Creating Bacterial Strains from Genomes That Have Been Cloned and Engineered in Yeast Carole Lartigue,1 Sanjay Vashee/t MikkeL A. Algire,1 Ray-Yuan Chuang,1 Gwynedd A. Benders,2 Li Ma,1 Vladimir N. Noskov,1 Evgeniya A. Denisova,1 Daniel G. Gibson,1 Nacyra Assad-Garcia,1 Nina Alperovich,1 David W. Thomas,1* Chuck Merryman,1 Clyde A. Hutchison III,2 Hamilton 0. Smith,2 ]. Craig Venter,1'2 John I. Glass1 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 (1-3) 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 "'The ]. Craig Venter Institute, 9704 Medical Center Drive, Rockville, MD 20850, USA. 2The ]. Craig Venter Institute, 10355 Science Center Drive, San Diego, CA 92121, USA. 'Present address: Biotechnology Industry Organization (BIO), 1201 Maryland Avenue SW, Washington, DC 20024, USA. |To whom correspondence should be addressed. E-mail: svashee@jcvi.org 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 thanM genitalium because of their faster growth rate. M. mycoides was transformed (7) with a vector containing a selectable tetracycline-resistance marker and a (5-galactosidase gene for screening. The vector also contained a yeast auxotrophic marker, a yeast centromere, and a yeast auton- omously replicating sequence, for selection and propagation in yeast as a yeast centromeric plasmid (YCp). Direct genomic sequencing (<5) of one clone (YCpMmycl .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 YCpMmycl.l throughout this paper, regardless of the cellular source. YCpMmycl .1 can refer to: (i) the original M. mycoides strain (the "native" M. myco ides YCpMmycl.l 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. YCpMmycl.l genomes were isolated from M. mycoides (9) and transformed into yeast sphe-roplasts (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 YCpMmycl.l in W303a. All appeared to contain complete genomes (fig. SI), which indicates that this bacterial genome is stable in yeast. Sequences between the am-plicons 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 YCpMmycl. 1 in yeast by creating a seamless deletion in a nonessential Type HI restriction endonuclease gene (Fig. 1). This modification cannot be made with the genetic tools available for this bacterium. We first transformed a YCpMmycl.l yeast clone with a cassette con- ■| typelllmod | typelllres- ||Gal1P| SCEI I URA3 I TR |||| Knock-out cassette 50 bp ii. YCpMmycl. 1-Aiype///::UR43 50 bp SD (-)His (-)Ura A |_| B in Yeast Post-transplant Fig- 1- Generation of Type III re- i. YCpMmycl .1 I 3.1 kb I Construct: i ii iii i ii iii striction enzyme deletions. (A) To make an M. mycoides Type III restriction enzyme gene (typelllres) 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 YCpMmycl.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 (Atypelllres::URA3). Galactose induction results in the expression of l-Sce I endonuclease, which cleaves the 18-bp l-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 typelllres gene (Atypelllres), 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 YCpMmycl.l clone in yeast (i) are shown for comparison. The expected sizes are obtained for each amplicon. typelllmod | |Gal1P| SCEI \ URA3 | TR Galactose induction iii. YCpMmycl A-Atypelllres P299 ♦ 5-FOA counterselection typelllmod 693 bp www.sciencemag.org SCIENCE VOL 325 25 SEPTEMBER 2009 1 taining 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 IE gene. Four of five clones contained intact genomes, and one contained a genome with a large deletion (YCpMmycl.l-A500kb) (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 IE restriction enzyme gene (Fig. 1 A). The changes to the genome were verified by PCR (Fig. IB). We isolated YCpMmycl.l 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 YCpMmycl.l donor DNA isolated from yeast (fig. S4). 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 YCpMmycl.l fiomM mycoides and transplanted it into wild-type M. capricolum andM 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 YCpMmycl.l 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 orM 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 YCpMmycl.l 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. YCpMmyc 1.1, as well as the engineered YCp genomes (YCpMmycl.l-AtypeIIIres::URA3 and YCpMmycl.l-Atypelllres), 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 (YCpMmycl.1-A500kb) 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. mycoWas-specific IS 1296 element as probe (Fig. 2A). We showed that the Type IE restriction gene was deleted in the engineered bacterium by PCR (Fig. IB), by Southern blot analysis using the Type IE restriction gene sequence as probe (Fig. 2B), and by sequencing the locus (Fig. 2C). To confirm that our transplants were entirely M. mycoides and not chimeras containing yeast sequences orM capricolum recipient cell sequences, we sequenced the genome of one transplant from which the Type IE 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 YCpMmyc 1.1-Atypelllres genome sequence was identical to M. mycoides YCpMmycl.l, 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. 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. SI. For untreated samples, yeast plugs were digested with (5-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 YCpMmycl.l. W303a yeast agarose plugs carried YCpMmycl.l, YCpMmycl.l that was engineered in yeast (YCpMmycl.l-Atypelllres::URA3 or YCpMmycl.l-zlfj/pe///res), or YCpMmycl.l-A500kb. The number of transplants is the average of at least three experiments. The error reported is the absolute mean deviation. Yeast strain Genome Methylation treatment Number of transplants (colonies or plugs) M. capricolum RE(-) Wild-type M. capricolum VL6-48N YCpMmycl.l Untreated 37 ± 3 0 M. capricolum 32 ± 13 9 ± 4 extracts M. mycoides 15 ± 8 22 ± 8 [13 ± 4]* extracts [10 ± 4]t Mock-methylated 34 ± 17 0 M. mycoides 20 ± 17 13 ± 10 purified methylases W303a YCpMmycl.l Untreated 22 ± 5 Not done YCpMmycl.l- Untreated 52 ± 10 Not done Atypelllres::URA3 YCpMmycl.l-Atypelllres Untreated 52 ± 12 Not done YCpMmycl.l-A500kb Untreated 0 Not done *Yeast plugs were cleared of yeast genomic DNA by digestion with a cocktail of Asi 51, Rsr II, and Fse I followed by pulsed-field gel electrophoresis. fYeast plugs were cleared of yeast genomic DNA by using pulsed-field gel electrophoresis. 694 25 SEPTEMBER 2009 VOL 325 SCIENCE www.sciencemag.org 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 indepen- dent 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. 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 theM 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 \ \ a- a- B \ \ \ C Sequence of typelllres deletion aggaaggctgtgtgaatttcattcgacaagataaaggaaa ttagaacatttactttagttacmataatgaaaataatat agggattgatgtttgctatgaacgcttgtttagaataaat aatggtaaatcaactgataataaaacagattttaaatgaa tagaaaaaaacgaaccctatctgtctaatttagatgtttt tgatattaaatactattcaactaaattgtttgaagataat agtgctaatgaaataatcaaaaaacaattcataaaaatgt taaatgatctaaaaataaatgctgatgatatttctacaat taaaactttaagacaattaactgcactaaaaccaatctca ggtggacaaaacjjatg|agattaac|raa|raaacaagaattt gtagtgcaagatattgttaaagaagattaiJtaaKttgtta aaataaaaaattacttatttaagcgggactatttctattt tatttaaagaagtcttatgattgaatgtaaaagttctaat gtattaacccctgttacattaaatacaaccatatctattt ttaattaatctcctgaaagtacaagttcttaattcctaac catgcctattaaatttaagatcaagagcagtttttctaac ccccgaaaaaaaactcacaacaaaaa Fig. 2. Southern blot analysis of M. mycoides transplants. (A) Hind Ill-restricted genomic DNA from representative YCpMmyd.l, YCpMmycl.l-Atypelllres::URA3 (next-to-last columns), or YCpMmyd.l-zlfype///res (last columns) M. mycoides genome transplants from yeast was probed with IS1296 sequences. The genomic DNAs of native M. mycoides YCpMmyd.l 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 YCpMmyd.l. (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 YCpMmyd.1-Atypelllres::URA3, and YCpMmyd.l-zlfype///res M mycoides cells derived by genome transplantation was probed with the typelllres gene sequence. The genomic DNAs of native M. mycoides YCpMmyd.1 cells and M. capricolum recipient cells were also probed for comparison. The typelllres gene is absent in the engineered genomes but present in the native M mycoides YCpMmyd.1 genome. (C) Sequencing of the seamless deletion region of the YCpMmyd.l-zlfype///res 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 typelllres gene remained after the deletion because of the overlap between the typelllmod and typelllres genes. The start and stop codons of the typelllres gene are boxed in red; the stop codon of the typelllmod gene is boxed in black. Native Transplants Native Transplants 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 systemfc). This cycle can be repeated starting from the newly engineered genome (dashed arrow). Insertion of yeast vector into bacterial genome Isolation Genome with a yeast vector CO Bacterial genome Methylation doned (if necessary) Isolation lci idi yer lur i le cloned a in yeast o \S« V*' M enc Ge nome engineering www.sciencemag.org SCIENCE VOL 325 25 SEPTEMBER 2009 1 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. References and Notes 1. A. ]. DaMassa, D. L. Brooks, H. E. Adler, Am. ]. Vet. Res. 44, 322 (1983). 2. L. Manso-Silvan et ai, Int. ]. Syst. Evol. Microbiol. 59, 1353 (2009). 3. Manso-Silvan et ai. (2) report the very recent renaming of Mycoplasma mycoides subspecies mycoides Large Colony to Mycoplasma mycoides subspecies capri. Power law statistics are a hallmark of critical phenomena. A less obvious characteristic of criticality is the emergence of universality classes that capture the similarity of seemingly disparate systems. For example, despite the fact that water and carbon dioxide have different chemical properties, they were observed to behave in the same manner when close to their respective critical points (1). This is because idiosyncrasies, such as the existence of electric dipoles or the ability to form hydrogen bonds, become irrelevant near the liquid/gas critical point. For physical systems, renormalization group theory (2, 3) has enabled researchers to understand the deep connection between the symmetries of a system and the mech- 1Department of Chemical and Biological Engineering, Northwestern University, Evanston, IL 60208, USA. 2Datascope Analytics, Evanston, IL 60201, USA. integrative Ecology Group, Estacidn Bioldgica de Donana, CSIC, 41092 Sevilla, Spain, "instituto Nacional de Pesquisas Espaciais, 12227-010 Sao ]ose' dos Campos, Sao Paulo, Brazil. 5Howard Hughes Medical Institute, Northwestern University, Evanston, IL 60208, USA. 6Northwestern Institute on Complex Systems, Northwestern University, Evanston, IL 60208, USA. *To whom correspondence should be addressed. E-mail: dean.malmgren@u.northwestern.edu (R.D.M.); amaral@ northwestern.edu (L.A.N.A.) 4. C. Lartigue et ai, Science 317, 632 (2007). 5. D. G. Gibson et ai, Science 319, 1215 (2008). 6. D. G. Gibson et ai, Proc. Natl. Acad. Sci. U.S.A. 105, 20404 (2008). 7. K. W. King, K. Dybvig, Plasmid 26, 108 (1991). 8. ]. I. Glass et ai, Proc. Natl. Acad. Sci. U.S.A. 103, 425 (2006). 9. Materials and methods are available as supporting material on Science Online. 10. N. Kouprina, V. Larionov, Nat. Protocols 3, 371 (2008). 11. V. Larionov, N. Kouprina, G. Solomon, ]. C. Barrett, M. A. Resnick, Proc. Natl. Acad. Sci. U.S.A. 94, 7384 (1997). 12. ]. D. Boeke, F. LaCroute, G. R. Fink, Mol. Gen. Genet. 197, 345 (1984). 13. C. Janis et ai, Appl. Environ. Microbiol 71, 2888 (2005). 14. L. Dedieu-Engelmann, Comp. Immunol. Microbiol. Inject. Dis. 31, 227 (2008). 15. F. Thiaucourt et ai, Dev. Biol. (Basel) 114, 147 (2003). 16. We thank R. Roberts for helpful information and the Synthetic Biology team at the ]. Craig Venter anisms that underlie its behavior. The similarity of different fluids near their respective liquid/gas critical points is often demonstrated by rescaling their statistics so that they collapse onto the same universal curves (often power law curves),which have particular scaling exponents. By grouping different substances into the same universality class, as identified by its scaling exponents, one discovers that fluids are described by the same statistical laws near the liquid/gas critical point as uniaxial magnets are near their paramagnetic critical point (1). One can also differentiate the behavior of these systems from the behavior of polymers near the sol/gel transition, which belong to a different universality class (1). In addition to describing critical phenomena, power law scaling has also been widely reported in biology, economics, and sociology (4-10). Re-normalization group theory therefore offers a tantalizing hypothesis for the prevalence of particular power law scaling exponents in social systems: Social systems, in analogy with physical systems, may operate near critical points and can therefore be classified into a small number of distinct universality classes. A heated debate has consequently ensued in the literature concerning the "universal- 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. Supporting Online Material www.sciencemag.org/cgi/content/full/1173759/DCl Materials and Methods Figs. SI to S7 Table SI References 18 March 2009; accepted 13 August 2009 Published online 20 August 2009; 10.1126/science.ll73759 Include this information when citing this paper. ity of human systems" (in the statistical physics meaning of the phrase). Is there enough statistical evidence for the asymptotic power law description of the heavy-tailed distributions reported in human systems (11-14)1 Is it reasonable to postulate that social systems, like their physical counterparts (2, 3,15), can be classified into universality classes according to scaling exponents (16)1 Human correspondence is a paradigmatic area where the matters of power law scaling and universality are contentious issues. One view that has recently received considerable attention in the literature (17, 18) posits that correspondence patterns are driven primarily by the need to respond to other individuals. This is formalized by a priority queuing model (19), which, under certain limiting conditions, reproduces the asymptotic scaling of empirically observed heavy-tailed correspondence statistics. In particular, the heavy-tailed statistical properties of e-mail correspondence are reportedly reproduced by a fixed-length queue with a single task type (19,20), whereas the heavy-tailed statistical properties of letter correspondence are reportedly reproduced by either a variable-length queue with a single task type (21, 20) or by a fixed-length queue with multiple task types (22). The fact that there are different exponents for the two modes of correspondence has been taken as evidence that human correspondence falls into one of two universality classes (20). When interpreted in the statistical mechanics sense of universality, one would conclude that e-mail and letter correspondence are fundamentally different activities. In contrast, we hypothesize that human correspondence patterns are not driven by responses to others but by more prosaic mechanisms: the circa-dian cycle, task repetition, and changing communication needs. We formalize these mechanisms with a cascading, nonhomogeneous Poisson process that we have previously shown to be statistically consistent with e-mail communication patterns (14). We hypothesize that the same model is capable of describing letter correspondence and that the heavy-tailed correspondence statistics primarily arise from the variation in an individual's communication needs over the course of his or her lifetime. On Universality in Human Correspondence Activity R. Dean Malmgren,1,2* Daniel B. Stouffer,1,3 Andriana S. L. O. Campanharo,1,4 Luis A. Nunes Amaral1'5'6* The identification and modeling of patterns of human activity have important ramifications for applications ranging from predicting disease spread to optimizing resource allocation. Because of its relevance and availability, written correspondence provides a powerful proxy for studying human activity. One school of thought is that human correspondence is driven by responses to received correspondence, a view that requires a distinct response mechanism to explain e-mail and letter correspondence observations. We demonstrate that, like e-mail correspondence, the letter correspondence patterns of 16 writers, performers, politicians, and scientists are well described by the circadian cycle, task repetition, and changing communication needs. We confirm the universality of these mechanisms by rescaling letter and e-mail correspondence statistics to reveal their underlying similarity. 696 25 SEPTEMBER 2009 VOL 325 SCIENCE www.sciencemag.org Science Creating Bacterial Strains from Genomes That Have Been Cloned and Engineered in Yeast Carole Lartigue, Sanjay Vashee, Mikkel A. Algire, Ray-Yuan Chuang, Gwynedd A. Benders, Li Ma, Vladimir N. Noskov, Evgeniya A. Denisova, Daniel G. Gibson, Nacyra Assad-Garcia, Nina Alperovich, David W. Thomas, Chuck Merryman, Clyde A. Hutchison III, Hamilton O. Smith, J. Craig Venter and John I. Glass Science 325 (5948), 1693-1696. DOI: 10.1126/science.1173759originally published online August 20, 2009 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 etal. (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. caphcolum to produce a bacterial cell with the genotype and phenotype of the altered M. mycoides large-colony genome. article tools http://science.sciencemag.org/content/325/5948/1693 MATEmALSNTARY http://science.sciencemag.org/content/suppl/2009/08/24/1173759.DC1 content http://science.sciencemag.org/content/sci/325/5943/928.full references This article cites 13 articles, 6 of which you can access for free http://science.sciencemag.Org/content/325/5948/1693#BIBL permissions http://www.sciencemag.org/help/rephnts-and-permissions Use of this article is subject to the Terms of Service Science (print ISSN 0036-8075; online ISSN 1095-9203) is published by the American Association for the Advancement of Science, 1200 New York Avenue NW, Washington, DC 20005. The title Science is a registered trademark of AAAS. Copyright © 2009, American Association for the Advancement of Science