© 1999 Macmillan Magazines Ltd

how the findings should be interpreted.Take, for instance, studies in which one orboth of us participated6,7. In one6 a mecha-nism known as the Matthew effect was foundto influence the growth rates of researchdepartments in the natural sciences andmedicine at Flemish universities. From amanagement point of view, these depart-ments are academic business units. Both interms of attracting external funding andpublication output, large departmentsbecame larger and small departmentsremained small, contrary to the findings ofPlerou and colleagues. Moreover, an over-emphasis on short-term objectives led to adecrease in the publication output per full-time equivalent spent on research .

In the other7, which involved an analysisof universities in the Netherlands during1980–96, changes in the distribution of stu-dents among universities and the outcomesof research evaluation studies seem to havebeen responsible for a trend towards a uni-formity of publication output in the naturaland life sciences. Smaller universities had ahigher growth rate of publication outputthan larger ones. This highlights the impor-tance of basic public funding of universities,based on student enrolments, in the Dutchacademic research system. In the UnitedStates, such basic funding is considerably lessimportant than in several Western Europeancountries, as most research activities arefunded through grants8. So differencesbetween academic systems in different coun-tries need to be taken into account.

Finally, Plerou and colleagues’ analysis1

can be taken further in two respects. First,

more sophisticated indicators of research per-formance are available9, and could be used.Second, in their study all scientific disciplinesare aggregated; yet there are big differencesbetween disciplines in the amount and originof funding, partly due to the increasingimportance of targeted research programmesin priority areas. The effects of the increasingmobility of scientists, of informal networksand of the growing links between universitiesand other parts of the R&D system should alsobe considered. Nonetheless, the observedsimilarities between universities and businessfirms will stimulate further debate andresearch on the effectiveness of nationalacademic research systems.Henk F. Moed is at the Centre for Science andTechnology Studies, Leiden University, PO Box9555, 2300 RB Leiden, The Netherlands.e-mail: moed@cwts.leidenuniv.nlMarc Luwel is at the Science and InnovationAdministration, Ministry of the FlemishCommunity, Boudewijnlaan 30, B-1000 Brussels,Belgium.e-mail: marc.luwel@wim.vlaanderen.be1. Plerou, V., Nunes Amaral, L. A., Gopikrishnan, P., Meyer, M. &

Stanley, H. E. Nature 400, 433–437 (1999). 2. Schelsky, H. Einsamkeit und Freiheit (Rowolt, Hamburg, 1963).3. Gibbons, M. et al. The New Production of Knowledge (Sage,

London, 1994). 4. OECD Group on the Science System University Research in

Transition (OECD, Paris, 1998).5. Roussel, P. A., Saad, K. A. & Erickson, T. J. Third Generation

R&D (Harvard Business School, Boston, 1991).6. Moed, H. F., Luwel, M., Houben, J. A., Spruyt, E. & Van Den

Berghe, H. Scientometrics 43, 231–255 (1998). 7. Moed, H. F., van Leeuwen, Th. N. & Visser, M. S. Res.

Evaluation 8, 60–67 (1999).8. National Science Board Science and Engineering Indicators

(National Science Foundation, Arlington, Virginia, 1998).9. van Raan, A. F. J. Handbook of Quantitative Studies of Science

and Technology (North Holland, Amsterdam, 1988).

tons) are translocated from the matrix to thecytosolic side of the membrane. In sum,then, eight charges are translocated acrossthe dielectric barrier in the protein.

It was thought that, to efficiently capturethe energy released by the oxygen chemistryfor proton translocation, the proton-pump-ing steps occur during the oxidative phase.Indeed, ten years ago, Wikström postulated9,from equilibrium measurements, that thetranslocation of all four pumped protons wascoupled to two specific reactions within theoxidative phase. This model, which has beenthe foundation for our thinking about pro-ton translocation ever since, was called intoquestion last year by Hartmut Michel10, whoproposed that one proton is pumped duringthe initial reduction of the enzyme, beforethe reaction with oxygen, and that the otherthree protons are pumped during theoxidative phase. But Michel’s model wasnot substantiated by experiment.

By making two types of measurement,Wikström and co-workers8 have now deter-mined the amount of charge that is trans-located during each phase of the enzymaticcycle. In the first set of experiments, theauthors measured the pH change of lipo-somes with cytochrome c oxidase embeddedin their membranes during the two halves ofthe catalytic cycle. In the second set, the time-resolved potential across a membrane con-taining cytochrome c oxidase was followedduring turnover. It came as a complete sur-prise to find that the translocation of charge isincomplete at the end of the oxidative phase.

So what happens to the remaining charge?This, it turns out, is translocated only whenthe reductive phase immediately follows theoxidative phase. About half of the charge istranslocated during each of the phases, indi-cating that part of the energy generated in theoxidative phase is conserved in the proteinand indirectly coupled to proton transloca-tion during the reductive phase. This remark-able result requires a shift in how we mustthink about proton pumping in cytochrome coxidase, because it invalidates the hypothesisthat proton translocation is directly coupledto the oxygen chemistry during the oxidativephase. But where is the energy stored at theend of the oxidative phase? And how is itconverted to proton translocation in thesubsequent reductive phase?

At the end of the oxidative phase, the fourmetal centres of the enzyme are fully oxi-dized. This state, termed O~, is a metastableintermediate that lasts for several secondsunder the conditions used by Wikström andcolleagues. The structure of O~ must be verydifferent from that of the resting oxidizedenzyme — the ‘O’ state — because reductionof O does not induce proton translocation8.The energy for driving the delayed chargetranslocation could be stored locally in a fewhighly strained chemical bonds near or atthe catalytic site. On reduction of the metal

news and views

412 NATURE | VOL 400 | 29 JULY 1999 | www.nature.com

The ‘central dogma’ of bioenergetics isthe chemiosmotic theory, which statesthat energy stored as a proton gradient

across biological membranes (the proton-motive force) is converted to useful chemicalenergy in the form of ATP. In eukaryotes, theproton-motive force is generated across theinner membrane of the mitochondria byharnessing the energy released from serialelectron-transfer events in membrane-bound proteins. This chain culminates in thefour-electron reduction of oxygen to waterby an enzyme called cytochrome c oxidase.Associated with this oxygen chemistry, fourprotons are pumped across the membrane inopposition to a proton gradient1.

The molecular mechanism by which oxy-gen reduction is coupled to proton transloca-tion — the redox linkage — remains a centralquestion. Guided by spectroscopic2,3, crystal-lographic4,5 and mutagenesis6 results, severalmodels for redox linkage have been proposed,

but each has its flaws. One feature that has beenlacking is a quantitative temporal relationshipbetween the chemical steps and the transloca-tion of protons7. On page 480 of this issue,Mårten Wikström and co-workers8 propose asurprising time frame for this relationship.

To study the reaction of cytochrome coxidase, all four of its redox centres (Fig. 1)are reduced, and the enzyme is exposed tooxygen. Complete turnover of the enzymerequires two phases. During the oxidativephase, four electrons are transferred to theoxygen molecule, the O–O bond is cleaved,and the enzyme becomes fully oxidized. Thisis followed by the reductive phase, in whichthe enzyme is re-reduced. Looking at thereaction as a whole, four protons (the chemi-cal protons) are taken up from the matrixside of the inner mitochondrial membraneand four electrons from cytochrome c areused in the formation of water at the catalyticsite. Four more protons (the pumped pro-

Bioenergetics

Two phases of proton translocationDenis L. Rousseau

© 1999 Macmillan Magazines Ltd

centres, this energy could be releasedthrough the charge translocation. The obvi-ous candidates for such bonds are thoseformed by the coordination of exogenousligands in the catalytic site (hydroxides, forexample) with the iron and the copper in theO~ state2,11. Or, they could be those formedby the intrinsic amino-acid residues that linkthe redox centres to the polypeptide. Indeed,on reduction of a closely related bacterialoxidase, one of the histidine bonds that coor-dinates to the copper atom in the catalyticsite has already been seen to rupture12. Thissuggests that bond breakage or formationcould be involved in the energy release.

But it is also possible that the conservedenergy in the O~ intermediate is distributedover many polypeptide bonds in the protein.Relaxation of the intermediate’s conforma-tion can be coupled to proton translocationon reduction of the metal centres. If thisis the case, we need to identify the criticalelement that triggers release of the energy,allowing the charge translocation to occur.

One interesting possibility comes fromconsideration of haemoglobin. Here, bind-ing of a ligand to the distal side of the haemmoiety is coupled to protein conformationalchanges through the proximal histidine–iron bond that links the haem to thepolypeptide. This cooperative interaction isdetected in the resonance Raman spectrumof the transient species generated by pho-todissociation of carbon-monoxide-boundhaemoglobin13. Changes similar to those

seen in transient spectra of haemoglobinhave already been detected in the spectra ofphotodissociated carbon-monoxide-boundcytochrome c oxidase14,15. It is conceivablethat, on reduction of the enzyme, release ofhydroxide from the haem moiety could trig-ger movement of the iron–histidine bond,thereby inducing the global conformationalchanges that are required for translocatingthe protons. If this process does turn out totrigger proton translocation, it is remarkablethat nature could use nearly identical com-munication mechanisms in such vastlydifferent proteins.Denis L. Rousseau is in the Department ofPhysiology and Biophysics, Albert Einstein Collegeof Medicine, Bronx, New York 10461, USA.e-mail: rousseau@aecom.yu.edu1. Babcock, G. T. & Wikström, M. Nature 356, 301–309 (1992).

2. Han, S., Ching, Y.-c. & Rousseau, D. L. Nature 348, 89–90 (1990).

3. Ferguson-Miller, S. & Babcock, G. T. Chem. Rev. 96, 2889–2907

(1996).

4. Tsukihara, T. et al. Science 269, 1069–1074 (1995).

5. Iwata, S., Ostermeier, C., Ludwig, B. & Michel, H. Nature 376,

660–669 (1995).

6. Gennis, R. B. Biochim. Biophys. Acta 1365, 241–248 (1998).

7. Gennis, R. B. Proc. Natl Acad. Sci. USA 95, 12747–12749 (1998).

8. Verhovsky, M. I., Jasaitis, A., Verhovskaya, M. L., Morgan, J. E.

& Wikström, M. Nature 400, 480–483 (1999).

9. Wikström, M. Nature 338, 776–778 (1989).

10.Michel, H. Proc. Natl Acad. Sci. USA 95, 12819–12824 (1998).

11.Mitchell, R., Mitchell, P. & Rich, P. R. Biochim. Biophys. Acta

1101, 188–191 (1992).

12.Osborne, J. P. et al. Biochemistry 38, 4526–4532 (1999).

13.Friedman, J. M. Science 228, 1273–1280 (1985).

14.Sassaroli, M., Ching, Y.-c., Argade, P. V. & Rousseau, D. L.

Biochemistry 27, 2496–2502 (1988).

15.Findsen, E. W., Centeno, J., Babcock, G. T. & Ondrias, M. R.

J. Am. Chem. Soc. 109, 5367–5372 (1987).

news and views

NATURE | VOL 400 | 29 JULY 1999 | www.nature.com 413

Daedalus

No difference at allSexual equality is a hot topic nowadays.Men and women are held to be identical inall respects; or if not, they should be.Bertold Brecht remarked, of theimplacable social engineering of the oldEast German regime: “Such a governmentshould dissolve the people, and appointanother.” This idealistic aim, saysDaedalus, is now feasible, at least in thesexual sphere.

The ‘default setting’ of a fetus is female.If a male Y chromosome is present,however, it stimulates the production offetal testosterone, which sets developmentalong the male path. In particular, thishormone powerfully influences thegrowing brain. Thus, pregnant womenwith toxaemia used to be given doses oftestosterone, to counter the condition. Thegirls born to such women often behaved inmale ways, being aggressive, unromantic,and with little interest in babies.Conversely, pregnant women treated withsupplementary oestrogens often give birthto boys whose later behaviour is typicallyfeminine. They are often shy, or lack aproper obsession with sport or technology.Much depends on the timing and the doseof the modifying hormone.

So, says Daedalus, the way seems opento true sexual equality. Provided that thesex of the growing fetus can be determinedearly enough, the mother could be put on awell-judged programme of hormonetreatment designed to cancel the mentalsexual bias of the fetus without affecting itsgenital development. Feminist motherswill rush to demand the treatment. Thenext generation will no longer show thedeplorable mental differences that opposea just and equal society. Everyone, male orfemale, will be moderately aggressive,moderately sexual, moderately interestedin both babies and technology, and capableof competing interchangeably in all walksof life.

The moral climate of society willimprove dramatically. The demand for sexfrom men will at last be equalled by thatfrom women, so prostitution will vanish.The rival art-forms of pornography andromantic fiction will merge into anintriguing new unisex romanticpornography. The long, traumatic sexualodyssey of so many adolescents will also beeliminated, for both sexes will share thesame outlook and aspirations. Marriageand the family will benefit enormouslyfrom this new harmony. And the old line,“My wife (or husband) doesn’t understandme”, will be totally discredited. David Jones

Figure 1 Proton pumping in cytochrome c oxidase, based on the results of Wikström and colleagues8. Thecatalytic reaction, which occurs at the Fea3–CuB binuclear centre, is initiated by presenting di-oxygen tothe fully reduced enzyme (R). During the oxidative phase, the O2 molecule is cleaved by accepting twoprotons from the matrix side of the membrane and four electrons from the metal centres of the enzyme.At the end of this phase the enzyme is fully oxidized. About half of the proton translocation occursduring this phase. Some of the chemical energy released by the oxygen chemistry is conserved in ametastable conformation of the oxidized enzyme (the O~ state). If the subsequent reductive phaseimmediately follows the oxidative phase, the energy stored in the oxidized enzyme is released bycompleting the proton translocation. At the same time, two more protons are taken up from the matrixside of the membrane for formation of the two water molecules.

CytCyt

Cyt Cyt

Cytosol

Reducedenzyme

Oxidizedenzyme

Matrix

O2

2H2O 2H+

2H+

OH OH

mH+

mH+

nH+

nH+

O~

4e—

R

CuA1+

CuA2+

CuB2+

CuB1+Fea3

2+

Fea33+

Fea3+

Fea2+ Oxidative

phase

Reductivephase