of LB plus 12.5 ?g/ml of chloramphenicol and 1X induction solution
BACs were purified using a Qiagen HiSpeed Plasmid Maxi Kit. DNA was eluted
using 500 ?l of Tris-EDTA (TE) buffer. Each insert was excised by restriction
digestion at the recommended temperature for 16 h and then terminated by
phenol-chloroform extraction and ethanol precipitation. The products were
electrophoresis with standards.
Yeast Spheroplast Transformation. In the yeast spheroplast transformation
procedure, cells are treated with zymolyase to remove the cell wall and then
made competent to take up foreign DNA by treatment with PEG and CaCl2. This
procedure was carried out using a previously published protocol with the VL6–
preparation of yeast spheroplasts. The 25 digested fragments were pooled by
adding 96 ng (0.8 ?l of 120 ng/?l) of each, then mixed with ?108yeast sphero-
plasts. After transformation, yeast spheroplasts were regenerated and selected
CSM-His agar plates as small patches and incubated for 2 days at 30 °C. After
screening and confirmation of the desired assembly, individual colonies for
analysis were obtained by restreaking these patches onto selective plates. Only
10%–20% of the restreaked colonies contained an intact genome. These were
used for further analysis.
Multiplex PCR Analysis. DNA was extracted from ?107yeast cells using an
instructions. Multiplex PCR was done using a Qiagen Multiplex PCR Kit. A 1/50
volume (2 ?l) of the DNA extract and 1 ?l of a 10X primer stock containing 20
oligos at 5 ?M each were included in each 10-?l reaction. Cycling parameters
were94 °Cfor15min,then35cyclesof94 °Cfor30s,52 °Cfor90s,and72 °Cfor
visualized using an Amersham Typhoon 9410 Fluorescence Imager.
from yeast clones in 1% agarose using a Bio-Rad Yeast DNA Plug Kit, following
the manufacturer’s instructions. After this protocol was completed, plugs were
submerged in 1X Tris-acetate-EDTA (TAE) buffer in a horizontal submarine elec-
and stored in TE buffer. Plugs were digested according to the Bio-Rad manual
England Biolabs. After incubation at 37 °C for 16 h, equal amounts of each plug
The FIGE parameters were forward: 90 V; initial switch, 5.0 s; final switch, 30 s,
with linear ramp, and reverse: 60 V; initial switch, 5.0 s; final switch, 30 s, with
linear ramp. Bands were visualized using an Amersham Typhoon 9410 Fluores-
cence Imager with 532-nm excitation and 610-nm emission wavelengths.
Analysis of Linearized Synthetic Genomes. Gel slices containing the linearized
synthetic genomes were excised, and the DNA was purified using a Qiagen
QIAquick Gel Extraction Kit according to the manufacturer’s instructions. A
kit, version 2.1, according to the manufacturer’s instructions at an annealing
for amplification were designed to be positioned outside the overlaps of the 25
A-series assemblies and to specifically generate a single product.
ACKNOWLEDGMENTS. We thank John I. Glass and Vladimir N. Noskov for
assistance. We also thank Fred Blattner for suggesting additional important
experiments that greatly improved the manuscript. This work was supported by
Synthetic Genomics, Inc.
1. Hinnen A, Hicks JB, Fink GR (1978) Transformation of yeast. Proc Natl Acad Sci U S A
2. Orr-Weaver TL, Szostak JW, Rothstein RJ (1981) Yeast transformation: A model system
for the study of recombination. Proc Natl Acad Sci U S A 78:6354–6358.
3. Ma H, Kunes S, Schatz PJ, Botstein D (1987) Plasmid construction by homologous
recombination in yeast. Gene 58:201–216.
4. Raymond CK, Pownder TA, Sexson SL (1999) General method for plasmid construction
using homologous recombination. Biotechniques 26:134–141.
5. Raymond CK, Sims EH, Olson MV (2002) Linker-mediated recombinational subcloning
of large DNA fragments using yeast. Genome Res 12:190–197.
Acad Sci U S A 92:11701–11705.
7. Ebersole T, et al. (2005) Rapid generation of long synthetic tandem repeats and its
application for analysis in human artificial chromosome formation. Nucleic Acids Res
8. Gibson DG, et al. (2008) Complete chemical synthesis, assembly, and cloning of a
Mycoplasma genitalium genome. Science 319:1215–1220.
9. Kouprina N, Larionov V (2008) Selective isolation of genomic loci from complex
genomes by transformation-associated recombination cloning in the yeast Saccharo-
myces cerevisiae. Nat Protoc 3:371–377.
10. Daley JM, Palmbos PL, Wu D, Wilson TE (2005) Nonhomologous end joining in yeast.
Annu Rev Genet 39:431–451.
11. Feller W (1957) An Introduction to Probability Theory and Its Applications (Wiley, New
York), pp 47–50.
12. Marschall P, Malik N, Larin Z (1999) Transfer of YACs up to 2.3 Mb intact into human
cells with polyethylenimine. Gene Ther 6:1634–1637.
13. Bell SP (1995) Eukaryotic replicators and associated protein complexes. Curr Opin
Genet Dev 5:162–167.
14. van Brabant AJ, Buchanan CD, Charboneau E, Fangman WL, Brewer BJ (2001) An
origin-deficient yeast artificial chromosome triggers a cell cycle checkpoint. Mol Cell
the integrity of mammalian YACs during transformation and propagation in yeast.
16. Larionov V, et al. (1994) Transformation-associated recombination between diverged
and homologous DNA repeats is induced by strand breaks. Yeast 10:93–104.
17. Holt RA, Warren R, Flibotte S, Missirlis PI, Smailus DE (2007) Rebuilding microbial
genomes. Bioessays 29:580–590.
Natl Acad Sci U S A 102:15971–15976.
19. Yonemura I, et al. (2007) Direct cloning of full-length mouse mitochondrial DNA using
a Bacillus subtilis genome vector. Gene 391:171–177.
Gibson et al.
December 23, 2008 ?
vol. 105 ?
no. 51 ?