npg
© 2014 Nature America, Inc. All rights reserved.
news and v iews hybridization technology allows simultaneous measurement of up to 800 genes in 12 cells per run6. However, none of these approaches provides TCR sequence information. Efforts to understand the role of T cells in health and disease will greatly benefit from both TCR sequence and phenotypic information for each T-cell clone. When a T cell is activated by an antigen, it divides and proliferates. The daughter cells all share the same TCR with their parent. Thus, the TCR sequence profile is important for identifying antigen specificities and the proliferation history of a TCR lineage. Phenotypic information is crucial for defining the effects of T cells on disease initiation and development, which has diagnostic and therapeutic value. Current approaches for obtaining both TCR sequence and phenotype information are inadequate because of limits on the numbers of markers that can be monitored. For example, single-cell TCR sequencing coupled with fluorescence-activated cell sorting (FACS) is limited to ~13 markers owing to fluorescence spectral overlapping. As shown in B cells, a larger number of markers can be analyzed by using microfluidic chip–based qPCR in combination with Sanger sequencing to determine immune receptor sequences 7. However, this method requires the use of specialized devices. In addition, Sanger sequencing is limited in its ability to resolve the second productive light (in B-cell) or α (in T-cell) chains. The method developed by Han et al.2 requires only common laboratory equipment and access to a next-generation sequencer, streamlining the correlated measurement of TCR sequences and phenotypic markers in thousands of cells. Using a nested PCR approach, the authors obtained the expression level for a set of 17 phenotypic markers along with TCR α- and β-chain sequences in over 80% of the cells in a population, the highest value reported so far2. Although their method is not designed to quantify transcript abundance, the presence or absence of a given transcript matched nicely with the protein secretion measured using FACS for a few cytokines. Having validated their method in peripheral blood mononuclear cells and a T-cell line, the authors went on to study tumor-infiltrating lymphocytes. CD4+ T cells were isolated from the colon tissue of a patient with colorectal cancer. Significant clonal expansion of T cells bearing the same TCR was observed in cells isolated from tumor tissue compared to those obtained from adjacent normal tissue. Interestingly, even within populations with identical TCRs, substantial heterogeneity in the expression of the 17 marker genes was 640
observed, underscoring the importance of single-cell analysis2. In colorectal carcinoma, there is considerable interest in the therapeutic and prognostic value of T helper 17 (Th17) cells because these cells have an inflammatory anti-microbial role that promotes cancer growth8. However, little is known about the lineage of tumorinfiltrating Th17 cells. By comparing the TCR sequences shared between different subpopulations of these cells, Han et al.2 found that FOXP3+RORC+IL17+ T cells and FOXP3–RORC+IL17+ T cells share a common ancestor. This suggests that the landscape of tumor-infiltrating lymphocytes is very complicated when both T-cell lineage and phenotype are considered, and that determining the origin of Th17 cells might reveal avenues for new therapies. Future research should use a larger patient cohort at different stages of tumor progression. T-cell biology holds many questions to be addressed in the future. What antigens do T cells bind to? What are the affinities of polyclonal TCRs, and how does TCR affinity affect T-cell development and differentiation? What is the regulatory network that controls T-cell development and differentiation?
T cells are implicated in many diseases such as cancers, infections and autoimmune diseases. Possessing correlated information on the TCR, its ligands and the phenotype of the T cell will likely be important in the diagnosis, treatment and prevention of a wide range of conditions. As high-throughput technologies for the analysis of antigen binding9, single-cell transcriptomes10 and other cellular attributes become available, the T-cell ID card will only become more informative. COMPETING FINANCIAL INTERESTS The authors declare no competing financial interests. 1. Arstila, T.P. et al. Science 286, 958–961 (1999). 2. Han, A., Glanville, J., Hansmann, L. & Davis, M.M. Nat. Biotechnol. 32, 685–693 (2014). 3. Davis, M.M. & Bjorkman, P.J. Nature 334, 395–402 (1988). 4. Turchaninova, M.A. et al. Eur. J. Immunol. 43, 2507–2515 (2013). 5. Bendall, S.C. & Nolan, G.P. Nat. Biotechnol. 30, 639–647 (2012). 6. Yosef, N. et al. Nature 496, 461–468 (2013). 7. Weinstein, J.A., Zeng, X., Chien, Y.H. & Quake, S.R. PLoS ONE 8, e67624 (2013). 8. Ladoire, S., Martin, F. & Ghiringhelli, F. Cancer Immunol. Immunother. 60, 909–918 (2011). 9. Birnbaum, M.E. et al. Cell 157, 1073–1087 (2014). 10. Shapiro, E., Biezuner, T. & Linnarsson, S. Nat. Rev. Genet. 14, 618–630 (2013).
A genealogy of the citrus family Riccardo Velasco & Concetta Licciardello Clarification of the genetic relationships among species opens new possibilities for enhancing citrus diversity and disease resistance. The family relationships among citrus fruits— which include sweet and sour oranges, mandarins, clementines, tangerines, grapefruits, pummelos, kumquats, lemons and limes—are famously obscure. Thousands of years of citrus cultivation and interbreeding have yielded some 25 species and at least 250 commercial varieties. Following on the recently published draft genome of sweet orange1, Wu et al.2 report in this issue the genome sequences of clementines, mandarins, pummelos, sweet oranges and sour oranges. By combining these data with some elegant sleuth work, they also make important progress in elucidating the Riccardo Velasco is at Fondazione Edmund Mach, di San Michele a/Adige, Trento, Italy. Concetta Licciardello is at the Consiglio di Ricerca e Sperimentazione in Agricoltura, CRA-ACM, Acireale, Italy. e-mail: [email protected]
phylogenetic history of citrus domestication. Given the dependence of the $9 billion/year global citrus industry on large-scale monoculture, the findings provide much-needed opportunities for de novo breeding of citrus varieties with increased resistance to pathogens and environmental extremes. Breeding guided by genomic information will also provide a greater diversity of fruit for consumers in terms of new combinations of taste, aroma, size, color, shape and ease of peeling. Wu et al.2 began prudently by choosing to assemble the 301-Mb genome of a clementine with a haploid genome to sidestep the complications of assembling outbred diploid genomes3. Whereas the 367-Mb sweet-orange genome, assembled using Illumina paired-end tags, was only 76% complete1, the haploid genome of the clementine assembled using Sanger sequencing is 96% complete. The clementine genome has an L50 value more than twice that of the sweet-orange genome (119 versus 49.9 Kb)
volume 32 number 7 JUly 2014 nature biotechnology
npg
and its L50 scaffolds are on average about four times as large (6.8 versus 1.69 Mb). The L50, which is analogous to the median scaffold length, provides a measure of the quality of genome assembly. To begin to examine taxonomic relationships between wild and domesticated species, Wu et al.2 next assembled genome sequences of an additional eight diploid heterozygote accessions by mapping sequences generated using Illumina technology (ranging from 17× to 110× coverage) to the haploid clementine reference genome. These were selected to represent a broad spectrum of genetic diversity across citrus species and to evaluate the adequacy of phylogenetic relationships and pedigrees that had been developed using traditional taxonomic criteria. The accessions chosen were four mandarins (Clementine mandarin, Ponkan mandarin, Willowleaf mandarin and W. Murcott mandarin), two pummelos (Chandler pummelo and low-acid pummelo), a sweet orange and a sour orange (Fig. 1). The authors then analyzed phylogenetic relationships between the nine accessions by generating neighbor-joining phylogenetic trees of the chloroplast genomes that define the two clades distinguished by maternal origin (Citrus maxima or Citrus reticulata) and by principal coordinate analysis of the nuclear genomes based on pairwise distances. All nine accessions are believed to be progeny of C. reticulata Blanco and C. maxima (Fig. 1). A more distantly related progenitor, Citrus medica, was not included in the analysis (Fig. 1). The authors’ comparative analysis helps to address several controversial hypotheses about the phylogenetic relationships between members of the citrus family and to map the evolution of citrus genomes from a paleohexaploid eudicotyledonous ancestor4. Although the analysis does support ancestral roles for C. reticulata and C. maxima, it calls into question the widely held assumption that traditional mandarins, which comprise a mixture of different species of ancient and recent hybrids5, are pure C. reticulata genotypes. Instead, Ponkan and Willowleaf mandarins are shown to be the product of considerable admixture between C. reticulata and C. maxima (Fig. 1). Another key insight that emerges from the authors’ analysis is that, despite many similar features, sweet orange and sour orange arose by very different breeding processes. Whereas sour orange is a pure F1 hybrid between C. maxima (egg donor) and C. reticulata (pollen donor) genotypes, sweet orange is a more complex admixture of C. maxima and C. reticulata genotypes. Finally, the new data negate the proposal that sweet orange might be derived
C. reticulata Blanco Mandarin
C. × sinensis • Sweet orange
C. maxima • Chandler pummelo • Low-acid pummelo
C. × aurantium • Sour orange
C. clementina • Haploid clementine (reference genome) • Clementine mandarin
C. medica Citron
C. × deliciosa • Willowleaf mandarin
C. paradisi Grapefruit
(C. reticulata) • Ponkan mandarin • W. Murcott mandarin
C. limon Lemon
C. micrantha
Kim Caesar/Nature Publishing Group
© 2014 Nature America, Inc. All rights reserved.
news and v iews
C. aurantifolia Mexican lime
Figure 1 The origin and evolution of select citrus species. The three major ancestors of citrus species included in the top row—Citrus reticulata, Citrus maxima and Citrus medica—contributed to the origins of all currently cultivated citrus species. Based on their high-quality genome sequence of a haploid clementine and resequencing of eight diploid species (indicated with bullet points against a yellow background), Wu et al.2 propose taxonomic relationships between members of the citrus family. Species not highlighted in yellow were not included in the current analysis. Lines with arrowheads indicate contributions to hybrids; lines without arrowheads represent simple introgression. The use of parentheses to denote Ponkan mandarin and W. Murcott mandarin as C. reticulata indicates that although both are almost pure C. reticulata, both resulted from a low level of introgression with C. maxima.
from a simple backcross between pummelo and mandarin ((C. maxima × C. reticulata) × C. reticulata)1. The large haplotypic block found in chromosome 2 of the two genotypes could be consistent with this proposal only if both putative parents had some pummelo ancestry. Instead, the authors show that sweet orange could be derived only from a cross between (C. maxima × C. reticulata) × C. maxima as egg donor and a male C. reticulata, with some introgression with C. maxima. Knowledge of the maternal and paternal donors that gave rise to the diverse range of citrus species will make it possible to implement an entirely new strategy for developing useful citrus varieties through breeding. By recapitulating step by step the domestication process used long ago to generate a particular variety, breeders should be able to screen progeny for subtle differences in their genetic composition. Inclusion of those lines in subsequent rounds of crossing or admixture should accelerate the rate with which breeders can emulate desirable traits while introducing genetic diversity elsewhere in the genome. Whereas C. maxima was the maternal donor in the
nature biotechnology volume 32 number 7 JUly 2014
lineage(s) that gave rise to pummelos, sweet oranges and sour oranges, C. reticulata was the maternal donor for all of the mandarins and the clementine variety studied (Fig. 1). Given that many commercially desirable varieties appear to have emerged after only a few generations of interbreeding between C. maxima and C. reticulata, it seems likely that sequencing-driven breeding using haplotype information should rapidly achieve other beneficial interspecies admixtures in a more targeted and rational way. The data of Wu et al.2 include >1.5 million fixed single-nucleotides that distinguish C. maxima from C. reticulata and define large haplotypic blocks characteristic of the genomes of mandarins, pummelos and oranges. These data will enable breeders to screen progeny for those most likely to improve commercial and agricultural traits while introducing genetic diversity. The availability of haplotype blocks also offers the opportunity to analyze the huge diversity of citrus germplasm in China6 to establish whether pure C. reticulata samples still exist, or to reveal the existence of additional novel species. New interspecies 641
news and v iews combinations between C. maxima, C. reticulata or other yet-to-be-identified wild species might provide superior varieties resistant to abiotic and biotic stresses7, such as that caused by the bacterium responsible for the citrus greening or Huanglongbing disease8, which is decimating orchards across the world and threatening the global citrus industry. Finally, the work of Wu et al.2 will lead to a better understanding of the genetic basis of the extraordinary diversity in the colors, flavors, sizes and aromas of citrus fruits and whether these might be engineered in novel varieties.
COMPETING FINANCIAL INTERESTS The authors declare no competing financial interests. 1. Xu, Q. et al. Nat. Genet. 45, 59–66 (2013). 2. Wu, G.A. et al. Nat. Biotechnol. 32, 656–662 (2014). 3. Velasco, R. et al. Nat. Genet. 42, 833–839 (2010). 4. Salse, J. Curr. Opin. Plant Biol. 15, 122–130 (2012). 5. Garcia-Lor, A. et al. Ann. Bot. (Lond.) 111, 1–19 (2013). 6. Gmitter, F.G. & Hu, X. Econ. Bot. 44, 267–277 (1990). 7. Bolger, M.E. et al. Curr. Opin. Biotechnol. 26, 31–37 (2014). 8. Gottwald, T.R. Annu. Rev. Phytopathol. 48, 119–139 (2010).
npg
© 2014 Nature America, Inc. All rights reserved.
Next-generation wearable electronics Michael J Cima New fabric-like sensors measure and transmit mechanical strain on the skin with unprecedented sensitivity. Motion, an act we take for granted until it is stolen by disease, age or injury, can now be measured by a growing array of portable electronic devices for consumer and medical use. Whether for fun, fitness or medicine, these simple wearable devices rely on tiny accel erometers that capture only the gross motion of a limb or entire body. Writing in Nature Nanotechnology, Son et al.1 describe a new class of integrated electronic sensors and memory elements packaged in fabric-like material that can record mechanical strains when applied to the skin, offering unprecedented fidelity for analyzing human motion. Devices built on this technology could find broad uses in medicine ranging from diagnostic tools to mechanically assisted prostheses. Portable electronics, originally devised for communication and entertainment, are expanding rapidly into the area of health metrics as they become more ‘wearable’ through form factors such as wristwatches, bracelets and eyewear. For example, Fitbit and similar products are worn like a watch and contain accelerometers or other sensors for tracking Michael J. Cima is David H. Koch Professor of Engineering, Faculty Director of the Lemelson-MIT Program, Department of Materials Science and Engineering, Koch Institute for Integrative Cancer Research, Cambridge Massachusetts, USA. e-mail: [email protected]
642
movement and recording vital signs. The clinical equivalent, such as systems by ActiGraph, can securely manage information across cohorts of study subjects or patients. These products presage a future in which measurements from wearable devices form the basis of clinical decisions. The transition to electronic health management builds on a long history of medical electronics. The first fully implanted pacemakers were developed in the 1960s, and since then similar devices have been developed for a large number of indications. Implanted devices for cardiac defibrillation, cardiac resynchronization, peripheral and vagus nerve stimulation and deep brain stimulation have saved many thousands of lives. All of these products were created by packaging existing electronic and microelectromechanical systems in ways appropriate to the application. An implanted device, for example, must have a package that is biocompatible and hermetic. This often takes the form of a welded titanium enclosure. Such devices are not truly integrated with tissue as they have completely different mechanical properties. The article by Son et al.1 is representative of a new trend in medical electronics: the design of discrete electronic and microelectromechanical systems devices for clinical measurement and actuation, embedded within packaging that can be worn directly on skin or contained in clothing. Such devices have been difficult
to develop because they must conform to the contours of the skin while also accommodating the large strains experienced during physical activity. Son et al.1 succeeded by incorporating device elements within an elastic matrix. The authors integrated mechanical strain sensors with arrays of data-storage elements, called resistive random access memory, into the elastic matrix (Fig. 1). When applied to the skin, the device’s onboard sensors record the strains that it experiences when the skin moves. These data are stored locally in the resistive random access memory and read at a later time. Several other groups are pursuing similar approaches, for example, multifunctional epidermal electronics2 from the group of John Rogers. Continuous strain measurements using the devices of Son et al.1 provide new opportunities for analyzing human motion. Mobility is a strong predictor of health outcomes in older adults3, and people who lose proficiency in balance and mobility are at increased risk of morbidity and mortality. Simple wearable accelerometers that measure gross motion of a limb or entire body have already been adopted in clinical settings. Indeed, a recent article in the Journal of the American Medical Association describes the use of accelerometer data as an outcome measure in evaluating structured physical activity to prevent mobility impairment in older adults4. Continuously monitored motion data generated by arrays of strain sensors1 could provide diagnostic information on the complicated interplay between individual muscles needed for motion and balance. Another application of strain-sensitive arrays would be to control input signals for exoskeleton assist devices—externally worn robotic hardware that assists human motion. The popular press depicts these technologies as part of the future, but they are already a reality in rehabilitation training for patients with stroke and other conditons5. Although chronic mechanical assist prostheses have improved substantially in recent years, the control systems integral to their operation6 are still inadequate. The primary control interface is surface electromyography, in which electrodes on the surface of the skin measure voltage signals within the depolarization zone of skeletal muscles. However, the relationships between the magnitude of the signals, the force exerted by a prosthetic and the displacement of the limb are difficult to determine. Strain-sensitive arrays could easily be integrated with surface electromyography to provide the high-quality measurements needed to more effectively control a prosthesis.
volume 32 number 7 JUly 2014 nature biotechnology
© 2014 Nature America, Inc. All rights reserved.
news and v iews hybridization technology allows simultaneous measurement of up to 800 genes in 12 cells per run6. However, none of these approaches provides TCR sequence information. Efforts to understand the role of T cells in health and disease will greatly benefit from both TCR sequence and phenotypic information for each T-cell clone. When a T cell is activated by an antigen, it divides and proliferates. The daughter cells all share the same TCR with their parent. Thus, the TCR sequence profile is important for identifying antigen specificities and the proliferation history of a TCR lineage. Phenotypic information is crucial for defining the effects of T cells on disease initiation and development, which has diagnostic and therapeutic value. Current approaches for obtaining both TCR sequence and phenotype information are inadequate because of limits on the numbers of markers that can be monitored. For example, single-cell TCR sequencing coupled with fluorescence-activated cell sorting (FACS) is limited to ~13 markers owing to fluorescence spectral overlapping. As shown in B cells, a larger number of markers can be analyzed by using microfluidic chip–based qPCR in combination with Sanger sequencing to determine immune receptor sequences 7. However, this method requires the use of specialized devices. In addition, Sanger sequencing is limited in its ability to resolve the second productive light (in B-cell) or α (in T-cell) chains. The method developed by Han et al.2 requires only common laboratory equipment and access to a next-generation sequencer, streamlining the correlated measurement of TCR sequences and phenotypic markers in thousands of cells. Using a nested PCR approach, the authors obtained the expression level for a set of 17 phenotypic markers along with TCR α- and β-chain sequences in over 80% of the cells in a population, the highest value reported so far2. Although their method is not designed to quantify transcript abundance, the presence or absence of a given transcript matched nicely with the protein secretion measured using FACS for a few cytokines. Having validated their method in peripheral blood mononuclear cells and a T-cell line, the authors went on to study tumor-infiltrating lymphocytes. CD4+ T cells were isolated from the colon tissue of a patient with colorectal cancer. Significant clonal expansion of T cells bearing the same TCR was observed in cells isolated from tumor tissue compared to those obtained from adjacent normal tissue. Interestingly, even within populations with identical TCRs, substantial heterogeneity in the expression of the 17 marker genes was 640
observed, underscoring the importance of single-cell analysis2. In colorectal carcinoma, there is considerable interest in the therapeutic and prognostic value of T helper 17 (Th17) cells because these cells have an inflammatory anti-microbial role that promotes cancer growth8. However, little is known about the lineage of tumorinfiltrating Th17 cells. By comparing the TCR sequences shared between different subpopulations of these cells, Han et al.2 found that FOXP3+RORC+IL17+ T cells and FOXP3–RORC+IL17+ T cells share a common ancestor. This suggests that the landscape of tumor-infiltrating lymphocytes is very complicated when both T-cell lineage and phenotype are considered, and that determining the origin of Th17 cells might reveal avenues for new therapies. Future research should use a larger patient cohort at different stages of tumor progression. T-cell biology holds many questions to be addressed in the future. What antigens do T cells bind to? What are the affinities of polyclonal TCRs, and how does TCR affinity affect T-cell development and differentiation? What is the regulatory network that controls T-cell development and differentiation?
T cells are implicated in many diseases such as cancers, infections and autoimmune diseases. Possessing correlated information on the TCR, its ligands and the phenotype of the T cell will likely be important in the diagnosis, treatment and prevention of a wide range of conditions. As high-throughput technologies for the analysis of antigen binding9, single-cell transcriptomes10 and other cellular attributes become available, the T-cell ID card will only become more informative. COMPETING FINANCIAL INTERESTS The authors declare no competing financial interests. 1. Arstila, T.P. et al. Science 286, 958–961 (1999). 2. Han, A., Glanville, J., Hansmann, L. & Davis, M.M. Nat. Biotechnol. 32, 685–693 (2014). 3. Davis, M.M. & Bjorkman, P.J. Nature 334, 395–402 (1988). 4. Turchaninova, M.A. et al. Eur. J. Immunol. 43, 2507–2515 (2013). 5. Bendall, S.C. & Nolan, G.P. Nat. Biotechnol. 30, 639–647 (2012). 6. Yosef, N. et al. Nature 496, 461–468 (2013). 7. Weinstein, J.A., Zeng, X., Chien, Y.H. & Quake, S.R. PLoS ONE 8, e67624 (2013). 8. Ladoire, S., Martin, F. & Ghiringhelli, F. Cancer Immunol. Immunother. 60, 909–918 (2011). 9. Birnbaum, M.E. et al. Cell 157, 1073–1087 (2014). 10. Shapiro, E., Biezuner, T. & Linnarsson, S. Nat. Rev. Genet. 14, 618–630 (2013).
A genealogy of the citrus family Riccardo Velasco & Concetta Licciardello Clarification of the genetic relationships among species opens new possibilities for enhancing citrus diversity and disease resistance. The family relationships among citrus fruits— which include sweet and sour oranges, mandarins, clementines, tangerines, grapefruits, pummelos, kumquats, lemons and limes—are famously obscure. Thousands of years of citrus cultivation and interbreeding have yielded some 25 species and at least 250 commercial varieties. Following on the recently published draft genome of sweet orange1, Wu et al.2 report in this issue the genome sequences of clementines, mandarins, pummelos, sweet oranges and sour oranges. By combining these data with some elegant sleuth work, they also make important progress in elucidating the Riccardo Velasco is at Fondazione Edmund Mach, di San Michele a/Adige, Trento, Italy. Concetta Licciardello is at the Consiglio di Ricerca e Sperimentazione in Agricoltura, CRA-ACM, Acireale, Italy. e-mail: [email protected]
phylogenetic history of citrus domestication. Given the dependence of the $9 billion/year global citrus industry on large-scale monoculture, the findings provide much-needed opportunities for de novo breeding of citrus varieties with increased resistance to pathogens and environmental extremes. Breeding guided by genomic information will also provide a greater diversity of fruit for consumers in terms of new combinations of taste, aroma, size, color, shape and ease of peeling. Wu et al.2 began prudently by choosing to assemble the 301-Mb genome of a clementine with a haploid genome to sidestep the complications of assembling outbred diploid genomes3. Whereas the 367-Mb sweet-orange genome, assembled using Illumina paired-end tags, was only 76% complete1, the haploid genome of the clementine assembled using Sanger sequencing is 96% complete. The clementine genome has an L50 value more than twice that of the sweet-orange genome (119 versus 49.9 Kb)
volume 32 number 7 JUly 2014 nature biotechnology
npg
and its L50 scaffolds are on average about four times as large (6.8 versus 1.69 Mb). The L50, which is analogous to the median scaffold length, provides a measure of the quality of genome assembly. To begin to examine taxonomic relationships between wild and domesticated species, Wu et al.2 next assembled genome sequences of an additional eight diploid heterozygote accessions by mapping sequences generated using Illumina technology (ranging from 17× to 110× coverage) to the haploid clementine reference genome. These were selected to represent a broad spectrum of genetic diversity across citrus species and to evaluate the adequacy of phylogenetic relationships and pedigrees that had been developed using traditional taxonomic criteria. The accessions chosen were four mandarins (Clementine mandarin, Ponkan mandarin, Willowleaf mandarin and W. Murcott mandarin), two pummelos (Chandler pummelo and low-acid pummelo), a sweet orange and a sour orange (Fig. 1). The authors then analyzed phylogenetic relationships between the nine accessions by generating neighbor-joining phylogenetic trees of the chloroplast genomes that define the two clades distinguished by maternal origin (Citrus maxima or Citrus reticulata) and by principal coordinate analysis of the nuclear genomes based on pairwise distances. All nine accessions are believed to be progeny of C. reticulata Blanco and C. maxima (Fig. 1). A more distantly related progenitor, Citrus medica, was not included in the analysis (Fig. 1). The authors’ comparative analysis helps to address several controversial hypotheses about the phylogenetic relationships between members of the citrus family and to map the evolution of citrus genomes from a paleohexaploid eudicotyledonous ancestor4. Although the analysis does support ancestral roles for C. reticulata and C. maxima, it calls into question the widely held assumption that traditional mandarins, which comprise a mixture of different species of ancient and recent hybrids5, are pure C. reticulata genotypes. Instead, Ponkan and Willowleaf mandarins are shown to be the product of considerable admixture between C. reticulata and C. maxima (Fig. 1). Another key insight that emerges from the authors’ analysis is that, despite many similar features, sweet orange and sour orange arose by very different breeding processes. Whereas sour orange is a pure F1 hybrid between C. maxima (egg donor) and C. reticulata (pollen donor) genotypes, sweet orange is a more complex admixture of C. maxima and C. reticulata genotypes. Finally, the new data negate the proposal that sweet orange might be derived
C. reticulata Blanco Mandarin
C. × sinensis • Sweet orange
C. maxima • Chandler pummelo • Low-acid pummelo
C. × aurantium • Sour orange
C. clementina • Haploid clementine (reference genome) • Clementine mandarin
C. medica Citron
C. × deliciosa • Willowleaf mandarin
C. paradisi Grapefruit
(C. reticulata) • Ponkan mandarin • W. Murcott mandarin
C. limon Lemon
C. micrantha
Kim Caesar/Nature Publishing Group
© 2014 Nature America, Inc. All rights reserved.
news and v iews
C. aurantifolia Mexican lime
Figure 1 The origin and evolution of select citrus species. The three major ancestors of citrus species included in the top row—Citrus reticulata, Citrus maxima and Citrus medica—contributed to the origins of all currently cultivated citrus species. Based on their high-quality genome sequence of a haploid clementine and resequencing of eight diploid species (indicated with bullet points against a yellow background), Wu et al.2 propose taxonomic relationships between members of the citrus family. Species not highlighted in yellow were not included in the current analysis. Lines with arrowheads indicate contributions to hybrids; lines without arrowheads represent simple introgression. The use of parentheses to denote Ponkan mandarin and W. Murcott mandarin as C. reticulata indicates that although both are almost pure C. reticulata, both resulted from a low level of introgression with C. maxima.
from a simple backcross between pummelo and mandarin ((C. maxima × C. reticulata) × C. reticulata)1. The large haplotypic block found in chromosome 2 of the two genotypes could be consistent with this proposal only if both putative parents had some pummelo ancestry. Instead, the authors show that sweet orange could be derived only from a cross between (C. maxima × C. reticulata) × C. maxima as egg donor and a male C. reticulata, with some introgression with C. maxima. Knowledge of the maternal and paternal donors that gave rise to the diverse range of citrus species will make it possible to implement an entirely new strategy for developing useful citrus varieties through breeding. By recapitulating step by step the domestication process used long ago to generate a particular variety, breeders should be able to screen progeny for subtle differences in their genetic composition. Inclusion of those lines in subsequent rounds of crossing or admixture should accelerate the rate with which breeders can emulate desirable traits while introducing genetic diversity elsewhere in the genome. Whereas C. maxima was the maternal donor in the
nature biotechnology volume 32 number 7 JUly 2014
lineage(s) that gave rise to pummelos, sweet oranges and sour oranges, C. reticulata was the maternal donor for all of the mandarins and the clementine variety studied (Fig. 1). Given that many commercially desirable varieties appear to have emerged after only a few generations of interbreeding between C. maxima and C. reticulata, it seems likely that sequencing-driven breeding using haplotype information should rapidly achieve other beneficial interspecies admixtures in a more targeted and rational way. The data of Wu et al.2 include >1.5 million fixed single-nucleotides that distinguish C. maxima from C. reticulata and define large haplotypic blocks characteristic of the genomes of mandarins, pummelos and oranges. These data will enable breeders to screen progeny for those most likely to improve commercial and agricultural traits while introducing genetic diversity. The availability of haplotype blocks also offers the opportunity to analyze the huge diversity of citrus germplasm in China6 to establish whether pure C. reticulata samples still exist, or to reveal the existence of additional novel species. New interspecies 641
news and v iews combinations between C. maxima, C. reticulata or other yet-to-be-identified wild species might provide superior varieties resistant to abiotic and biotic stresses7, such as that caused by the bacterium responsible for the citrus greening or Huanglongbing disease8, which is decimating orchards across the world and threatening the global citrus industry. Finally, the work of Wu et al.2 will lead to a better understanding of the genetic basis of the extraordinary diversity in the colors, flavors, sizes and aromas of citrus fruits and whether these might be engineered in novel varieties.
COMPETING FINANCIAL INTERESTS The authors declare no competing financial interests. 1. Xu, Q. et al. Nat. Genet. 45, 59–66 (2013). 2. Wu, G.A. et al. Nat. Biotechnol. 32, 656–662 (2014). 3. Velasco, R. et al. Nat. Genet. 42, 833–839 (2010). 4. Salse, J. Curr. Opin. Plant Biol. 15, 122–130 (2012). 5. Garcia-Lor, A. et al. Ann. Bot. (Lond.) 111, 1–19 (2013). 6. Gmitter, F.G. & Hu, X. Econ. Bot. 44, 267–277 (1990). 7. Bolger, M.E. et al. Curr. Opin. Biotechnol. 26, 31–37 (2014). 8. Gottwald, T.R. Annu. Rev. Phytopathol. 48, 119–139 (2010).
npg
© 2014 Nature America, Inc. All rights reserved.
Next-generation wearable electronics Michael J Cima New fabric-like sensors measure and transmit mechanical strain on the skin with unprecedented sensitivity. Motion, an act we take for granted until it is stolen by disease, age or injury, can now be measured by a growing array of portable electronic devices for consumer and medical use. Whether for fun, fitness or medicine, these simple wearable devices rely on tiny accel erometers that capture only the gross motion of a limb or entire body. Writing in Nature Nanotechnology, Son et al.1 describe a new class of integrated electronic sensors and memory elements packaged in fabric-like material that can record mechanical strains when applied to the skin, offering unprecedented fidelity for analyzing human motion. Devices built on this technology could find broad uses in medicine ranging from diagnostic tools to mechanically assisted prostheses. Portable electronics, originally devised for communication and entertainment, are expanding rapidly into the area of health metrics as they become more ‘wearable’ through form factors such as wristwatches, bracelets and eyewear. For example, Fitbit and similar products are worn like a watch and contain accelerometers or other sensors for tracking Michael J. Cima is David H. Koch Professor of Engineering, Faculty Director of the Lemelson-MIT Program, Department of Materials Science and Engineering, Koch Institute for Integrative Cancer Research, Cambridge Massachusetts, USA. e-mail: [email protected]
642
movement and recording vital signs. The clinical equivalent, such as systems by ActiGraph, can securely manage information across cohorts of study subjects or patients. These products presage a future in which measurements from wearable devices form the basis of clinical decisions. The transition to electronic health management builds on a long history of medical electronics. The first fully implanted pacemakers were developed in the 1960s, and since then similar devices have been developed for a large number of indications. Implanted devices for cardiac defibrillation, cardiac resynchronization, peripheral and vagus nerve stimulation and deep brain stimulation have saved many thousands of lives. All of these products were created by packaging existing electronic and microelectromechanical systems in ways appropriate to the application. An implanted device, for example, must have a package that is biocompatible and hermetic. This often takes the form of a welded titanium enclosure. Such devices are not truly integrated with tissue as they have completely different mechanical properties. The article by Son et al.1 is representative of a new trend in medical electronics: the design of discrete electronic and microelectromechanical systems devices for clinical measurement and actuation, embedded within packaging that can be worn directly on skin or contained in clothing. Such devices have been difficult
to develop because they must conform to the contours of the skin while also accommodating the large strains experienced during physical activity. Son et al.1 succeeded by incorporating device elements within an elastic matrix. The authors integrated mechanical strain sensors with arrays of data-storage elements, called resistive random access memory, into the elastic matrix (Fig. 1). When applied to the skin, the device’s onboard sensors record the strains that it experiences when the skin moves. These data are stored locally in the resistive random access memory and read at a later time. Several other groups are pursuing similar approaches, for example, multifunctional epidermal electronics2 from the group of John Rogers. Continuous strain measurements using the devices of Son et al.1 provide new opportunities for analyzing human motion. Mobility is a strong predictor of health outcomes in older adults3, and people who lose proficiency in balance and mobility are at increased risk of morbidity and mortality. Simple wearable accelerometers that measure gross motion of a limb or entire body have already been adopted in clinical settings. Indeed, a recent article in the Journal of the American Medical Association describes the use of accelerometer data as an outcome measure in evaluating structured physical activity to prevent mobility impairment in older adults4. Continuously monitored motion data generated by arrays of strain sensors1 could provide diagnostic information on the complicated interplay between individual muscles needed for motion and balance. Another application of strain-sensitive arrays would be to control input signals for exoskeleton assist devices—externally worn robotic hardware that assists human motion. The popular press depicts these technologies as part of the future, but they are already a reality in rehabilitation training for patients with stroke and other conditons5. Although chronic mechanical assist prostheses have improved substantially in recent years, the control systems integral to their operation6 are still inadequate. The primary control interface is surface electromyography, in which electrodes on the surface of the skin measure voltage signals within the depolarization zone of skeletal muscles. However, the relationships between the magnitude of the signals, the force exerted by a prosthetic and the displacement of the limb are difficult to determine. Strain-sensitive arrays could easily be integrated with surface electromyography to provide the high-quality measurements needed to more effectively control a prosthesis.
volume 32 number 7 JUly 2014 nature biotechnology
