have to obtain a delicate balance in the reac-tivities of the alkylidene intermediates formed during the reactions. To achieve this, they chose alkene reactants that included chemical groups that would bias the reactivities of the alkylidene intermediates through electronic and steric effects (repulsion that occurs when bulky groups are brought too closely together).

The first reactants they used were enol ethers — compounds in which a C=C bond is connected to the rest of the molecule by an oxygen atom (see Fig. 2 of the paper2). The C=C bonds in enol ethers have a higher electron density than those in simple alkenes, which makes them more reactive for meta thesis. The authors found that, for the best conversion of reactants to products and for maximum yields of Z alkenes, the number of moles of enol ether used in the reactions should be ten times that of the other alkene reactant. By this means, they obtained high yields of heterocoupled products in reactions of enol ethers with simple alkenes, with exceptionally good Z selectivity.

Why are the reactions so successful? Dur-ing metathesis reactions, a key intermediate known as a methylidene is produced. This methylidene can initiate reverse-CM reac-tions with Z-alkene products and generate unwanted E isomers. But in Hoveyda and colleagues’ reactions, the methylidene reacts mainly with the enol ether reactants because there is an excess of them and because they are more reactive. What’s more, the electronic properties of the resulting alkylidene inter-mediate are such that it preferentially reacts with an alkene rather than with another electron-rich enol ether. Heterocoupling is therefore the main reaction, rather than homocoupling between enol ethers.

The enol ether reactants used by the authors were cheap and readily available, but some enol ethers are much more expensive, in which case using a large excess of them would be un desirable. Fortunately, Hoveyda et al. found that unwanted methylidene formation can be minimized if the volatile ethylene by-product is removed by evaporation as it is produced. In this way, the authors maintained high Z selec-tivity in their reactions using just twice as much enol ether as alkene. They demonstrated the effectiveness of this protocol by using it to synthesize an unusual lipid containing a Z enol ether substructure.

The second reactants examined by the authors were allylic amides — compounds in which an amide group is attached at the carbon atom adjacent to a C=C bond (see Fig. 3 of the paper2). Unlike enol ethers, the C=C bonds in allylic amides are electroni-cally similar to those in simple alkenes, mak-ing homo coupling between the amides a potential problem. But Hoveyda et al. used a compound that incorporates a bulky amide group, which prevents such homocoupling6 — although homo coupling between the other

alkene molecules in the reaction could still be a problem. Never theless, the authors obtained excellent yields and Z selectivities in their CM reactions of allylic amides with alkenes.

The factor that determines the final prod-uct in these reactions is the relatively slow reverse CM of the heterocoupled product that is formed initially, compared with the faster reverse reaction of homocoupled alkenes. Removal of the ethylene by-product from the reactions was critical for efficient Z-selective CM, because this minimized the formation of methylidene species that convert Z isomers of products to their E isomers. The authors went on to use CM of an allylic amide with an alkene as a key step in the synthesis of the antitumour agent KRN7000. The use of this reaction allowed the agent to be made in just nine steps — the shortest synthetic sequence for the compound reported so far, under lining the effectiveness of the method.

Several aspects of these new metathesis reac-tions2 remain to be further refined: the range of alkene substrates that can be used should be broadened beyond enol ethers and allylic amides, for example, and ways should be

found to avoid using an excess of one of the reagents. The conversion of starting materi-als to products in the reactions is currently not complete, so achieving complete conver-sion and higher yields of products without sacri ficing the Z selectivity is also desirable. Never theless, these Z-selective CM reactions are highly promising and will potentially be of use for the preparation of numerous com-pounds, with far-reaching consequences for the future of metathesis chemistry. ■

Daesung Lee is in the Department of Chemistry, University of Illinois at Chicago, Chicago, Illinois 60607-7061, USA. e-mail: dsunglee@uic.edu

1. http://nobelprize.org/nobel_prizes/chemistry/laureates/2005

2. Meek, S. J., O’Brien, R. V., Llaveria, J., Schrock, R. R. & Hoveyda, A. H. Nature 471, 461–466 (2011).

3. Crowe, W. E. & Goldberg, D. R. J. Am. Chem. Soc. 117, 5162–5163 (1995).

4. Solans-Monfort, X., Clot, E., Copéret, C. & Eisenstein, O. J. Am. Chem. Soc. 127, 14015–14025 (2005).

5. Ibrahem, I., Yu, M., Schrock, R. R. & Hoveyda, A. H. J. Am. Chem. Soc. 131, 3844–3845 (2009).

6. Chatterjee, A. K., Choi, T.-L., Sanders, D. P. & Grubbs, R. H. J. Am. Chem. Soc. 125, 11360–11370 (2003).

R E P R O D U C T I V E B I O L O G Y

In vitro sperm maturation Anticancer therapies can impair male fertility. Whereas men can opt to freeze their sperm before treatment, young boys don’t produce mature sperm and so lack this choice. Work in mice offers hope for such patients. See Letter p.504

M A R C O S E A N D E L & S H A H I N R A F I I

The blueprint for producing mature, functional spermatozoa in a labora-tory dish — all the way from stem cells

to flagellated sperm — has eluded reproduc-tive biologists for decades. With a view to the eventual in vitro production of human sperm for clinical use, the main criterion of nor-mal gamete function is the ability to support fertilization, with the subsequent development of normal offspring. Reporting on page 504 of this issue, Sato et al.1 meet this challenge in mice.

The process of spermatogenesis in mam-mals persists throughout almost all of adult-hood. It starts with spermatogonial stem cells, which differentiate from type A spermato-gonia into type B spermatogonia, and then into spermatocytes. The spermatocytes undergo meiotic cell division to form spermatids, and finally spermatozoa2. Complete maturation takes more than a month in most mammals.

Previous efforts to nurture sperm in vitro

have either managed to recapitulate only parts of this complex differentiation process or failed to demonstrate the production of normal, fertile offspring. Although the organ-culture methods developed in the 1960s allowed germ cells to progress to meiosis3, until now no methods could support the entire process.

Sato et al.1 reasoned that the orchestration of cell-maturation signals during spermatogen-esis should be achievable in a dish by providing nearly all of the cellular components found in the testes. To this end, they cultured testicular fragments to maintain the proper microenvi-ronment for cell differentiation. This involved suspending fragments of immature testis on a semi-solid support, such as agarose, and partially bathing them in liquid. This organ-fragment culture system — referred to as the gas–liquid interface method — balances the delivery of nutrients from the culture medium to the maturing cells with the need for efficient gas exchange to maintain spermatogenesis for more than two months.

To visualize and score maturing germ cells

2 4 M A R C H 2 0 1 1 | V O L 4 7 1 | N A T U R E | 4 5 3

NEWS & VIEWS RESEARCH

© 2011 Macmillan Publishers Limited. All rights reserved

50 Years Ago“The great majority of school children are not only robust and healthy, but are taller and heavier than their predecessors”, states Sir John Charles, the chief medical officer of the Ministry of Education, in his report for 1958 and 1959 … Only about five per cent of school children contracted a notifiable infectious disease in the years under review. Tuberculosis continued to decline. Poliomyelitis reached its lowest level for thirteen years, and vaccination against it was undertaken vigorously everywhere … The main cause of death among children is through accident … In 1958, 869 children of 5–15 years died from accidents, including 395 from accidents involving motor vehicles, 174 from drowning, and 58 from burns and scalds.From Nature 25 March 1961

100 Years AgoI have just been told a very interesting story by Mr. James Day of this town. Many years ago he and his father … noticed a fox come searching along the hedgerows … they saw that he was collecting the sheep’s wool caught on the thorns and brambles. When he had gathered a large bunch he went down to a pool … and backed slowly brush first into the water, until he was all submerged except his nose and the bunch of wool, which he held in his mouth. He remained thus for a short time, and then let go of the wool, which floated away; then he came out, shook himself, and ran off. Much astonished at this strange proceeding, they took a shepherd’s crook … and pulled the wool out. They found that it was full of fleas, which, to save themselves from drowning, had crept up and up the fox’s brush and body and head and into the wool, and that thus the wily fox had got rid of them. T. McKenny HughesFrom Nature 23 March 1911

surprising, given that germ-cell maturation is known to depend on somatic-cell signals. In fact, even when embryonic stem cells have been used to produce germ cells in vitro7, sig-nals contributed by differentiating testicular somatic cells in the culture seem to be required.

The exact nature of the external signals that enhance sperm maturation is not the only remaining mystery. It is also not known whether the resulting offspring — especially those produced from cryopreserved tissue — are generally healthy. Indeed, fertility of the offspring is just a crude indicator of whether gametes are ‘normal’. Investigations should be made into whether the progeny Sato and co-workers generated by IVF are healthy in other ways (with respect to ageing, immune function, behaviour and so on).

As for the consequences of in vitro sperma-to genesis at the molecular and cellular levels, previous data have indicated8 that adverse epi genetic effects occur when cells, especially gametes, are maintained in culture. Whether DNA repair, which is essential for spermato-genesis in vivo, functions normally in vitro is another concern. Subtle genetic or epigenetic changes could be pivotal for the well-being of subsequent generations. These caveats aside,

under different conditions, the authors used a marker for spermatogenesis4: their mice were genetically engineered to express green fluorescent protein (GFP) under the control of regulatory elements for genes that are activated only when germ cells progress into meiosis and beyond. As expected, the tissues that they col-lected from the testes of newborn mice showed little or no baseline expression of GFP.

The authors had previously optimized4 various parameters involved in in vitro sperm formation, including temperature and the choice of basic ingredients for the medium. Notably, fetal bovine serum (FBS) seemed to be an important component, because its absence resulted in negligible maturation — as evidenced by the lack of the GFP signal.

In their present work, Sato et al.1 confirm the importance of FBS but, borrowing from the field of embryonic-stem-cell biology5, they find that an alternative to FBS known as knockout serum replacement (KSR) is even more effective. This finding is counter- intuitive, because KSR is commonly used to maintain stem cells in an undifferentiated state. A clue to the mechanism involved comes from the fact that the lipid-rich albumin component of KSR is itself highly effective in boosting differentiation.

The authors used in vitro fertilization (IVF) techniques to demonstrate the authenticity of the sperm collected from their cultures: they obtained male and female offspring that were themselves fertile.

The preservation of fertility is a major con-cern for patients requiring therapy, such as chemotherapy or radiation therapy, that can inadvertently destroy germ cells. In men, this problem can be mitigated by banking sperm before treatment. The solution is less straightforward in pre-pubescent boys. On the basis of pioneering work in animals by Brinster6 and others, the idea of transplanting cryopreserved spermatogonial stem cells is a reasonable strategy, although it has not yet been rigorously assessed in humans. Further-more, the technology for the long-term culture and expansion of human spermatogonial stem cells has not been standardized, nor has the safety of the approach been tested.

Sato and colleagues’ results suggest a viable alternative. In this scheme, boys would under go testicular biopsy before chemotherapy or radiation therapy, to obtain tissue for cryo-preservation (Fig. 1). If infertility occurs, the testicular fragments could be thawed and sperm obtained from organ culture for IVF. Such a protocol would bypass the need for surgical spermatogonial stem-cell transplantation.

It remains unclear whether the success of this system is due to signalling molecules released by the germ cells themselves, or to molecules released by the surrounding somatic (non-germ) cells, or to both. Nonetheless, the integrity of somatic cells, especially Sertoli cells, seems to be essential. However, this is not

Figure 1 | Possible scenario for preserving fertility, based on the new work1. After a boy has been diagnosed with cancer, a biopsy would be performed to obtain testicular fragments for long-term cryopreservation. Later in life, when the individual wants to start a family, fertility would be assessed, and if he cannot produce functional sperm, the stored tissues could be thawed for organ culture. Sperm formed by this ex vivo method could then be used for in vitro fertilization (IVF).

Diagnosis

Testisbiopsy

Long-termstorage

Organculture

IVFChildbearing

Infertility

Therapy

4 5 4 | N A T U R E | V O L 4 7 1 | 2 4 M A R C H 2 0 1 1

NEWS & VIEWSRESEARCH

© 2011 Macmillan Publishers Limited. All rights reserved

The implication of cAMP-mediated signal-ling in schizophrenia is particularly note worthy because this molecule has several crucial func-tions, including effects on neuronal activity and gene transcription8. Changes in VPAC2 activity are one route by which this signalling pathway might be affected in schizophrenia. Unfortu-nately, how 7q36.3 microduplications affect the structure of VPAC2 and its coupling to adenylyl cyclase remains unknown.

VPAC2 is expressed throughout the nervous system, and so it is essential to identify where in the brain of patients with schizophrenia the microduplications in VIP2R are active and how they alter neuronal behaviour. Candidate sites include the amygdala, hippocampus and suprachiasmatic nuclei, which are respectively involved in emotion, learning and memory, and circadian rhythms. Interestingly, mice with targeted disruptions in VIP–VPAC2 signalling show alterations in neural processes affected in schizophrenia, including cognition9, responses to sensory stimuli10 and sleep11.

Vacic and colleagues’ study2 coincides with another investigation12 that also points to CNVs at 7q36.3 as potential risk factors for schizophrenia. Thus, converging evidence sup-ports voluntary testing for CNVs at this locus to detect the likelihood of a person develop-ing schizophrenia. Of course, such an endeav-our is fraught with concerns over whether a person really wants to know that they are at risk of developing schizophrenia, or indeed over how this CNV influences particular symptoms associated with the disorder.

These studies2,12 also identify VPAC2 as a potential target for therapy. The distribution of this receptor in mammalian tissues is rea-sonably well mapped, and VPAC2-selective compounds have been, and continue to be, developed. Potentially, therefore, drugs could be tailored to regulate VIP–VPAC2 signalling in distinct regions of the brain. A possible, though surmountable, problem is that neuro-peptides other than VIP bind to VPAC2, and that VIP can signal through other receptors. Another cause for caution is that the relative incidence of CNVs at the 7q36.3 locus — although significant — is low and may account for only a small proportion of the risk. It is more likely, therefore, that VIP inhibitors could form part of a cocktail of drugs to treat the symptoms associated with schizophrenia. ■

Hugh D. Piggins is in the Faculty of Life Sciences, University of Manchester, Manchester M13 9PT, UK. e-mail: hugh.d.piggins@manchester.ac.uk

1. van Os, J. & Kapur, S. Lancet 374, 635–645 (2009).

2. Vacic, V. et al. Nature 471, 499–503 (2011).3. Howes, O. D. & Kapur, S. Schizophrenia Bull. 35,

549–562 (2009).4. Tam, G. W. C., Redon, R., Carter, N. P. & Grant, S. G. N.

Biol. Psychiatry 66, 1005–1012 (2009). 5. Mulle, J. G. et al. Am. J. Hum. Genet. 87, 229–236

(2010).

the organ-culture approach1 represents a cru-cial experimental advance along the thorny path to the clinical use of sperm developed in vitro. ■

Marco Seandel is in the Department of Surgery, and Shahin Rafii is in the Department of Genetic Medicine and at the Ansary Stem Cell Institute, Weill Cornell Medical College, New York, New York 10065, USA. e-mails: mseandel@med.cornell.edu; srafii@med.cornell.edu

1. Sato, T. et al. Nature 471, 504–507 (2011).2. Russell, L. D., Ettlin, R. A., Hikim, A. P. S. & Clegg, E. D.

Histological and Histopathological Evaluation of the Testis (Cache River Press, 1990).

3. Staub, C. J. Androl. 22, 911–926 (2001).4. Gohbara, A. et al. Biol. Reprod. 83, 261–267

(2010).5. Price, P. J., Goldsborough, M. D. & Tilkins, M. L.

International patent application: PCT/US98/00467 (1998).

6. Brinster, R. L. & Zimmermann, J. W. Proc. Natl Acad. Sci. USA 91, 11298–11302 (1994).

7. Kee, K., Angeles, V. T., Flores, M., Nguyen, H. N. & Reijo Pera, R. A. Nature 462, 222–225 (2009).

8. Nayernia, K. et al. Dev. Cell 11, 125–132 (2006).

S C H I Z O P H R E N I A

Zooming in on a gene Genome-wide association studies are often criticized for providing little insight of immediate physiological relevance. The finding of one such study, which implicates a signalling molecule in schizophrenia, is welcome news. See Letter p.499

H U G H D . P I G G I N S

Schizophrenia affects between 3 and 7 in every 1,000 individuals1. Yet, as for most mental disorders, the origins and mech-

anisms of this multi-symptom psychiatric illness remain elusive. In this respect, Vacic and colleagues’ report2 on page 499 of this issue — the first genome-wide association study to implicate a single gene in this complex disease — is enlightening.

Throughout the late twentieth century, schizophrenia research focused mainly on alterations in the levels of the neurotransmit-ter dopamine and in particular of its receptors3. However, this proved to be only one piece of the puzzle.

Hopes were again raised with the emergence of tools that can detect and map genetic varia-tions and alterations. One such modification, known as copy number variation (CNV), involves structural changes in the genome that result from disruptions, duplications or dele-tions of sections of DNA. Although research-ers tend to look for common CNVs that are associated with complex disorders, large-scale and sensitive genome-wide association studies, in which thousands of genes in patient sam-ples can be assessed, also allow the detection and mapping of rare CNVs and the genomic regions (loci) enriched in them.

One idea that is rapidly gaining momentum is that the accumulation of rare CNVs is asso-ciated with a high risk of developing schizo-phrenia and other mental disorders such as autism4. Indeed, CNVs on several chromo-somal loci (microdeletions at loci 1q21.1, 3q29, 15q13.3 and 22q11.2, as well as micro-duplication at the 16p11.2 locus) are associ-ated with the risk of schizophrenia4,5. Together, however, these CNVs account for only 2–4%

of schizophrenia cases and do not extensively contribute to our understanding of the mecha-nism of this disorder.

Vacic et al.1 carried out a genome-wide association study of 8,290 patients with schizophrenia. They find that 0.35% of these patients carry rare CNVs in the chromosomal locus 7q36.3. These rare copy number gains, or microduplications, were much less frequent (0.03%) among 7,431 healthy controls. It is particularly intriguing that all of the micro-duplications overlapped with — or occurred within 89 kilobases upstream of — a single gene, VIPR2.

The VIPR2 gene encodes the VPAC2 recep-tor, which binds a brain chemical known as vasoactive intestinal peptide (VIP). The VIP–VPAC2 signalling pathway has been implicated in several neural and behavioural processes6, and is known to influence circadian rhythms7 and cognition6. Vacic and co-workers’ identification of several duplications in the protein-coding sequences at or near VIPR2 suggests that alteration in VIP signalling is also a risk factor for schizophrenia.

The activity of the VPAC2 receptor is posi-tively coupled with that of the intracellular enzyme adenylyl cylase. And stimulation of this receptor increases the intracellular levels of the signalling molecule cyclic AMP, which is gener-ated by adenylyl cyclase. To determine whether the VIPR2-associated CNVs have a physiologi-cal role, the authors2 examined immune cells called lymphocytes obtained from subjects who had 7q36.3 microduplications and from control subjects. Specifically, they measured cAMP levels in response to the addition of VIP or another selective VPAC2 activator. Lym-phocytes from patients carrying the micro-duplications showed significantly higher levels of cAMP than did cells from controls.

2 4 M A R C H 2 0 1 1 | V O L 4 7 1 | N A T U R E | 4 5 5

NEWS & VIEWS RESEARCH

© 2011 Macmillan Publishers Limited. All rights reserved