R EPORTS tecting the nascent DNA from degradation, 7. A. K. Ganesan, J. Mol. Biol. 87, 103 (1974). 22. K. Hanada, M. Iwasaki, S. Ihashi, H. Ikeda, Proc. Natl. the amount of X-shaped intermediates is se- 8. Z. Horii, A. J. Clark, J. Mol. Biol. 80, 327 (1973). Acad. Sci. U.S.A. 97, 5989 (2000). 9. J. Courcelle, C. Carswell-Crumpton, P. C. Hanawalt, 23. T. Tsujimura, V. M. Maher, A. R. Godwin, R. M. Liskay, verely reduced in recF and recR mutants Proc. Natl. Acad. Sci. U.S.A. 94, 3714 (1997). J. J. McCormick, Proc. Natl. Acad. Sci. U.S.A. 87, 1566 (Fig. 3C). Furthermore, although recF and 10. J. Courcelle, P. C. Hanawalt, Mol. Gen. Genet. 262, (1990). recR mutants still fail to recover replication 543 (1999). 24. M. Lopes et al., Nature 412, 557 (2001). in the absence of either the RecQ or 11. S. Rangarajan, R. Woodgate, M. F. Goodman, Mol. 25. J. M. Sogo, M. Lopes, M. Foiani, Science 297, 599 Microbiol. 43, 617 (2002). (2002). the RecJ nuclease, the nascent DNA at the 12. Z. Horii, K. Suzuki, Photochem. Photobiol. 8,93 26. P. McGlynn, A. A. Mahdi, R. G. Lloyd, Nucleic Acids replication fork remains intact (10) and, in (1968). Res. 28, 2324 (2000). these mutants, the X-shaped intermediates 13. B. B. Konforti, R. W. Davis, Proc. Natl. Acad. Sci. 27. L. Postow et al., J. Biol. Chem. 276, 2790 (2001). U.S.A. 84, 690 (1987). 28. I. Mellon, P. C. Hanawalt, Nature 342, 95 (1989). are clearly restored (Fig. 3; fig. S2), indicat- 14. B. L. Webb, M. M. Cox, R. B. Inman, Cell 91, 347 29. C. P. Selby, A. Sancar, Science 260, 53 (1993). ing that a portion of the intermediates in the (1997). 30. We thank R. S. Lloyd for providing TEV, A. R. Poteete cone region are formed through the regres- 15. Q. Shan, J. M. Bork, B. L. Webb, R. B. Inman, M. M. Cox, for providing bacterial strains, and D. J. Crowley and J. Mol. Biol. 265, 519 (1997). A. K. Ganesan for their comments. Supported by the sion of the replication fork and the extrusion 16. J. Courcelle, P. C. Hanawalt, Proc. Natl. Acad. Sci. National Science Foundation, grant MCB0130486. of the nascent DNA (Fig. 4). Thus, DNA U.S.A. 98, 8196 (2001). lesions that block replication fork progression 17. J. K. Karow, L. Wu, I. D. Hickson, Curr. Opin. Genet. Supporting Online Material Dev. 10, 32 (2000). www.sciencemag.org/cgi/content/full/1081328/DC1 induce a regressed intermediate that is main- 18. K. Hanada et al., Proc. Natl. Acad. Sci. U.S.A. 94, Materials and Methods tained by RecA and RecF, -O, and -R. In the 3860 (1997). Figs. S1 and S2 absence of these proteins, the intermediate is 19. K. L. Friedman, B. J. Brewer, Methods Enzymol. 262, References 613 (1995). degraded by RecQ and RecJ, which have 20. L. Martin-Parras, P. Hernandez, M. L. Martinez-Robles, 9 December 2002; accepted 15 January 2003 been previously shown to process the nascent J. B. Schvartzman, J. Mol. Biol. 220, 843 (1991). Published online 23 January 2003; lagging strand before the resumption of rep- 21. Materials and methods are available as supporting 10.1126/science.1081328 lication (10). Recently, rad53 checkpoint mu- material on Science Online. Include this information when citing this paper. tants in have been shown to contain nascent lagging strand ab- normalities and accumulate a similar interme- diate after hydroxyurea treatment (24, 25), Taming of a Poison: Biosynthesis suggesting that similar mechanisms may op- erate in both prokaryotes and eukaryotes. It of the NiFe-Hydrogenase should be emphasized that not all lesions may block replication or be processed by the same mechanisms, and it is likely that a portion of Cyanide Ligands the X-shaped intermediates in wild-type cells Stefanie Reissmann,1 Elisabeth Hochleitner,2 Haofan Wang,3 are produced through recombinational ex- Athanasios Paschos,1 Friedrich Lottspeich,2 Richard S. Glass,3 changes. This is especially true in the case of August Bo¨ck1* uvrA mutants in which multibranched mole- cules accumulate. It remains to be determined NiFe-hydrogenases have an Ni-Fe site in which the iron has one CO and two whether fork regression in vivo is promoted CN groups as ligands. Synthesis of the CN ligands requires the activity of enzymatically or occurs spontaneously via two hydrogenase maturation proteins: HypF and HypE. HypF is a carbamoyl- positive supercoiling in front of the replica- that transfers the carbamoyl moiety of carbamoyladenylate to tion fork (26, 27). We propose that the re- the COOH-terminal cysteine of HypE and thus forms an -thiocar- gressed intermediate allows repair bamate. HypE dehydrates the S-carbamoyl moiety in an adenosine triphos- or alternative polymerases to gain access to phate–dependent process to yield the enzyme thiocyanate. Chemical model the replication blocking lesions, thereby al- reactions corroborate the feasibility of this unprecedented biosynthetic lowing processive replication to resume once route and show that thiocyanates can donate CN to iron. This finding the block to replication has been removed or underscores a striking parallel between biochemistry and organometallic bypassed. By analogy, the recovery of tran- chemistry in the formation of an iron-cyano complex. scription has been shown to require the re- moval of the RNA polymerase and nascent Hydrogenases catalyze the reversible oxi- because of the distinctive chemical nature transcript before repair enzymes can effect dation of molecular hydrogen into protons of their substrate and the reaction mecha- repair (28, 29). Consistent with this idea, the and electrons. They are widely distributed nism but also because of their potential appearance and duration of the regressed rep- among microorganisms, and they provide biotechnological applications (1). lication fork intermediate coincides with the them with the capacity either to use hydro- Two major classes of hydrogenases can be time it takes for the DNA lesions to be re- gen as an energy source or to dissipate differentiated according to the metal content moved and robust replication to resume. excess reducing equivalents in the form of of the : Fe-hydrogenases molecular hydrogen. These enzymes have and NiFe-hydrogenases. Although the overall References and Notes attracted considerable attention, not only structure of their metal centers differs, they 1. R. B. Setlow, P. A. Swenson, W. L. Carrier, Science 142, share one unusual feature: diatomic, nonpro- 1464 (1963). 2. J. Courcelle, D. J. Crowley, P. C. Hanawalt, J. Bacteriol. 1Department Biologie I, Mikrobiologie, University of teinaceous iron ligands, namely, carbon mon- 181, 916 (1999). Munich, Maria-Ward-Strasse 1a, D-80638 Munich, oxide and cyanide. In NiFe-hydrogenases, the 3. R. B. Setlow, W. L. Carrier, Proc. Natl. Acad. Sci. Germany. 2Max-Planck Institut fu¨r Biochemie, Abtei- iron of the center carries two cyanide and one U.S.A. 51, 226 (1963). lung Proteinchemie, Am Klopferspitz, D-82152 Mar- carbon monoxide moieties (2). The presence 4. W. D. Rupp, P. Howard-Flanders, J. Mol. Biol. 31, 291 tinsried, Germany. 3Department of Chemistry, Uni- (1968). versity of Arizona, Post Office Box 210041, Tucson, of these ligands stabilizes iron in a low oxi- 5. W. D. Rupp, C. E. I. Wilde, D. L. Reno, P. Howard- AZ 85721-0041, USA. dation and spin state. Flanders, J. Mol. Biol. 61, 25 (1971). 6. A. K. Ganesan, K. C. Smith, Mol. Gen. Genet. 113, 285 *To whom correspondence should be addressed. E-mail: In metal center synthesis and incorpora- (1971). [email protected] tion into proteins, important issues are con-

www.sciencemag.org SCIENCE VOL 299 14 FEBRUARY 2003 1067 R EPORTS trol of the fidelity of insertion of the correct toxicity in the free state and, in particular, bamoyl (CP) as a substrate and metal into the cognate target protein, mainte- because addition of CO and CN to the metal catalyzes both a CP phosphatase reaction in nance of a folding state of the protein com- reflects organometallic chemistry unprece- the absence of any other substrate and a petent for metal addition, and conformational dented in biology. CP-dependent hydrolysis of ATP into AMP

change to internalize the assembled metal A total of seven auxiliary proteins are and inorganic pyrophosphate (PPi). Because center (3). In this context, the synthesis of the required for the synthesis and incorporation the latter reaction is reversible, as shown CN and CO ligands of NiFe-hydrogenases of the metal center in NiFe-hydrogenases (4). by a CP-dependent pyrophosphate-ATP ex- poses intriguing questions because of their One of them, the HypF protein, accepts car- change, it was postulated that an adenylated CP derivative is an intermediate in the reac- tion and that the CP phosphatase activity may Fig.1. Modification of reflect a side reaction followed in the absence the hydrogenase mat- uration protein HypE of ATP (5). The necessity of CP for the by the HypF protein. synthesis of the metal center was then clearly (A) Binding of radio- proven by the inability of mutants devoid of actively labeled CP to CP synthetase activity to mature NiFe- HypF and/or HypE pro- hydrogenases (6). teins assessed by bind- The formation of an adenylated CP deriv- ing of the proteins to nitrocellulose filters. ative by HypF prompted studies on catalytic HypE⌬C is a HypE vari- self-carbamoylation of the protein. To deter- ant lacking the COOH- mine whether a stable carbamoyl derivative terminal cysteine resi- of purified HypF protein from Escherichia ␮ ␮ due. (B) HypE (2 M) and HypF (2 M) were reacted with radioactive CP in the absence of ATP (bar coli is formed, a filter binding assay was 1), in the presence of ADP-CH -P (bar 2) or AMP-CH -PP (bar 3). Bar 4 presents the result of a 2 2 ␮ devised (7) (Fig. 1). The radioactivity of standard assay with ATP as substrate, but the assays contained 250 g bovine serum albumin 14 instead of HypE and HypF. (C and D) Lane 1 contains only HypF; lane 2 contains only HypE. (Lanes C-labeled CP was not stably incorporated 3 to 6): HypF protein (2 ␮M) was reacted with 6 ␮M HypE protein in the assay mixture given above, into the HypF protein per se (Fig. 1A, second which contained as (NTP) (a) ATP, (b) AMP-CH2-PP, or (c) ADP-CH2-P. On bar). HypF and HypE from Helicobacter py- the right, F in (C) denotes HypF; E indicates HypE, and E*, a dimer of HypE. Denaturation with SDS lori reportedly interact with each other (8), in the absence of dithiothreitol (DTT) allows more radioactivity to be recovered, which indicates and HypE shares sequence similarity with the that the adduct is labile to thiols. The altered migration possibly reflects an oxidized form of HypE. PurM protein, which catalyzes an ATP- dependent dehydration reaction during purine Fig.2. Mass and tan- biosynthesis (9), a reaction that formally dem mass spectro- could also transform a carbamoyl into a metric analysis of cyano moiety. Consequently, transcarbamoy- HypE protein modified by HypF. (A) Spec- lation of HypE catalyzed by HypF was stud- trum of a reaction of ied. Purified HypE protein from E. coli in- HypE, HypF, CP, and deed was found to hydrolyze ATP into ADP ATP and (B) like (A) and inorganic phosphate in the absence of but with ADP-CH2-P any other substrate (10). When HypE was instead of ATP, digest- incubated with the HypF protein from the ed with endoprotein- ase Asp-N. Peptide same organism with different molar ratios in 311-322 (m/z 1360.7) the presence of ATP and CP (Fig. 1A), a modified with (A) CN stable adduct between radiolabeled CP and (m/z 1385.7) and (B) the HypE protein was formed. Its radioactiv- CONH2 (m/z 1403.7) ity quantitatively correlated with the amount is enlarged. (C) depicts of HypE, but not HypF, protein used in the the tandem mass spec- trum of unmodified assay. The stoichiometry between CP- peptide 311-322 (m/z derived radioactivity and HypE protein was 1360.7), (D) of pep- between 50 and 60%. These results support tide 311-322 modified the conclusion that HypF is required for the with CONH2 (m/z transfer of the carbamoyl group of CP to 1403.7), and (E)ofthe HypE. This transfer depends on the cleavage peptide modified with CN (m/z 1385.7). y of ATP into AMP and PPi (Fig. 1B). With the ions modified with nonhydrolyzable analog ADP-CH2-P (in CONH2 and CN are which a methylene group replaces the phos- labeled with (*) and phoanhydride bond) in place of ATP, transfer (#), respectively. For still took place, but in the presence of AMP- clarity, only the spec- CH -PP, it was blocked. These results are trum of the m/z range 2 from 640 to 1500 is consistent with HypF functioning as a car- depicted. The peptide bamoyltransferase that requires prior adeny- designation follows lation of CP, which is not possible with the standard nomen- AMP-CH2-PP as a substrate. clature (22) [for de- To prove that the HypE protein indeed tails see the insert in acts as the acceptor and to study whether Fig. 2C and (7)]. binding of the putative carbamoyl residue is covalent, HypE and HypF were incubated

1068 14 FEBRUARY 2003 VOL 299 SCIENCE www.sciencemag.org R EPORTS with [14C]CP and ATP, and then the assay was found that, in the MS/MS spectrum of treatment of S-(n-decyl) thiocarbamate with mixture was denatured by SDS under re- the 311-322 peptide derived from a reaction ethyl polyphosphate, PPE (12, 13), gave a ducing and nonreducing conditions, separat- in which the substrates were present, all y 55% yield of the corresponding thiocyanate ed by SDS–polyacrylamide gel electrophore- ions but no b ions carried the modification (14) [Fig. 4 (reaction 1)]. Similarly, (␩5- sis (SDS-PAGE), and blotted onto a nitrocel- (Fig. 2, D and E). Moreover, one peptide, C5H5)Fe(CO)2CONH2 (FpCONH2)(15) can lulose membrane. The blot was autoradio- a11/b11, corresponds to a fragment of amino be dehydrated, and sequential treatment with graphed (Fig. 1D) and then stained (Fig. 1C) acids 311 to 321 (lacking the cysteine). Be- PPE followed by triethylamine results in a (7). The results show that the radioactivity is cause it also lacks the modification, it is clear 57% yield of FpCN [Fig. 4 (reaction 2)], transferred to HypE during the reaction and that that each of the modifications is bound to the despite the previously reported (16) facile the binding is stable under denaturing condi- COOH-terminal cysteine. Finally, the spec- fragmentation of FpCONH2 cleaving the tions, which suggests covalent attachment. trum shown in Fig. 2E demonstrates that iron-to-carbon bond. Nevertheless, the bio- Apart from the sequence similarity with HSCN is cleaved off the precursor; therefore, chemical results show that the dehydration the PurM protein, HypE carries a PRIC motif it is the COOH-terminal cysteine that is either of the carbamoyl moiety occurs when as the last four amino acids (Pro-Arg-Ile- cyanated or carbamoylated. bound to sulfur, rather than iron, followed Cys), which is invariable in all members of The steps in CN formation catalyzed by the by transfer of the CN moiety to iron. Con- the HypE family. This cysteine residue at the hydrogenase maturation proteins HypF and sequently, chemical modeling of such a COOH-terminus prompted speculation that HypE are summarized in Fig. 3. HypF forms transfer was studied. Reduction of phenyl- its thiol group might be the target for car- the postulated AMPOCONH2 from which the thiocyanate by phenylthiolate in the pres- bamoylation. Indeed, the purified HypE vari- carbamoyl group is transferred to the COOH- ence of FpBr generated a 50% yield of ant with a deletion of this amino acid could terminal cysteine of HypE (step 1). HypE- FpCN [Fig. 4 (reaction 3)]. This reaction not accept the putative carbamoyl group carbamoylation (step 1) is followed by the validates the proposed model (Fig. 3, step transferred by HypF (Fig. 1A, last bar). ATP-dependent dehydration by HypE (step 3) in which a nucleophilic CN is transferred To identify the amino acid residue of 2) and the transfer to the iron atom (step 3). to an electrophilic iron center. Because the HypE that is modified, samples of the reac- Chemical models were studied to deter- identity of the iron species in the bio- tion mixtures were digested with endopro- mine the chemical feasibility of steps 2 and 3 logical system is unknown, another chem- teinase Asp-N, and the peptides were ana- in Fig. 3 and related reactions. The reactions ical model system was studied in which the lyzed by mass spectrometry (Fig. 2) (7). The carried out are shown in Fig. 4 (7). Thus, iron is nucleophilic and the CN electro- peptide pattern of an assay containing HypF, HypE, and CP plus ATP differed from that of a control assay lacking CP (11) in the appear- Fig.3. Biosynthesis of ance of a single peptide with a mass of the cyanide ligand. (Step 1), formation of 1385.7. This peptide corresponds to the carbamoyladenylate COOH-terminal peptide of the HypE protein, by HypF (F) as delin- ranging from amino acid 311 to 322 (mass eated from the ATP- 1360.7), and the mass of the additional pep- PPi exchange reaction, tide is compatible with the presence of a and carbamoyl trans- cyano group replacing a hydrogen from an fer to the COOH-ter- minal cysteine of HypE amino acid side chain (Fig. 2A). (E).(Step2),ATP-depen- Because HypE has intrinsic adenosine dent dehydration in- triphosphatase activity and exhibits se- volving phosphoryla- quence similarity to PurM, an enzyme cat- tion of the carbonyl alyzing an ATP-dependent dehydration re- oxygen of CP followed action, it seemed feasible that the CN mod- by dephosphorylation. (Step 3), chemical transfer of the cyano group to iron. ification derives from the dehydration of a carbamoyl precursor. This conjecture was tested by analyzing the peptide pattern of an assay in which ATP was replaced by the nonhydrolyzable ADP-CH2-P. Under this condition, the transcarbamoylation by HypF (requiring cleavage of ATP into

AMP and PPi) should still be possible, but the dehydration should be blocked. The mass increment of peptide 311-322 increas- es from 1360.7 to 1403.7 (Fig. 2B); this proves that HypE indeed carries a car- bamoyl modification. To identify the amino acid residue carrying the carbamoyl or the cyano modification, tandem mass spec- trometry (MS/MS) of this peptide from as- says lacking CP and ATP (Fig. 2C) was performed. The pattern was compared with those from reactions delivering carbamoy- Fig.4. Chemical models for the proposed biosynthesis of LnFeCN. (1) Dehydration of S-(n-decyl) thiocarbamate to the corresponding thiocyanate with ethyl polyphosphate (PPE). (2) Dehydration lated (in the presence of CP and ADP-CH2- ␩5 of iron carboxamido complex to iron cyanide where Fp is ( -C5H5)Fe(CO)2. (3) Transfer of a P) (Fig. 2D) or cyanated (in the presence of nucleophilic CN group to electrophilic iron atom. (4) Transfer of an electrophilic CN group to CP and ATP) (Fig. 2E) HypE protein. It nucleophilic iron atom.

www.sciencemag.org SCIENCE VOL 299 14 FEBRUARY 2003 1069 R EPORTS

philic. Treating t-butylthiocyanate with Fp– 27. We thank the Deutsche Forschungsgemeinschaft and Supporting Online Material brought about a 20% yield of FpCN (17) the Fonds der Chemischen Industrie for supporting www.sciencemag.org/cgi/content/full/299/5609/1067/DC1 [Fig. 4 (reaction 4)]. These chemical mod- the research of A.B. and the donors of the Petroleum Materials and Methods Research Fund, administered by American Chemical References and Notes els validate the feasibility of the proposed Society, for support of this research (ACS-PRF 37202- biosynthesis of the CN ligand and its trans- AC3, R.S.G. and H.W.). 2 December 2002; accepted 15 January 2003 fer from sulfur to iron (18). In sum, the cysteine residue of HypE under- goes unprecedented biotransformations in serv- ing as the site for dehydration of a carboxamido Single-Neuron Activity and moiety to CN, in carrying the CN residue, and in transferring the CN to iron. Previously, thio- Tissue Oxygenation in the cyanates were rarely encountered in nature. Thiocyanate synthesis in the marine sponge Axinyssa n. sp. has been proposed to involve Cerebral Cortex sulfuration of cyanide to give –SCN followed Jeffrey K. Thompson, Matthew R. Peterson, Ralph D. Freeman* by reaction with a sesquiterpene cation (19), although cyanation of a thiol moiety has also Blood oxygen level–dependent functional magnetic resonance imaging uses been suggested (20, 21). alterations in brain hemodynamics to infer changes in neural activity. Are these hemodynamic changes regulated at a spatial scale capable of resolving func- References and Notes tional columns within the cerebral cortex? To address this question, we made 1. P. M. Vignais, B. Billoud, J. Meyer, FEMS Microbiol. simultaneous measurements of tissue oxygenation and single-cell neural ac- Rev. 25, 455 (2001). 2. M. Frey, J. C. Fontecilla-Camps, A. Volbeda, in Hand- tivity within the visual cortex. Results showed that increases in neuronal spike book of Metalloproteins, A. Messerschmidt, R. Huber, rate were accompanied by immediate decreases in tissue oxygenation. We used K. Wieghardt, T. Poulos, Eds. (Wiley, New York, 2001), this decrease in tissue oxygenation to predict the orientation selectivity and pp. 880–896. 3. R. P. Hausinger, in Mechanisms of Metallocenter As- ocular dominance of neighboring neurons. Our results establish a coupling sembly, R. P. Hausinger, G. L. Eichhorn, L. G. Marzilli, between neural activity and oxidative metabolism and suggest that high- Eds. (Wiley-VCH, New York, 1996), pp. 1–18. resolution functional magnetic resonance imaging may be used to localize 4. M. Blokesch et al., Biochem. Soc. Trans. 30, 674 (2002). neural activity at a columnar level. 5. A. Paschos, A. Bauer, A. Zimmermann, E. Zehelein, A. Bo¨ck, J. Biol. Chem. 277, 49945 (2002). Functional magnetic resonance imaging dip exhibit sharper patches and stripes [char- 6. A. Paschos, R. S. Glass, A. Bo¨ck, FEBS Lett. 488,9 (fMRI) is a powerful tool used for the non- acteristic of orientation and ocular dominance (2001). 7. Materials and methods are available as supporting invasive mapping of neural activity in the columns (13)] relative to functional maps that material on Science Online. brain (1–3). Blood oxygen level–dependent are based on other signals (6, 8, 9). It has 8. J. C. Rain et al., Nature 409, 211 (2001). (BOLD) fMRI infers neural activity by mea- been proposed that the initial dip reflects an 9. C. Li, T. J. Kappock, J. Stubbe, T. M. Weaver, S. E. Ealick, Structure Fold Des. 7, 1155 (1999). suring small changes in deoxyhemoglobin increase in oxygen consumption by active 10. A. Paschos, S. Reissmann, A. Bo¨ck, unpublished data. within the brain’s vasculature. Therefore, the cells and is consequently better localized to 11. E. Hochleitner, S. Reissmann, F. Lottspeich, A. Bo¨ck, coupling between local cerebral hemodynam- the site of neural activity (6, 8, 9). unpublished observations. 12. G. Burkhardt, M. P. Klein, M. Calvin, J. Am. Chem. Soc. ics and the underlying neural activity is crit- Despite its potential, the existence and in- 87, 591 (1965). ically important to the interpretation and de- terpretation of the initial dip are controversial 13. J. R. Van Wazer, S. Norval, J. Am. Chem. Soc. 88, 4415 sign of fMRI studies. Increases in neural because it is not always observed using fMRI (1996). 14. Dehydration of citrullines to cyanamides with p- activity elicit a delayed decrease in deoxyhe- (14, 15), and its detection using imaging spec- toluenesulfonyl chloride and pyridine, i.e., RNH- moglobin corresponding to a positive change troscopy may be confounded by the difficulty 3 CONH2 RNHCN, was reported in a chemical sim- in the BOLD signal (1–4). Most fMRI inves- in correcting for the wavelength dependence of ulation of arginine biosynthesis and the urea cycle (23). tigations use this positive BOLD response to light scattering in tissue (16, 17). An alternative 15. For evidence supporting the relevancy of modeling map neural activity. However, because the hypothesis is that changes in blood volume, the active site of [NiFe] hydrogenase with Fp com- response is also observed from draining veins rather than oxygen consumption, could account plexes, see (24). that are displaced from the sites of activation, for the initial dip (18). Although simultaneous 16. A. Jungbauer, H. Behrens, J. Organomet. Chem. 186, 361 (1980). the spatial resolution is limited (5–7). Optical recordings of fMRI signals and neural activity 17. Although this reaction may involve nucleophilic dis- imaging and imaging spectroscopy findings were recently reported in the monkey’s primary placement by iron of thiolate at the cyano carbon atom, suggest that a fast increase in deoxyhemoglo- visual cortex (4), the analysis focused on the an electron transfer mechanism is also possible. 18. There are two possibilities for the origin of the CO bin, occurring before the delayed decrease, positive BOLD response and used stimuli that ligands: hydrolysis of a CN ligand and transfer of a may be a better indicator of neural activity (8, were designed to activate a relatively large uni- carbamoyl group to iron followed by deamination. 9). This component of the hemodynamic re- form area of the cortex. Therefore, the studies Hydrolysis of a CN ligand is analogous to the known hydrolysis of iron isocyanides to iron carbonyl com- sponse is often referred to as the “initial dip.” did not address neural and hemodynamic cou- plexes (25). Deamination of carboxamido iron com- Recent fMRI studies and direct measure- pling at a submillimeter spatial resolution. plexes is also known (26). ments of oxygen tension within the microcir- In our study, we made simultaneous colo- 19. J. S. Simpson, M. J. Garson, Tetrahedron Lett. 39, 5819 (1998). culation have also identified the initial dip (6, calized measurements of tissue oxygenation 20. A. T. Pham et al. Tetrahedron Lett. 32, 4843 (1991). 7, 10–12). In studies of the visual cortex, and single-cell neural activity in area 17 of 21. C. Jimenez, P. Crews, Tetrahedron 47, 2097 (1991). brain imaging maps that emphasize the initial the cat’s visual cortex. Tissue oxygenation is 22. P. Roepsdorff, J. Fohlman, Biomed. Mass Spectrom. linked to deoxyhemoglobin through the local 11, 601 (1984). oxygen concentration gradient in tissue and 23. D. Ranganathan, R. Rathi, J. Org. Chem. 55, 2351 (1990). Group in Vision Science, School of Optometry, Helen 24. C.-H. Lai et al., J. Am. Chem. Soc. 120, 10103 (1998). Willis Neuroscience Institute, University of California, through the oxygen-hemoglobin dissociation 25. V. Riera, J. Ruiz, J. Organomet. Chem. 384, 339 Berkeley, CA 94720–2020, USA. curve and is therefore expected to reflect the (1990). 26. L. Busetto, R. J. Angelici, Inorg. Chim. Acta 2, 391 *To whom correspondence should be addressed. E- hemodynamic changes measured with fMRI. (1968). mail: [email protected] A quantitative model of cerebral hemody-

1070 14 FEBRUARY 2003 VOL 299 SCIENCE www.sciencemag.org