Acquisition and Utilization of Transition Metal Ions by ... · Recent advances in sampling and...

Acquisition and Utilization of Transition Metal Ions by Marine OrganismsAuthor(s): Alison ButlerSource: Science, New Series, Vol. 281, No. 5374 (Jul. 10, 1998), pp. 207-210Published by: American Association for the Advancement of ScienceStable URL: http://www.jstor.org/stable/2896008 .

Accessed: 02/04/2014 17:54

Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at .http://www.jstor.org/page/info/about/policies/terms.jsp

.JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range ofcontent in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new formsof scholarship. For more information about JSTOR, please contact [email protected].

.

American Association for the Advancement of Science is collaborating with JSTOR to digitize, preserve andextend access to Science.

http://www.jstor.org

This content downloaded from 128.210.126.199 on Wed, 2 Apr 2014 17:54:29 PMAll use subject to JSTOR Terms and Conditions

http://www.jstor.org/action/showPublisher?publisherCode=aaas

http://www.jstor.org/stable/2896008?origin=JSTOR-pdf

http://www.jstor.org/page/info/about/policies/terms.jsp


CHEMISTRY AND BIOLOGY OF THE OCEANS SP EC IA L SECT IO N

Acquisition and Utilization of Transition Metal

Ions by Marine Organisms Alison Butler

REVIE W

Recent research has revealed that trace metals, particularly transition metals, play important roles in marine productivity. Most of the work has been on iron, which shows a nutrient- depleted profile in the upper ocean. Marine organisms have a variety of means for acquiring iron and other transition metal ions that differ from those of terrestrial organisms.

Metalloproteins comprise a third to a half of all known proteins. Metals function in catalysis, play structural roles, and activate biochemical processes. Many essential life processes, including photosynthesis, respiration, and nitrogen fixation, involve multi-electron transformations. The essential steps in all of these processes are catalyzed by metalloenzymes (Fig. 1, A and B). These enzymes contain iron and other transition metal ions that can exist in multiple oxidation states. Other essential life processes, such as proteolysis and the equilibration of carbon dioxide and bicarbonate ion are hydrolytic transformations that are also catalyzed by metalloenzymes. Usually these metalloenzymes contain active-site transition metal ions that do not undergo oxidation state changes [for example, Zn(II)] but which function as Lewis acid-type catalysts. The essential transition metal ions for terrestrial organisms include vanadium to zinc of the first-row transition metal series and molybdenum in the second-row series. Iron is the most abundant transition metal ion in most terrestrial organisms. Iron levels are also high in most lakes, estuaries, streams, and rivers, whereas levels of other transition metals vary widely. In contrast, the transition metal composition of the open ocean differs dramatically from that of terrestrial environments (Fig. 2) (1, 2). Recent advances in sampling and analytical techniques have permitted the metal ion composition and speciation (oxidation state and degree and type of ligation) of the ocean to be defined (1). Molybdenum is the most abundant transition metal in surface seawater at 100 nM, followed by vanadium at 20 to 35 nM (3). By contrast iron levels in surface seawater are extremely low, 0.02 to 1 nM (4-7). These metals show a nutrient-like distribution profile in that the elements are depleted in surface waters, where most primary production occurs. Despite its relative scarcity, iron is essential to marine organisms, and iron levels represent one of the key limitations in marine ecosystems. In this article, I cover some of the mechanisms by which marine organisms acquire iron and use other essential metal ions, compared to their better understood terrestrial counterparts.

We know relatively little about marine bioinorganic chemistry, but recent studies are beginning to unravel some of the mysteries. Marine microorganisms acquire iron through novel siderophores (8, 9) and new siderophore-mediated and other processes (10, 11). Phytoplank- ton substitute flavodoxin for ferredoxin in times of severe iron stress, replacing the iron-sulfur cluster with an organic cofactor (12, 13). The carbonic anhydrases (CA) of diatoms, which have no homology to other known CA, are particularly apt to substitute Co for Zn in the active site and some even contain Cd (14). Vanadium enzymes catalyze peroxidative halogenation reactions, which in most terrestrial organisms is accomplished by Fe(III)-heme enzymes (15), and tung-

Alison Butter is in the Department of Chemistry, University of California, Santa Barbara, CA 93106-9510, USA.

sten is required for catalysis of oxidoreductases in hyperthermophilic archea (16). Given the unique metal ion composition and speciation of ocean waters, it seems likely that many other bioinorganic processes that are specific to oceanic organisms remain to be discovered.

A G S-Cys

S~~~~~~~~Y-

Cy s-s-slS-Cys

F S-Cys

B C

H is 404 V0

H20

jr HiF9

H NH

H OH 0 1 N O0 OH

NH2 HN H HN O

H H 0 H HN

N

HO' 50 HO

H~~~ H OH HNOHNH2

OHO0 H 0 H 0 H CO N KNQK_AN N H

N2 H H H6"~ H COOH HO COOH

Fig. 1. Metal sites of (A) nitrogenase (Azotobacter vinlandii) [left: Cofac- tor (46); right: P-cluster (47); cluster drawings adapted from (49)]; (B) bovine heart cytochrome c oxidase (48); (C) vanadium chloroperoxidase (Curvularia inaequalis) (33); and (D) alterobactins A (1) and B (2) (8).

www.sciencemag.org SCIENCE VOL 281 10 JULY 1998 207



S PE C I AL SECT IO N CHEMISTRY AND BIOLOGY OF THE OCEANS

Iron is arguably the most prevalent metal ion among the different classes of terrestrial metalloproteins (Fig. 1, A and B). Its importance is magnified by its insolubility under the neutral pH conditions of biological systems. At pH 7, only 10-18 M aqueous Fe3+ is present in solution; yet, bacteria typically require micromolar levels of total iron for growth (17). Microorganisms have evolved an elaborate mechanism to acquire iron, which is precisely regulated and exquis- itely selective (18). Under aerobic conditions, bacteria produce siderophores to solubilize and sequester iron (III). Siderophores are low molecular weight compounds that coordinate Fe(III) with a high affinity (18). They are produced under conditions of iron demand in conjunction with their outer membrane receptor proteins. Under conditions of excess available Fe(III), the biosyntheses of siderophores and outer membrane receptor proteins are repressed.

Iron is also arguably the most important transition metal ion in the ocean, precisely because of its relatively low abundance. Surprisingly, a consensus on the solubility of ferric species in the ocean has not been reached, although the speciation of Fe(III) in surface waters is beginning to be understood (10). Most (>99%) ferric ion is complexed by organic ligands, having conditional stability constants in seawater on the order of 01 9 to 1022 M- 1 (19-21). A major challenge will be to elucidate the nature of these ligands or ligand classes. The low levels of iron have been shown to limit primary production of phytoplankton in vast regions of the world's oceans, including the subarctic Pacific, the equatorial Pacific, and the Southern Ocean (22, 23). These ocean waters are characterized by high nitrate and low chlorophyll (HNLC) levels; they are replete in major nutrients but depressed in rates of primary production, and most importantly, are deficient in iron (approximately 20 to 50 pM). Two separate iron addition experiments carried out by the IronEx I and II studies demonstrated that primary production increased when iron was added at a level of 1 to 2 nM. A surprising result of the IronEx II study was

that the concentration of the organic ligand (that is, 0.5 nM at ambient levels) increased dramatically within a day or two of the mesoscale injection of iron (to 2 nM) (24). This rapid increase raises the question whether these Fe(III) chelating ligands are siderophores or derived from siderophores. Little is known about siderophore production by eukaryotic phytoplankton, although it has been reported (25). Sid- erophore production by bacteria, phytoplankton or other marine microorganisms, could be an adaptive response of these ambient iron- deficient populations (24).

In addition to phytoplankton, heterotrophic marine bacteria have been shown to be limited by the low levels of iron in the ocean (26, 27). Heterotrophic bacteria compete successfully with phytoplankton for iron. Thus it is of tremendous interest to elucidate the molecular mechanisms that are used by oceanic bacteria and phytoplankton to acquire iron. Many oceanic bacteria have been shown to produce siderophores, although few structures have yet been determined (28- 30). We have identified aerobactin, a known terrestrial siderophore from an oceanic Vibrio species (30). We have also isolated and determined the structures of alterobactins A (1) (Fig. 1) and B (2) from Alteromonas luteoviolacea, found in oligotrophic and coastal seawater; alterobactin A has a very high thermodynamic stability constant for ferric ion (8). The conditional stability constant of the Fe(III) alterobactin A complex at pH 8 is at least equal to or greater than the ferric complex of the oceanic ligands. Thus alterobactin A can clearly function well in the ocean. While we are beginning to elucidate the siderophore-mediated uptake of some oceanic bacteria, even less is known about the molecular mechanisms phytoplankton use to sequester iron. Phytoplankton may use cell-surface reductases to obtain iron from chelated complexes, including possibly marine siderophores from other microorganisms (10, 11).

Can parallels be drawn from the behavior of iron to other metal ions? On the one hand, iron is unique in the neutral pH conditions of biological

Sc * Ti + x Cr Mn Fe Co

* @ * * * F Total *

* 0 | ' l + vlx U U

U

*'hNi 'Cu *"mZn C N SiP

* ~~~~~~~~~Inorganic' NO3 m~~~~~~~

Fig. 2. Vertical profiles of the first-row transition metal ions and selected other elements in the North Pacific Ocean. Speciation is not included. Data compiled by Y. Nozaki (2). References for plotted data include: Sc (40), Ti (41), V (3), Cr (42), Mn (43), Fe (44), Co (44), Ni (45), Cu (45), and Zn (45). For a recent

Element review, which includes speciation, see (1).

* pmol/kg * nmol/kg No jimol/kg

~mmoI/kg*

208 10 JULY 1998 VOL 281 SCIENCE www.sciencemag.org



CHEMISTRY AND BIOLOGY OF THE OCEANS S P E C I A L S E C T I O N

systems because of its poor solubility. The siderophore-mediated system for Fe(III) uptake used by microorganisms is essential because it solubi- lizes ferric hydroxide, which would otherwise be unavailable (50). But a second advantage of the siderophore system is specificity. If no other organism can recognize a particular microorganism's ferric siderophore, then the iron is reserved for that species, resulting in productive chemical warfare. On the other hand, because of the much greater solubility of other first-row transition metal ions, which are largely present in the divalent oxidation state, specific metal ligands are not required to increase the bioavailability of these metals. The consensus in terrestrial systems is that microorganisms can acquire these metal ions through other uptake pathways.

In seawater, many of the first-row transition metal ions are par- tially or fully complexed by, as yet, undefined organic ligands (1). The initial evidence points to metal-specific organic ligands, distinct from the organic ligand or ligand classes complexing Fe(III) in the ocean (31). Thus marine microorganisms living in surface seawater may have evolved specific uptake systems for metals other than iron, as an effective competition strategy to enhance acquisition of these dilute metal ions. But it is also possible that the organic complexing ligands may mitigate the toxicity of the uncomplexed, aquatic species (for example, Cu). At this point we know little about the nature of the ligands, other than the conditional stability constant of the metal- ligand complexes.

In contrast to the deficit of iron in surface seawater, vanadium is abundant, and the bioinorganic chemistry of vanadium is diverse. Most marine tunicates acquire vanadium in large quantities, but the functional significance of the sequestered vanadium has eluded in- vestigators since its discovery in seawater near the tum of the century (32). Vanadate is also important in halide metabolism in marine algae. Vanadium haloperoxidases, found in virtually all classes of marine algae, catalyze halogenation reactions and are thought to be responsible for the vast array of halogenated marine natural products (14). The x-ray structure of vanadium chloroperoxidase from Curvularia inaequalis, a fungus with ultimate marine origins which is similar to the marine vanadium bromoperoxidase, reveals a remarkably simple active site (Fig. IC) (33). The protein complexes hydrogen vanadate ion, HVO42-, with one histidine ligand. Yet vanadate ion itself, which is the second most abundant transition metal ion in seawater, does not catalyze peroxidative halogenation under ambient seawater conditions (for example, pH 8; -1-10 -IM H202; 0.5 M Cl-, -1-10 mM Br-, and -1-10 pFM I). On a molecular level we are beginning to understand the mechanism of halide oxidation and the peroxidative halogenation process (14, 33). On a global level, algal blooms produce massive quantities of volatile chlorinated and brominated hydrocarbons (34), which may lead to depletion of the Arctic ozone layer (35). However, on a functional level, the significance of these halogenated compounds is not entirely understood. In contrast to marine organisms, relatively few halogenated compounds are known in terrestrial organisms which usually have Fe(III)-heme halogenating enzymes [for example, chloroperoxidase (Caldariomyces fumago), eosinophil peroxidase, lactoperoxidase, myeloperoxidase, thyroid peroxidase, and so forth].

Seeking consequences of low iron levels in much of the world's oceans clearly inspired the exploration of the biogeochemistry of this element on a global scale and also initiated investigations into its bioinorganic chemistry at a molecular level. Insofar as it controls oceanic primary production, Fe must have a direct effect on the global C cycle (39) and may have played a role in its variations over geologic times (38). Iron availability must also affect the N cycle because it is involved in the catalysis of all nitrogen redox transformations. In particular, the reduction and assimilation of nitrate by phytoplankton and the fixation of N2 by cyanobacteria (which involves nitrogenase, Fig. 1A) may both be limited by low Fe (36, 51). The low oceanic concentration of Zn (and substituted metals Co and Cd) may have a

direct effect on the C cycle by limiting the ability of primary producers to acquire and fix inorganic C (37). The mechanistic investigations of the marine vanadium haloperoxidase, the enzyme containing vanadium has influenced global considerations of the natural halogen cycle. Many details of the underlying bioinorganic or bioorganic (that is, not involving metal ions) processes remain unknown. Most of what we know about global element cycles comes from studies of terrestrial organisms, where metalloenzymes are involved in most steps of these cycles. As the examples of iron and vanadium discussed above have shown, the bioinorganic mechanisms in marine organisms often differ from the terrestrial ones, and this influences global element cycles in new ways. Detailed studies into the bioinorganic chemistry of other transition metal ions, together with increased knowledge of the concentration and speciation of these ions in the oceans, are required to obtain a complete picture of global element cycles comprising both terrestrial and oceanic components.

References and Notes 1. J. R. Donat and K. W. Bruland, in Trace Elements in Natural Waters, B. Salbu and E.

Steinnes, Eds. (CRC Press, Boca Raton, Fl, 1995), pp. 247-281. 2. For a graphical representation of the elements in the North Pacific Ocean in the form

of the periodic table, see Y. Nozaki, Eos 78, 221 (1997). The article and periodic table can also be seen at: www.agu.org/eos-elecas97025e.html

3. R. W. Collier, Nature 309, 441 (1984). 4. K. W. Bruland, J. R. Donat, D. A. Hutchens, Limnol. Oceanogr. 36, 1555 (1991). 5. J. H. Martin, R. M. Gordon, S. E. Fitzwater, ibid., p. 1793. 6. P. M. Saager, H. J. W. de Baar, P. H. Burkill, Geochim. Cosmochim. Acta 53, 2259

(1989). 7. J. Wu, G. W. Luther l1l, Geochim. Cosmochim. Acta 60, 2729 (1996). 8. R. T. Reid, D. H. Live, D. J. Faulkner, A. Butler, Nature 366, 455 (1993). 9. B. L. Lewis et aL, Mar. Chem. 50, 179 (1995).

10. M. L. Wells, N. P. Price, K. W. Bruland, ibid., p. 157. 11. S. Soria-Dengg and U. Horstmann, Mar. Ecol. Progr. Ser. 127, 269 (1995). 12. G. J. Doucette, D. L. Erdner, M. L. Peleato, J. J. Hartman, D. M. Anderson, ibid. 30, 269

(1996). 13. J. La Roche, P. W. Boyd, R. M. L. McKay, R. J. Geider, Nature 382, 802 (1996). 14. J. G. Lee, , S. B. Roberts, F. M. M. Morel, Limnol. Oceanogr. 40, 1056 (1995); S. B.

Roberts, T. Lane, F. M. M. Morel,J. Phycol. 33, 845 (1997); D. Yee and F. M. M. Morel, Limnol. Oceanogr. 41, 573 (1996).

15. A. Butler, Curr. Opin. Chem. Biol: Bioinorg. Chem. 2, 279 (1998). 16. M. K. Johnson, D. C. Rees, M. W. W. Adams, Chem. Rev. 96, 2817 (1996). 17. C. E. Lankford, CRC Crit. Rev. Microbiol. 2, 273 (1973). 18. C. B. Matzanke, G. Muller-Matzanke, K. N. Raymond, in Iron Carriers and Iron

Proteins., T. M. Loehr, Ed. (VCH, New York, 1989), pp. 1-121. 19. M. Gledhill and C. M. G. van den Berg, Mar. Chem. 47, 41 (1994). 20. E. L. Rue and K. W. Bruland, Mar. Chem. 50, 117 (1995). 21. J. Wu and G. W. Luther IlIl, ibid., p. 159. 22. K. H. Coale et al., Nature 383, 495 (1996). 23. J. H. Martin et al., ibid. 371, 1239 (1994). 24. E. L. Rue and K. W. Bruland, Limnol. Oceanogr. 42, 901 (1997). 25. C. G. Trick, R. J. Andersen, N. M. Prince, A. Gillam, P. J. Harrison, Mar. Biol. 75, 9

(1983). 26. P. D. Tortell, M. T. Maldonado, N. M. Price, Nature 383, 330 (1996). 27. J. D. Pakulski, R. B. Coffin, C. A. Kelley, S. L. Holder, R. Downer, P. Aas, M. M. Lyons,

W. H. Jeffrey, ibid., p. 133. 28. C. G. Trick, Curr. Microbiol. 18, 375 (1989). 29. E. R. Goyne and E. J. Carpenter, Limnol. Oceanogr. 19, 840 (1974). 30. M. G. Haygood, P. D. Holt, A. Butler, ibid. 38, 1091 (1993). 31. K. W. Bruland, J. R. Donat, D. A. Hutchins, ibid. 36, 1555 (1991). 32. M. J. Smith, D. E. Ryan, K. Nakanishi, P. Frank, K. 0. Hodgson, Metal Ions Biol. Syst.

31, 423 (1995). 33. A. Messerschmidt and R. Wever, Proc. NatI. Acad. Sci. U.S.A. 93, 392 (1996); A.

Messerschmidt, L. Prade, R. Wever, Biol. Chem. 378, 309 (1997). 34. P. M. Gschwend, J. K. MacFarlane, A. Newman, Science 227, 1033 (1989). 35. R. Wever, Nature 335, 501 (1988); L. A. Barrie, J. W. Bottenheim, R. C. Schnell, P. J.

Crutzen, R. A. Rasmussen, ibid. 334, 138 (1988). 36. P. G. Falkowski, ibid. 387, 272 (1997). 37. F. M. M. Morel et al., ibid. 369, 740 (1994); F. M. M. Morel and J. R. Reinfelder, ibid.

373, 28 (1995). 38. J. H. Martin, Paleoceanography 5, 1 (1990). 39. D. J. Cooper, A. J. Watson, P. D. Nightingale, Nature 383, 511 (1996). 40. P. G. Brewer, D. W. Spencer, D. E. Robertson, Earth Planet. Sci. Lett. 16, 111 (1972).

Data are actually from the Sargasso Sea. 41. K J. Orians, E. A. Boyle, K. W. Bruland, Nature 348, 322 (1990). 42. E. Nakayama, H. Kumamoto, H. Tokoro, T. Fujinaga, ibid. 290, 768 (1981). 43. W. M. Landing and K. W. Bruland, Earth Planet. Sci. Lett. 49, 45 (1980). 44. J. H. Martin, R. M. Gordon, S. Fitzwater, W. W. Broenkow, Deep-Sea Res. 36, 649

(1989).

www.sciencemag.org SCIENCE VOL 281 10 JULY 1998 209



IS PE C IA L SECT IO N CHEMISTRY AND BIOLOGY OF THE OCEANS

45. K. W. Bruland, Earth Planet. Sci. Lett. 47, 176 (1980). 46. M. K. Chan, J. S. Kim, D. C. Rees, Science 260, 792 (1993). 47. J. W. Peters et al., Biochemistry 36, 1181 (1997). 48. T. Tsukihara et al., Science 269, 1069 (1995); figure is from S. Yoshikawa et al.,

Science 280, 1723 (1998). 49. H. Beinert, R. H. Holm, E. Munck, ibid. 277, 653 (1997). 50. H. W. Rich, F. M. M. Morel, Limnol. Oceanogr. 35, 652 (1990). 51. N. M. Price, B. A. Ahner, F. M. M. Morel, ibid. 39, 520 (1994).

52. I thank K. W. Bruland, M. G. Haygood, G. W. Luther, III, and F. M. M. Morel for valuable discussions and M. T. Simpson for assistance with graphics. Sup- ported by grants from the National Science Foundation (CHE 9529374), the National Institutes of Health (GM38130) and the AP Sloan Foundation. Partial support for my work on marine haloperoxidases and siderophores is also sponsored by NOAA, U.S. Department of Commerce under grant number NA66RGO447, project number R/MP-76 through the California Sea Grant Col- lege System and in part by the California State Resources Agency.

Climate-Ocean Variability and Ecosystem

Response in the Northeast Pacific John A. McGowan,*t Daniel R. Cayan,* LeRoy M. Dorman*

R E v I E W

The role of climatic variation in regulating marine populations and communities is not well understood. To improve our knowledge, the sign, amplitude, and frequency of climatic and biotic variations should be compared as a necessary first step. It is shown that there have been large interannual and in- terdecadal sea-surface temperature changes off the West Coast of North America during the past 80 years. Interannual anomalies appear and disappear rather suddenly and synchronously along the entire coastline. The frequency of warm events has increased since 1977. Although extensive, serial, biological observations are often incomplete, it is clear that climate-ocean variations have disturbed and changed our coastal ecosystems.

The biological consequences of climatic variability of the atmosphere and ocean are largely unknown. This is probably because of the mismatch between the scales of important atmospheric and oceano- graphic processes and the spatial and temporal dimensions of biological research programs (1). However, there is a widespread consensus that marine populations respond to climatic events and that major changes have taken place in the past 20 years in the marine ecosystems of the Pacific (2). Much of the biological, observational evidence is disconnected spatially and often discontinuous temporally, but because the potential consequences of large-scale ecosystem disturbance and disruption are uncertain and possibly detrimental, we must accept less than ideal data in our attempt to understand what is happening. Atmospheric and certain hydrographic properties are much better sampled, especially sea-level pressure (SLP) and sea- surface temperature (SST). By using these two measures, we are learning that the relation between large-scale, low-frequency climatic variability (3) and that of ecosystem and population biology is close.

Temperature variations not only affect an organism's metabolic rates directly but also influence other equally important variables such as sea level and therefore exposure of intertidal organisms, local currents and the movement of planktonic larvae, erosional regimes and therefore substrate structure, photosynthetic light intensity (cloud- iness), and water-column stratification and nutrient cycling and therefore production. These environmental variables affect population and community dynamics strongly and, over time, community structure and function. The use of departures of temperature from long-term daily or monthly means (nonseasonal anomalies) can indicate physical

J. A. McGowan and L. M. Dorman are at Scripps Institution of Oceanography, University of California, San Diego, La Jolla, CA 92093, USA. D. R. Cayan is at Scripps Institution of Oceanography and the U.S. Geological Survey, University of California, San Diego, La Jolla, CA 92093, USA.

*These authors contributed equally to this study. tTo whom correspondence should be addressed.

perturbations of the kind that act as ecosystem disturbances. Disturbance theory is well developed and is central to our under-

standing of the maintenance of community structure and patterns of diversity (4). But what types of hydrographic perturbations disturb marine ecosystems and what types do not? For example, does a single large, brief warm (cold) event have greater effects than, say, a decadal trend? We cannot answer this question because many of our concepts of the biological consequences of physical perturbations are based on brief, process-oriented studies. Large, low-frequency changes are simply not well detected by short, small-scale studies (1, 5).

There are long-term daily (since 1916) coastal SST time-series measurements over much of the Pacific coast of North America (6) and a shorter (since 1947) oceanic SST series and good, but inter- rupted, biological measurements of zooplankton, fish catch, and kelp forest communities for two to six decades. Departures from nonseasonal, long-term average SSTs (anomalies) have varied considerably between years and over decades (7). Thus, ecosystem disturbances as indexed by SST changes can be inferred back some 80 years, and their relation to basin-wide patterns of SST and SLP can be studied. Some of these low-frequency coastal temperature anomalies are connected to warm El Ninios and cool La Ninias. Particularly strong tropical events during 1957 to 1958 and 1982 to 1983 had noticeable effects on Pacific coast marine populations (8-10). Here, we used anomalies from long-term coastal SSTs to describe environmental perturbations and what is known of the biological consequences. Such knowledge will be necessary for the further development of conceptual models of marine ecosystem dynamics and of fisheries management.

Physical Changes

Interannual scales. Daily SST has been measured for decades at 17 stations along the Pacific coast (6) (Fig. 1). SSTs episodically varied from monthly, nonseasonal means over large areas by up to 3?C (Fig. 2). Large-scale heating and cooling occurred rapidly and apparently synchronously in many instances along the entire 1130 km of coastline. Many of the warm episodes lasted only a couple of months, sometimes less. Remarkable warm events associated with the 1957 to 1958 and 1982 to 1983 tropical El Nifios stand out. In both of these cases, the warming off California persisted long after the tropical signal had died out.

Conventional wisdom and a well-established theory of coastally trapped Kelvin waves (11) would lead us to expect that warming episodes should propagate poleward along the West Coast. This signal should be especially pronounced during large equatorial El Nifios. Sea-level changes do apparently progress from south to north (12), but there is no consistent evidence in the data (Fig. 2) of south to north movement of warm anomalies. Because monthly averages of SST anomalies might not resolve a south-to-north signal if Kelvin waves pass along the coast in less than 1 month, we plotted daily anomalies from long-term daily means for each station, for segments of our record during which there was a large-scale warm period (the tropical

210 10 JULY 1998 VOL 281 SCIENCE www.sciencemag.org



ARTICLES

Structure and metal exchange in the cadmiumcarbonic anhydrase of marine diatomsYan Xu1*, Liang Feng2*{, Philip D. Jeffrey2*, Yigong Shi2 & Francois M. M. Morel3

Carbonic anhydrase, a zinc enzyme found in organisms from all kingdoms, catalyses the reversible hydration of carbondioxide and is used for inorganic carbon acquisition by phytoplankton. In the oceans, where zinc is nearly depleted, diatomsuse cadmium as a catalytic metal atom in cadmium carbonic anhydrase (CDCA). Here we report the crystal structures ofCDCA in four distinct forms: cadmium-bound, zinc-bound, metal-free and acetate-bound. Despite lack of sequencehomology, CDCA is a structural mimic of a functional b-carbonic anhydrase dimer, with striking similarity in the spatialorganization of the active site residues. CDCA readily exchanges cadmium and zinc at its active site—an apparently uniqueadaptation to oceanic life that is explained by a stable opening of the metal coordinating site in the absence of metal. Giventhe central role of diatoms in exporting carbon to the deep sea, their use of cadmium in an enzyme critical for carbonacquisition establishes a remarkable link between the global cycles of cadmium and carbon.

Vertical profiles of cadmium (Cd) concentrations in the oceans showthat this metal is cycled in the water column like an algal nutrient: it isimpoverished at the surface by phytoplankton uptake and regener-ated at depth by remineralization of sinking organic matter1,2. Partof the explanation for this nutrient-like behaviour is the use of Cdas a catalytic metal atom in carbonic anhydrase (CA) in marinediatoms3–5. These organisms, which are responsible for some 40%of net marine primary production6, have adapted to life in a mediumcontaining vanishingly small concentrations of essential metals. Theexpression of a CDCA is a remarkable example of this adaptation.

CA, which catalyses the reversible hydration of carbon dioxide(CO2), was one of the first proteins for which a crystal structurewas obtained7–9. In phytoplankton, CA plays an essential part inthe acquisition of inorganic carbon for photosynthesis10,11. CA iscategorized into three main classes: a, b and c, which share no sig-nificant similarity in primary sequence or overall structure12, butwhich all rely on Zn for activity. Whereas a- and c-CA use threehistidine residues to coordinate the Zn atom, b-CA uses two cysteineresidue and one histidine residue12–16. Two new classes of CA havebeen discovered in marine diatoms, both isolated from the modelspecies Thalassiosira weissflogii: d-CA, represented by TWCA1, withsimilar enzymes identified in other classes of phytoplankton12,17,18

and an active site similar to that of a-CA19; and f-CA, representedby CDCA1, which naturally uses Cd as its catalytic metal12,20. CDCA1consists of three tandem CA repeats (R1–R3), which share 85% iden-tity in their primary sequences20. Genes coding for similar proteinshave been identified in other cultured diatoms and in natural samplesof sea water21.

Here we report the high-resolution crystal structures of CDCA1repeats and associated biochemical characterization. AlthoughCDCA1 was initially isolated as a Cd enzyme, it is actually a cambia-listic enzyme—that is, it can use either Zn or Cd for catalysis—andspontaneously exchanges the two metals. Structural analysis reveals aplausible explanation for this facile metal exchange. Though lessefficient than the Zn form, Cd-bound CDCA is a fast enzyme thatcan support the catalytic needs of fast growing diatoms.

Overall structure and active site

We crystallized the Cd-bound second repeat of CDCA1 (CDCA1-R2,residues 223–432) and determined its structure at 1.45 A resolutionusing single anomalous dispersion signal from the bound Cd atom(Supplementary Table 1). The structure of CDCA1-R2 has an overallellipsoidal shape, with seven a-helices and nine b-strands (Fig. 1a andSupplementary Fig. 1). Seven b-strands are located in the centre ofthe structure, constituting two b-sheets that appear to be contiguouswith each other. One b-sheet contains four strands (b1, b6, b8 andb9), whereas the other b-sheet comprises three strands (b3, b4 andb5). All seven a-helices are located on one side of the contiguousb-sheets. The structure can be divided into two lobes, each nucleatedby a b-sheet. One lobe (residues 223–257 and 370–432) consists ofthe four-stranded b-sheet, strands b2 and b7, and helices a1, a6 anda7. The other lobe (residues 258–369) contains the three-strandedb-sheet, and helices a2–a5. On the surface of the ellipsoid betweenthe two lobes, there is a deep cleft that leads to the active site pocket.

In the active site, Cd is coordinated by three invariant residues inCDCA of all diatom species21: Cys 263, His 315 and Cys 325 (Fig. 1band Supplementary Fig. 1). The distances between Cd and the metal-binding atoms of these residues (Sc of Cys 263, Ne of His 315 and Scof Cys 325) are 2.46 A, 2.34 A and 2.51 A, respectively (Fig. 1c, leftpanel). Notably, despite their conservation in other CDCAs, His 310and His 318 are not involved in binding to Cd. The tetrahedralcoordination of Cd is completed by a water molecule 2.35 A away.A second water molecule also contributes to Cd binding, with adistance of 2.65 A. These two water molecules are hydrogen-bondedto a third water molecule, which is connected to a number of well-ordered water molecules above the active site.

A search of the Protein Data Bank using the program Dali22 did notyield any entry with significant structural similarity; thus CDCA1appears to represent a previously unreported protein fold. Despitea lack of sequence homology and overall structure, the active siteconformation in CDCA1-R2 closely resembles that of the b-CA fromPisum sativum23 (Fig. 1b). There are five highly conserved residues atthe active site of CDCA1: three involved in coordinating Cd and an

*These authors contributed equally to this work.

1Department of Ecology and Evolutionary Biology, 2Department of Molecular Biology, 3Department of Geosciences, Princeton University, New Jersey 08544, USA. {Present address:Laboratory of Molecular Neurobiology and Biophysics, Rockefeller University, New York 10065, USA.

Vol 452 | 6 March 2008 | doi:10.1038/nature06636

56Nature Publishing Group©2008

www.nature.com/nature


www.nature.com/doifinder/10.1038/nature06636

Asp 265–Arg 267 pair (Supplementary Fig. 1). The Ca atoms ofthese five residues can be superimposed with those of b-CA23 withan r.m.s. deviation of 0.73 A. The locations of the catalytic metal ionand its tetrahedrally bound water molecule are also nearly identicalbetween CDCA1 and b-CA. In addition, the role of the Asp–Arg pairis also identical. The carboxylate side chain of Asp 265 accepts onehydrogen bond from the tetrahedrally bound water molecule andtwo hydrogen bonds from the guanidinium group of Arg 267, whichis further buttressed by a fourth hydrogen bond from the amidenitrogen of Arg 267 to the side chain of Asp 265 (Fig. 1b). CDCA1is inactivated by mutation of any of these five conserved residues(data not shown). These structural features suggest a common cata-lytic mechanism.

Mimicry of a b-CA dimer

Although CDCA1-R2 shares little structural similarity with a b-CAmonomer, it can be superimposed with a functional dimer of b-CA23

with an r.m.s. deviation of 1.93 A over 102 Ca atoms (Fig. 2a).Strikingly, the three-stranded lobe only superimposes with oneb-CA monomer whereas the four-stranded lobe aligns well withthe adjacent b-CA monomer. The functional unit of b-CA is ahomodimer23; yet each CA repeat of CDCA1 exhibits robust catalyticactivity as a monomer (see below). Thus, a single molecule ofCDCA1-R2 is a structural and functional mimic of a b-CA dimer.This structural mimicry emanates from the active site between thetwo lobes of CDCA1-R2 and becomes less apparent with increasingdistance away from the active site.

a

b c

His 315 His 315Cys 325 Cys 325Cys 263 Cys 263

Figure 1 | Structure of the second CA repeat of CDCA1 (CDCA1-R2).a, Overall structure of the Cd-bound CDCA1-R2. Two lobes of the structureare coloured blue and green. Cd is highlighted in red, and Cd-coordinatingresidues are coloured yellow. b, Comparison of the active site conformationbetween CDCA-R2 (green) andb-CA (blue). The active site residues in CDCA-R2 and b-CA are coloured yellow and orange, respectively. Hydrogen bonds

are represented by red dashed lines. Cd and the water ligand are shown in largeand small red spheres, respectively. (A stereo view is shown in SupplementaryFig. 1b.) c, Comparison of Cd- and Zn-coordination in CDCA1-R2. Metalcoordination and hydrogen bonds are indicated by red and green dashed lines,respectively; numbers indicate bond lengths in A. W1–W3, water molecules.Structural images were prepared using MOLSCRIPT44 and GRASP45.

a b

(Phe 179)(Tyr 205)

Phe 393

His 315

Cys 325

Leu 421

Thr 417β-

β-

β-

β-

Figure 2 | CDCA1-R2 is a structural mimic of a functional dimer of b-CA.a, Superposition of the CDCA1-R2 structure (green) with that of afunctional b-CA dimer. The two monomers of b-CA (b-CA-m1 and b-CA-m2) are coloured dark and light blue, respectively. For clarity, only thealigned portion of b-CA structure is shown. Zn-binding residues are shown

in the b-CA dimer to indicate the orientation of the molecules. b, A close-upview of the active site comparison between Cd-bound CDCA1-R2 (green)and a functional b-CA dimer bound to acetate23 (blue). (Stereo views areshown in Supplementary Fig. 2a and b.)

NATURE | Vol 452 | 6 March 2008 ARTICLES


Analysis of the active site conformation in CDCA1-R2 revealsdistinct contributions from the two lobes and hence provides arational basis for why CDCA1 needs to be a structural mimic of ab-CA dimer. The three metal-coordinating residues as well as theconserved Asp–Arg pair are all located in the three-stranded lobe.However, the putative substrate-binding residues are mostly con-tained within the four-stranded lobe. The residues Phe 179 andTyr 205, which directly interact with the substrate analogue acetatein b-CA23, are replaced by Phe 393, Leu 421 and Thr 417 in CDCA1-R2, all located in the four-stranded lobe (Fig. 2b). In confirmation ofthis analysis, missense mutation of any of these three residues crip-ples catalytic activity (Supplementary Fig. 2). Thus the two-lobearchitecture of CDCA1-R2, and hence its structural mimicry of ab-CA dimer, are necessitated by the catalytic activity.

Substrate binding

To support the structural analysis, we crystallized the Cd-boundCDCA1-R2 in the presence of acetate, a known substrate analogueand weak inhibitor of b-CA, and determined its structure at 1.4 Aresolution (Supplementary Table 2). As reported for b-CA23, acetateis bound at the bottom of the catalytic cleft (Supplementary Fig. 3),and is well-ordered (as judged by the electron density; Fig. 3a). Thebound acetate replaces one of the two water molecules that werehydrogen-bonded to the metal ion in the acetate-free CDCA1-R2structure (Figs 1c, 3b). Consequently, one oxygen atom of acetateis 2.47 A away from Cd, and the carboxylate moiety of acetate iswithin hydrogen-bond distances of the remaining water molecule(Fig. 3b). This water molecule is presumably deprotonated owingto Cd coordination and a hydrogen bond from Asp 265. The resultinghydroxide ion is the nucleophile that attacks the electrophilic carbonatom in CO2. The orientation of acetate is further stabilized by ahydrogen bond from the backbone amide of Phe 327.

Cd is located at the bottom of a funnel-shaped active site pocket.Intriguingly, the funnel continues to traverse the hydrophobic core ofCDCA1-R2 through an elongated channel (Fig. 3c). This channel hasan inner diameter of 4–5 A and is surrounded exclusively by 12 hydro-phobic amino acids which are conserved in CDCA (SupplementaryFig. 1): Ile 229 and Leu 233 on helix a1, Val 264 and Phe 327 on a4,Val 374 and Ile 376 on b6, Phe 393 and Val 395 on b8, and Leu 411,Ala 414, Val 418 and Leu 421 ona7. The acetate molecule is situated atone end of the hydrophobic channel, whose conserved nature suggestsfunctional significance. The channel being large enough to accom-modate a carbon dioxide molecule, we speculate that it might be theentry or escape route for the reaction substrate or product.

Facile metal exchange

Although Cd was the catalytic metal ion in CDCA1 isolated from T.weissflogii, it can be substituted by Zn, yielding an even more efficient

enzyme. In the absence of supplemental Cd in growth medium(which contains 15 mM Zn and ,3 nM Cd), Escherichia coli incorpo-rates Zn into CDCA1 and its monomers. We determined the struc-ture of Zn-bound CDCA1-R2 at 1.8 A resolution (SupplementaryTable 2). As anticipated, the overall structure remains nearly identicalto that of the Cd-bound CDCA1-R2, with an r.m.s. deviation of0.11 A over all 209 Ca atoms. However, the distances between Znand the metal-binding atoms of the three coordinating residues arereduced by approximately 0.16–0.19 A (Fig. 1c).

With increasing concentrations of Cd in the medium, Cd graduallyreplaces Zn in the enzyme until an essentially pure Cd form(Zn , 2% of Cd) is obtained at a Cd concentration of 500 mM(Fig. 4a). Either metal can also be incorporated in vitro into theapoprotein. The b-CA of the diatom Phaeodactylum tricornutum,PtCA124, which has a metal binding centre practically identical tothat of CDCA, incorporates Cd only to a small extent when over-expressed in E. coli in the presence of 500 mM Cd.

The dissociation constants of CDCA1-R2, measured by competi-tion with 4-(2-pyridylazo) resorcinol, are 10–8.9 M for Zn21 andless than 10210 M for Cd21. In the presence of Cd21 chelated withan excess of nitrilotriacetic acid (NTA), the Zn-bound CDCA1 spon-taneously exchanges Zn for Cd (Fig. 4b). But in the reciprocal experi-ment, Zn bound to NTA failed to replace Cd in the enzyme (Fig. 4c).In the presence of excess phytochelatin dimer (PC2 5 (c-Glu-Cys)2-Gly, whose thiol groups confer a high affinity for Cd21), Zn21 orCd21 was able to replace a large fraction of the other metal in theprotein within 24 h (Fig. 4b, c). Previous studies have documentedvery slow metal exchange kinetics in CAs25. The presence of someligands can greatly accelerate the removal of the metal from the activecentre of CAs, but metal re-insertion appears to necessitate an equi-molar or higher metal-to-ligand ratio26,27. We observed inefficientreplacement of Zn by Cd in PtCA1 in the presence of Cd bound toexcess NTA or PC2 (Fig. 4b). The metal-binding and metal exchangeproperties of CDCA thus appear unusual: it is a cambialistic enzymethat can incorporate either Zn or Cd as its metal centre and readilyexchange one metal for the other. As the synthesis of phytochelatins isconstitutive in diatoms and responds to very low concentrations ofCd and Zn (refs 5, 28), it seems possible that these peptides play a partin incorporating metals into CDCA in vivo.

To elucidate the structural basis for the facile metal exchange inCDCA, we determined the structure of metal-free CDCA1-R1 at1.45 A resolution (Supplementary Table 2). For comparison, we alsodetermined the structure of the Cd-bound CDCA1-R1 at 1.7 A reso-lution (Supplementary Table 2). Although most of the structuralelements are extremely well superimposed between these two struc-tures (r.m.s. deviation of 0.515 A over 202 Ca atoms), the active sitepocket takes on two distinct conformations (Fig. 5a, b). Compared tothe Cd-bound CDCA1-R1, the sequence element between the last

a b c

Cys 263

Cys 263Cys 325 Cys 325

A

Cys 325

His 315

His 315 His 315

Figure 3 | Structure of CDCA1-R2 bound to substrate analogue acetate.a, A close-up view of the region around acetate. The Fo 2 Fc electron densitymap surrounding acetate was calculated using simulated annealing with theomission of acetate and was contoured at 5s. b, A close-up view of the activesite conformation. Metal coordination and hydrogen bonds are indicated by

red and green dashed lines, respectively. Relevant distances are indicated(A). c, A hydrophobic channel traverses through CDCA1-R2. Cd ishighlighted in red. Acetate is shown in yellow. The conserved hydrophobicresidues that line the channel are shown in magenta. (Stereo views of a andb are shown in Supplementary Fig. 3b and c.)

ARTICLES NATURE | Vol 452 | 6 March 2008


two Cd-binding residues (amino acids 106–114) in the metal-freestructure has a much more open conformation. Cys 125 is trans-located from its metal-coordinating position by approximately 4 A

and undergoes a 90u rotation. In addition, the Sc atom of Cys 53 isflipped away from its metal-binding position (Fig. 5b). Thus, incontrast to other known CAs such as CAII29, the active site conforma-tion of metal-free CDCA1 is stable and different from that of themetal-bound form. Accordingly, the free energy minimum of themetal-free CDCA1-R1 corresponds to an active site conformationthat is different from that of the metal-bound form. This uniquethermodynamic characteristic probably facilitates metal exchangein CDCA1 by stabilizing the transitional, metal-free conformation.

To allow the opening of the metal-binding site, the linker sequencebetween the metal-coordinating His 105 and Cys 115 must be suffi-ciently flexible. Significantly, the two ends of this sequence are madeup of Gly residues (Supplementary Fig. 4), which are known to pro-vide flexibility to polypeptide chains30 and may serve as hinges for theconformational change. Supporting this analysis, mutation of eitherGly 316 or Gly 324 to Ala in CDCA1-R2 resulted in a loss of ability toexchange the metals (Fig. 5c).

Enzyme kinetics

We performed a preliminary kinetic analysis of CDCA1 and studiedthe rate of CO2–HCO3

2 interconversion at equilibrium by massspectrometry31. Both the CA repeats and the full-length CDCA1exhibit high CA activity with either Cd or Zn as the catalytic metal(Fig. 6 and Supplementary Fig. 5). The catalytic efficiency of thezinc enzyme is remarkably high, perhaps even higher than CAII32,approaching the diffusion limit at high pH (kcat/Km 5 8.7 3108 M21 s21 for CO2 hydration, where kcat is the turnover numberand Km the half saturation constant). The decreasing efficiency atlower pH is typical of many CAs and resembles an acidimetric titra-tion with pKa values around 7 and 9 (refs 13, 33, 34). Though lowerthan that of Zn-CDCA1, the catalytic efficiency of the Cd-CDCA1protein is still very high (kcat/Km 5 1.4 3 108 M21 s21 for CO2 hydra-tion at pH 9). The catalytic efficiency reported for Cd-substitutedCAs is generally very low, about 2% or less of the Zn enzyme atcircumneutral pH35,36. In assays of PtCA1 with partial substitutionof Cd for Zn, the reduced activity corresponded to the fraction ofZn in the enzyme. We observed no effect on CDCA1 activity of theconcentration of HEPES used as a buffer over the range 2.5 to 50 mM(data not shown).

Discussion

Collectively, the five related structures of CDCA1 we report herereveal a coherent molecular basis for the function of CDCA. Thestructural mimicry between CDCA1 and a functional b-CA dimerensures the formation of a complete active site. Structural analysisprovides a plausible explanation for spontaneous exchange of Zn andCd in the active site of CDCA1. A conserved sequence between twometal-coordinating residues adopts a stable open conformation in

a b c

Arg 57

Time (h)

Zn/

pro

tein

Cys 125

His 115Asp 55

Cys 53

-

Figure 5 | Structural basis of efficient metal exchange in CDCA1. a, Overlayof the structure of metal-free CDCA1-R1 (magenta) with that of the Cd-boundCDCA1-R1 (green). Details of the area indicated by the blue rectangle is shownin b. b, Comparison of the active site conformation between metal-freeCDCA1-R1 (magenta) and the Cd-bound CDCA1-R1 (green). (A stereo viewis shown in Supplementary Fig. 4a.) c, Effect of the mutation of Gly 316 and

Gly 324 flanking the linkage sequence between two metal binding residues,His 315 and Cys 325, on Zn replacement for Cd in Cd-bound CDCA1-R2. Datashown for wild-type (open diamonds), G316A (green triangles) and G324A(blue squares). 12mM Zn-PC2 complex was incubated with 10mM Cd-CDCA1-R2 and Cd-mutants in the presence of excess phytochelatin (PC2).Error bars (s.d.) represent duplicate measurements of a single experiment.

a

b

c

Cd

/pro

tein

Cd

/pro

tein

Zn/

pro

tein

Time (h)

Cd added (µM)

Figure 4 | Facile metal exchange in CDCA1. a, Cd incorporation in CDCA1(circles) and PtCA1 (squares) in E. coli expression system. b, In vitro Cdreplacement for Zn in Zn-CDCA1 (circles) and Zn-PtCA1 (squares) overtime. 12 mM Cd-NTA (open symbols) or Cd-phytochelatin (filled symbols)complex was incubated with 3.3 mM Zn-CDCA1 or 5 mM Zn-PtCA1. c, Invitro Zn replacement for Cd in Cd-CDCA1 over time. 12 mM Zn-NTA (opencircles) or Zn-PC2 (filled circles) complex was incubated with 3.3 mM Cd-CDCA1. In b and c, the ligands were in excess of the metals by a factor of 2(phytochelatin) or 3 (NTA). Error bars (s.d.) represent duplicatemeasurements of a single experiment except PtCA1 in a, which shows themean and s. d. from three separate experiments.



the metal-free protein which effectively lowers the free energy penaltyof releasing the bound metal and hence facilitates metal exchange.

The CDCA of diatoms readily exchanges metals at its catalyticcentre and retains better activity in the Cd form than other CAs.These properties have presumably evolved in response to the lowmetal environment of the oceans. But our kinetic data show thatthe replacement of Zn by Cd results nonetheless in a decrease incatalytic efficiency. Whereas the addition of Cd is clearly beneficialto Zn-limited laboratory cultures, increasing both in vivo CA activityand growth rate3, how effective can Cd replacement of Zn be inthe ocean where Cd is even more scarce than Zn? Phytoplanktongrowing at, say, 0.5 d21 need a photosynthetic carbon turnover rateof nearly 2 d21 (52 3 1025 s21) during daylight to support boththeir growth and their light and dark respiration. On the basis ofthe catalytic efficiency of Cd-CDCA1 (taken at kcat/Km 5 3 3

107 M21 s21, Fig. 6), and an intracellular CO2 concentration of,1 mM (taken as the half saturation constant for photosynthesis37),we estimate that a cellular CDCA1 concentration of 0.6 mmol enzymeper mol C is needed to catalyse the hydration/dehydration of allinorganic carbon fixed photosynthetically. This corresponds to acellular Cd/C ratio of 2 mmol Cd per mol C, similar to the Cd/C ratiomeasured in the phytoplankton biomass38. The replacement of Zn byCd in CDCA in the ocean can indeed satisfy a substantial fraction ofthe catalytic needs of fast growing diatoms.

The remarkable ability to make use of an element previouslyknown only for its toxicity is presumably a significant competitiveadvantage for diatoms in the metal-poor environment of the oceans.CA is a key enzyme in the carbon uptake machinery of these organ-isms, which are responsible for a large fraction of the carbon exportfrom the atmosphere to the deep ocean. The biochemical use ofCd may thus have played a part in the global radiation of diatomsduring the Cenozoic era and the concomitant decrease in atmo-spheric CO2.

METHODS SUMMARY

The various CA coding sequences were overexpressed in E. coli and purified

by metal affinity chromatography and gel filtration. Metal content of purified

protein was measured by inductively coupled plasma-mass spectrometry. The

kinetics of 18O exchange from triply labelled CO2 (13C18O18O) were followed

in a membrane-inlet mass spectrometer and the rate constants calculated as

reported31. The metal binding affinity assay was as described39 with minor

modifications. The metal exchange experiment was performed by incubating

CAs with CdY or ZnY (Y 5 NTA or PC2) with tracer 109Cd or 65Zn and then

removing aliquots for protein assay and radioactivity counting.

Various CA repeats were crystallized by the hanging-drop vapour-diffusion

method. Diffraction data were processed using the HKL suite40. The structure of

the Cd-bound CDCA1-R2 was determined by Cd-SAD (single wavelength

anomalous diffraction) using SHELX41; all other structures were solved by

molecular replacement. Model building and refinement were performed using

ARP/wARP42 and CNS43, respectively.

Full Methods and any associated references are available in the online version ofthe paper at www.nature.com/nature.

Received 3 July 2007; accepted 10 January 2008.

1. Boyle, E. A., Sclater, F. & Edmond, J. M. Marine geochemistry of cadmium. Nature263, 42–44 (1976).

2. Bruland, K. W., Knauer, G. A. & Martin, J. H. Cadmium in Northeast Pacific waters.Limnol. Oceanogr. 23, 618–625 (1978).

3. Lane, T. W. & Morel, F. M. M. A biological function for cadmium in marinediatoms. Proc. Natl Acad. Sci. USA 97, 4627–4631 (2000).

4. Morel, F. M. M. et al. Zinc and carbon co-limitation of marine phytoplankton.Nature 369, 740–742 (1994).

5. Price, N. M. & Morel, F. M. M. Cadmium and cobalt substitution for zinc in amarine diatom. Nature 344, 658–660 (1990).

6. Falkowski, P. G. et al. The evolution of modern eukaryotic phytoplankton. Science305, 354–360 (2004).

7. Fridborg, K. et al. Crystal structure of human erythrocyte carbonic anhydrase C.3.Molecular structure of enzyme and of one enzyme-inhibitor complex at 5.5 Aresolution. J. Mol. Biol. 25, 505–516 (1967).

8. Kannan, K. K. et al. Crystal structure of human erythrocyte carbonic anhydraseC.6. 3-dimensional structure at high resolution in relation to other mammaliancarbonic anhydrases. Cold Spring Harb. Symp. Quant. Biol. 36, 221–231 (1971).

9. Liljas, A. et al. Crystal structure of human carbonic anhydrase C. Nature New Biol.235, 131–137 (1972).

10. Badger, M. The roles of carbonic anhydrases in photosynthetic CO2 concentratingmechanisms. Photosynth. Res. 77, 83–94 (2003).

11. Reinfelder, J. R., Kraepiel, A. M. L. & Morel, F. M. M. Unicellular C4 photosynthesisin a marine diatom. Nature 407, 996–999 (2000).

12. Tripp, B. C., Smith, K. & Ferry, J. G. Carbonic anhydrase: New insights for anancient enzyme. J. Biol. Chem. 276, 48615–48618 (2001).

13. Cronk, J. D. et al. Identification of a novel noncatalytic bicarbonate binding site ineubacterial beta-carbonic anhydrase. Biochemistry 45, 4351–4361 (2006).

14. Mitsuhashi, S. et al. X-ray structure of beta-carbonic anhydrase from the red alga,Porphyridium purpureum, reveals a novel catalytic site for CO2 hydration. J. Biol.Chem. 275, 5521–5526 (2000).

15. Sawaya, M. R. et al. The structure of beta-carbonic anhydrase from thecarboxysomal shell reveals a distinct subclass with one active site for the price oftwo. J. Biol. Chem. 281, 7546–7555 (2006).

16. Strop, P., Smith, K. S., Iverson, T. M., Ferry, J. G. & Rees, D. C. Crystal structure ofthe ‘‘cab’’-type beta class carbonic anhydrase from the archaeonMethanobacterium thermoautotrophicum. J. Biol. Chem. 276, 10299–10305 (2001).

17. Roberts, S. B., Lane, T. W. & Morel, F. M. M. Carbonic anhydrase in the marinediatom Thalassiosira weissflogii (Bacillariophyceae). J. Phycol. 33, 845–850(1997).

18. Sotoj, A. R. et al. Identification and preliminary characterization of two cDNAsencoding unique carbonic anhydrases from the marine alga Emiliania huxleyi. Appl.Environ. Microbiol. 72, 5500–5511 (2006).

19. Cox, E. H. et al. The active site structure of Thalassiosira weissflogii carbonicanhydrase 1. Biochemistry 39, 12128–12130 (2000).

20. Lane, T. W. et al. A cadmium enzyme from a marine diatom. Nature 435, 42(2005).

21. Park, H., Song, B. & Morel, F. M. M. Diversity of the cadmium-containing carbonicanhydrase in marine diatoms and natural waters. Environ. Microbiol. 9, 403–413(2007).

22. Holm, L. & Sander, C. Protein structure comparison by alignment of distancematrices. J. Mol. Biol. 233, 123–138 (1993).

23. Kimber, M. S. & Pai, E. F. The active site architecture of Pisum sativum beta-carbonic anhydrase is a mirror image of that of alpha-carbonic anhydrases. EMBOJ. 19, 1407–1418 (2000).

24. Satoh, D., Hiraoka, Y., Colman, B. & Matsuda, Y. Physiological and molecularbiological characterization of intracellular carbonic anhydrase from the marinediatom Phaeodactylum tricornutum. Plant Physiol. 126, 1459–1470 (2001).

25. Coleman, J. E. Human carbonic anhydrase. Protein conformation and metal ionbinding. Biochemistry 4, 2644–2655 (1965).

26. Ejnik, J., Munoz, A., Gan, T., Shaw, C. F. & Petering, D. H. Interprotein metal ionexchange between cadmium-carbonic anhydrase and apo- or zinc-metallothionein. J. Biol. Inorg. Chem. 4, 784–790 (1999).

27. Pocker, Y. & Fong, C. T. O. Kinetics of inactivation of erythrocyte carbonicanhydrase by sodium 2,6-pyridinedicarboxylate. Biochemistry 19, 2045–2050(1980).

28. Ahner, B. A., Price, N. M. & Morel, F. M. M. Phytochelatin production by marinephytoplankton at low free metal ion concentrations — Laboratory studies andfield data from Massachusetts Bay. Proc. Natl Acad. Sci. USA 91, 8433–8436(1994).

29. Hakansson, K., Carlsson, M., Svensson, L. A. & Liljas, A. Structure of native andapo carbonic anhydrase II and structure of some of its anion ligand complexes.J. Mol. Biol. 227, 1192–1204 (1992).

109k c

at/K

m (M

–1 s

–1)

108

107

106

6 7 8pH

Cd-CDCA1

Zn-CDCA1

9 10

Figure 6 | pH dependence of kcat/Km for the Cd-bound (circles) and Zn-bound (squares) CDCA1. Error bars (s.d.) represent duplicatemeasurements of a single experiment.

ARTICLES NATURE | Vol 452 | 6 March 2008



30. Okoniewska, M., Tanaka, T. & Yada, R. Y. The pepsin residue glycine76contributes to active-site loop flexibility and participates in catalysis. Biochem. J.349, 169–177 (2000).

31. Silverman, D. N. Carbonic anhydrase - O18 exchange catalyzed by an enzyme withrate contributing proton transfer steps. Methods Enzymol. 87, 732–752 (1982).

32. Christianson, D. W. & Cox, J. D. Catalysis by metal-activated hydroxide in zinc andmanganese metalloenzymes. Annu. Rev. Biochem. 68, 33–57 (1999).

33. Alber, B. E. et al. Kinetic and spectroscopic characterization of the gamma-carbonic anhydrase from the methanoarchaeon Methanosarcina thermophila.Biochemistry 38, 13119–13128 (1999).

34. Fisher, S. Z. et al. Speeding up proton transfer in a fast enzyme: Kinetic andcrystallographic studies on the effect of hydrophobic amino acid substitutions inthe active site of human carbonic anhydrase II. Biochemistry 46, 3803–3813(2007).

35. Coleman, J. E. Metal ion dependent binding of sulphonamide to carbonicanhydrase. Nature 214, 193–194 (1967).

36. Tibell, L. & Lindskog, S. Catalytic properties and inhibition of Cd21-carbonicanhydrases. Biochim. Biophys. Acta 788, 110–116 (1984).

37. Burkhardt, S., Amoroso, G., Riebesell, U. & Sultemeyer, D. CO2 and HCO32 uptake

in marine diatoms acclimated to different CO2 concentrations. Limnol. Oceanogr.46, 1378–1391 (2001).

38. Kuss, J. & Kremling, K. Spatial variability of particle associated trace elements innear-surface waters of the North Atlantic (30 degrees N/60 degrees W to 60degrees N/2 degrees W), derived by large volume sampling. Mar. Chem. 68,71–86 (1999).

39. Liu, J. B., Stemmler, A. J., Fatima, J. & Mitra, B. Metal-binding characteristics of theamino-terminal domain of ZntA: Binding of lead is different compared tocadmium and zinc. Biochemistry 44, 5159–5167 (2005).

40. Otwinowski, Z. & Minor, W. Processing of X-ray diffraction data collected inoscillation mode. Methods Enzymol. 276, 307–326 (1997).

41. Sheldrick, G. A short history of SHELX. Acta Crystallogr. A 64, 112–122 (2008).

42. Perrakis, A., Morris, R. & Lamzin, V. S. Automated protein model buildingcombined with iterative structure refinement. Nature Struct. Biol. 6, 458–463(1999).

43. Brunger, A. T. et al. Crystallography & NMR system: A new software suite formacromolecular structure determination. Acta Crystallogr. D 54, 905–921 (1998).

44. Kraulis, P. J. Molscript: A program to produce both detailed and schematic plots ofprotein structures. J. Appl. Crystallogr. 24, 946–950 (1991).

45. Nicholls, A., Sharp, K. A. & Honig, B. Protein folding and association: Insights fromthe interfacial and thermodynamic properties of hydrocarbons. Proteins Struct.Funct. Genet. 11, 281–296 (1991).

Supplementary Information is linked to the online version of the paper atwww.nature.com/nature.

Acknowledgements We thank A. Saxena at the NSLS for assistance and PatrickMcGinn for help with CA assays. This work was supported by start-up funds fromPrinceton University (to Y.S.), the NSF and the NSF-funded Center forEnvironmental Bioinorganic Chemistry (to F.M.M.M.).

Author Contributions Y.X. performed all the biochemical experiments; L.F.crystallized all forms of CDCA1; P.D.J. solved the structures; and F.M.M.M. and Y.S.supervised the work and wrote the paper. All authors discussed the results andcommented on the manuscript.

Author Information Atomic coordinates have been deposited with the ProteinData Bank with the accession numbers 3BOB (R2-Cd), 3BOC (R2-Zn), 3BOE(R2-Cd-Acetate), 3BOH (R1-Cd-acetate) and 3BOJ (R1-metal free). Reprints andpermissions information is available at www.nature.com/reprints.Correspondence and requests for materials should be addressed to F.M.M.M.([email protected]) or Y.S. ([email protected]).




www.nature.com/reprints

mailto:[email protected]

mailto:[email protected]

METHODSProtein preparation. The coding sequences for CDCA1, repeat 1, 2 and 3, and

PtCA1 were amplified from T. weissflogii cDNA and P. tricornutum cDNA,

respectively. PCR products were cloned into a pET15b expression vector

(Novagen) and transformed in E. coli strain BL21(DE3). The identity of indi-

vidual clones was verified through double-stranded plasmid sequencing. The

overexpressed protein was purified from the cell lysate by metal (Ni21) affinity

chromatography and gel filtration. To produce the Cd form of CAs, 0.5 mM

CdCl2 was added into LB medium before induction by IPTG. The Zn form of

CAs was otherwise produced without Cd addition. Metal content of purifiedprotein was measured by inductively coupled plasma-mass spectrometry.

Crystallization. Purified protein from gel filtration chromatography (10 mg ml21

CDCA1 repeats in 20 mM Tris, pH 8.0, 2 mM DTT, 150 mM NaCl) was used for

crystallization. Crystals were produced using the hanging-drop vapour-diffusion

method. The Cd-bound CDCA1-R1 was crystallized in 1.2 M NaH2PO4, 1.4 M

K2HPO4, and 0.1 M sodium acetate pH 3.5. The metal-free CDCA1-R1 was crystal-

lized in 0.5 M NaH2PO4, 1.4 M K2HPO4 and 0.1 M sodium acetate pH 4.5. These

crystals were flash frozen in cryoprotectant containing 20% glycerol plus mother

liquid. The Zn-bound and Cd-bound CDCA1-R2 were crystallized in 0.1 M

sodium citrate pH 5.5 and 20% (w/v) PEG3350. Cd-loaded CDCA1-R2 bound

to acetate was crystallized in the presence of 200 mM sodium acetate with 0.1 M

sodium citrate pH 5.5 and 20% (w/v) PEG3350. Crystals were flash frozen in

cryoprotectant containing 15% glycerol plus mother liquid.

Data collection and structure determination. All data were collected either at

the X29 beam line of the National Synchrotron Light Source at Brookhaven

National Laboratory on a Quantium Q315 detector or on a Rigaku Ru3H

generator equipped with an Rigaku RAXIS-411 detector. Data were processed

using the HKL suite40 (Supplementary Table 2). CDCA1-R2 was crystallized inspace group P21212 with typical cell dimensions a 5 64.3 A, b 5 78.2 A,

c 5 37.1 A, and a 5 b 5 c 5 90u. The metal-free CDCA1-R1 was crystallized

in space group P212121 with cell dimensions a 5 46.3 A, b 5 59.2 A,

c 5 79.6 A, and a 5 b 5 c 5 90u. The Cd-bound CDCA1-R1 was crystallized in

space group C2 with cell dimensions a 5 125.3 A, b 5 43.7 A, c 5 76.9 A, and

a 5 c 5 90u, b 5 92.5u.The structure of CDCA1-R2 was determined using Cd SAD at a wavelength

(1.61 A) away from the inaccessible Cd edge, from a signal of approximately 5

electrons in 209 amino acids. The single Cd site was found and initial phases

calculated using SHELX41. The phases were further improved using SHARP and

the solvent-flattened map was calculated to a resolution of 2.0 A. The structure

was built using ARP/wARP42, using the 2.0 A solvent-flattened SAD phases for

the initial map and native (CAD-Cd) data to 1.45 A resolution. ARP/wARP built

208 of 209 residues in 100 cycles. The structure was rebuilt in O and refined using

CNS43. The Cd ion and water molecules were added to difference density in the

model-phased map. The zinc-substituted CDCA1-R2 models and the Cd-acetate

complex were determined using the Cd-bound CDCA1-R2 model as the starting

point. The Cd-bound CDCA1-R1 and metal-free CDCA1-R1 were crystallized in

different space groups to that of CDCA1-R2. In each case the structure was

solved by molecular replacement using PHASER46 and either the Cd-bound

CDCA1-R2 or metal-free CDCA1-R1 as the search models, with metal ion and

water molecules removed. Structures were rebuilt in O and refined using CNS43,

including addition of metal ions and water molecules where appropriate. Final

refinement statistics are summarized in Supplementary Table 2.

CA activity assay. The rate constants k for CO2 hydration and k9 for HCO32

dehydration catalysed by carbonic anhydrase at equilibrium were measured by18O exchange from triply labelled CO2 (13C18O18O) in a membrane-inlet mass

spectrometer. Following Silverman31, the constants k and k9 were calculated from

the slope, l, of the line ln(mass 49) versus time (corrected for the uncatalysed

blank) by the equation 2l 5 k 1 k9 2 [(k 1 k9)2 2 4/3kk9]1/2 and the ratio

k/k9 5 [HCO32]/[CO2] at the pH of interest. Assay solution contained

0.4 mM total dissolved inorganic carbon, 25 mM HEPES, 0.1 M NaSO4 at pH

8 and 20 uC unless otherwise indicated. kcat/Km was obtained from k by norma-

lizing to the molar protein concentrations using 67.5 kDa for CdCA1 and

22.8 kDa for CdCA1 R2.

Binding affinity assay. The metal binding affinity assay was as described39 with

some modifications. The buffer was chelexed to remove extraneous metal salts

and bubbled with argon gas to remove oxygen. Reduced apo-CAs were prepared

by treating proteins with 2 mM DTT and 50 mM EDTA for 6 to 8 h at room

temperature. DTT and EDTA were removed by passing through three conse-

cutive 10DG desalting columns (Bio-Rad) in an anaerobic chamber. 100mM

PAR and 15mM Cd/Zn chloride were mixed in 10 mM Tris pH 7.0 and 0.1 M

NaCl first, and then aliquots of apo-CAs were added to the mixture under argon

flow. Following each addition, the absorption spectrum between 350 nm and

650 nm was recorded in a UV-Vis spectrophotometer. The dissociation constant

was computed by Hyperquad200647.

Metal exchange. The buffer of 20 mM Tris pH 8 with 0.1 M Na2SO4 was chelexed

to remove extraneous metal salts. Before a metal exchange experiment, CdCl2 or

ZnCl2 solution was incubated with NTA at a ratio of 1:3, or with phytochelatin

(PC2) at a ratio of 1:2 in the presence of 5 mM TCEP overnight. Carrier free109Cd or 65Zn was added at tracer amount. To study Cd replacing Zn in the Zn

form of CAs, 3.3mM CDCA1, 10 mM CDCA1-R2 or mutants, and 5mM PtCA1

were incubated in the buffer and 2 mM TCEP with either 12 mM Cd-NTA in one

experiment or Cd-PC2 in another experiment at room temperature. Aliquots

were taken over time, concentrated using Microcon YM-10 (Millipore) and

resuspended in the buffer with 1 mM EDTA twice, then concentrated and resus-

pended in the buffer before protein assay and counting radioactivity. To study

Zn replacing Cd in Cd form of CAs, experiments were performed similarly.

46. McCoy, A. J., Grosse-Kunstleve, R. W., Storoni, L. C. & Read, R. J. Likelihood-enhanced fast translation functions. Acta Crystallogr. D 61, 458–464 (2005).

47. Gans, P., Sabatini, A. & Vacca, A. Investigation of equilibria in solution.Determination of equilibrium constants with the HYPERQUAD suite of programs.Talanta 43, 1739–1753 (1996).

doi:10.1038/nature06636

Nature Publishing Group©2008

www.nature.com/doifinder/10.1038/nature06636



The Biological Chemistry of the Elements: The Inorganic Chemistry of Life

J.J.R. Fraústo da Silva and R.J.P. Williams

Second Edition

The role of vanadium bromoperoxidase in the biosynthesis ofhalogenated marine natural products

Alison Butler* and Jayme N. Carter-FranklinDepartment of Chemistry and Biochemistry, University of California, Santa Barbara,CA 93106-9510, USA

Received (in Cambridge, UK) 11th November 2003First published as an Advance Article on the web 21st January 2004

Covering: 1998–2003

Halogenated natural products are frequently reported metabolites in marine seaweeds. These compounds span arange from halogenated indoles, terpenes, acetogenins, phenols, etc., to volatile halogenated hydrocarbons that areproduced on a very large scale. In many cases these halogenated marine metabolites possess biological activities ofpharmacological interest. Given the abundance of halogenated marine natural products found in marine organismsand their potentially important biological activities, the biogenesis of these compounds has intrigued marinenatural product chemists for decades. Over a quarter of a century ago, a possible role for haloperoxidase enzymeswas first suggested in the biogenesis of certain halogenated marine natural products, although this was long beforehaloperoxidases were discovered in marine organisms. Since that time, FeHeme- and Vanadium-haloperoxidases(V-HPO) have been discovered in many marine organisms. The structure and catalytic activity of vanadiumhaloperoxidases is reviewed herein, including the importance of V-HPO-catalyzed bromination and cyclizationof terpene substrates.

1 Introduction2 Vanadium haloperoxidases2.1 Characteristics of vanadium bromoperoxidase2.2 The vanadium site2.3 Proposed catalytic cycle for vanadium

bromoperoxidase3 Selectivity of vanadium bromoperoxidase3.1 Vanadium bromoperoxidase reactivity with indole

substrates3.2 Vanadium bromoperoxidase reactivity with organic

sulfides4 Biogenesis of halogenated marine natural products4.1 Biosynthesis of halogenated cyclic terpenes4.2 Biosynthesis of halogenated cyclic ethers5 Summary6 Acknowledgements7 References

1 Introduction

Halogenated natural products are frequently reported metab-olites in marine seaweeds, particularly in red macroalgae(Rhodophyceae).1–4 These compounds span a range from halo-genated indoles, terpenes, acetogenins, phenols etc., to volatile

halogenated hydrocarbons (e.g., bromoform, chloroform,dibromomethane, etc.) that are produced on a very large scale[Fig. 1].3,5,6 In many cases these halogenated marine metabolitespossess biological activities of pharmacological interest, includ-ing antifungal, antibacterial, antiviral, and anti-inflammatoryactivities. Given the abundance of halogenated marine naturalproducts found in marine organisms and their potentiallyimportant biological activities, the biogenesis of these com-pounds has intrigued marine natural product chemists fordecades. Over a quarter of a century ago, a possible role forhaloperoxidase enzymes was first suggested in the biogenesis ofcertain halogenated marine natural products,4,7 although this waslong before haloperoxidases were discovered in marine organ-isms. Since that time, haloperoxidases have been discovered inmany marine organisms, including FeHeme-vanadium haloper-oxidases,8–10 and vanadium()-bromoperoxidases,11–13 howeverthe Vanadium-dependent haloperoxidases (V-HPOs) appear tobe the most prevalent.14 V-HPOs have now been isolated, puri-fied and cloned from marine algae that produce halogenatedsecondary metabolites such as the chiral halogenated sesqui-terpenes, acetogenins, and indole derivatives. The scope ofthis review includes the reactivity of vanadium bromoperoxidase(V-BrPO) 15 and the applications of V-BrPO in the biosynthesisof halogenated marine natural products.14,16

Alison Butler

Alison Butler is a professor ofChemistry and Biochemistryat UC Santa Barbara. Herresearch interests focus onmetallobiochemistry, includingthe bioinorganic chemistry ofthe marine environment.

Jayme N. Carter-Franklin

Jayme N. Carter-Franklin isa postdoctoral fellow in theDepartment of Chemistry andBiochemistry at UC SantaBarbara. She obtained herPhD degree with Alison Butlerworking on the role ofvanadium bromoperoxidasein the biosynthesis ofhalogenated marine naturalproducts.

DO

I:1

0.1

03

9/ b

30

23

37

k

180 N a t . P r o d . R e p . , 2 0 0 4 , 2 1, 1 8 0 – 1 8 8 T h i s j o u r n a l i s © T h e R o y a l S o c i e t y o f C h e m i s t r y 2 0 0 4

2 Vanadium haloperoxidases

Vanadium haloperoxidases catalyze the oxidation of halides(iodide, bromide, chloride) by hydrogen peroxide. Vanadiumhaloperoxidases are classified according to the most electro-negative halogen oxidized. Thus vanadium chloroperoxidases(V-ClPOs) can oxidize chloride, bromide and iodide, whilevanadium bromoperoxidases (V-BrPOs) can oxidize bromideand iodide. Hydrogen peroxide does not have the driving forceto oxidize fluoride, however a fluorinating enzyme, fluorinase,has recently been isolated and is proposed to act by an SN2mechanism.17 Vanadium chloroperoxidases have been isolatedprimarily from dematiaceous hyphomycete fungi, and have yetto be isolated from marine organisms, whereas V-BrPO hasbeen isolated and characterized from all the different classes ofmarine algae, including chlorophyta (green algae), phaeophyta(brown algae), and rhodophyta (red algae).16 The overall reac-tion that vanadium haloperoxidases catalyze is:

In the first step V-HPOs catalyze the oxidation of halides(X�) by hydrogen peroxide producing a two electron oxidizedhalogen intermediate (i.e., “Br�” or biological equivalent).18 Inthe second step, the oxidized intermediate can halogenate anappropriate organic substrate or react with another equivalentof hydrogen peroxide, forming dioxygen in the singlet-excitedstate (1O2,

1∆g) [Scheme 1].19,20

2.1 Characteristics of vanadium bromoperoxidase

The X-ray crystal structures of native V-BrPO from the brownalga Ascophyllum nodosum and the red alga Corallina officinalis

Fig. 1 Examples of halogenated marine natural products.

H2O2 � X� � R–H � H� R–X � 2H2O (1)

Scheme 1 Overall reaction scheme of vanadium haloperoxidases.19,65

have been reported at 2.0 Å and 2.3 Å resolution, respect-ively.21,22 The structure of V-BrPO from A. nodosum is a homo-dimeric protein with approximate dimensions of 90 Å × 77 Å ×75 Å, and an estimated molecular weight of 120,400 Da [Fig. 2].The most recently reported structure of V-BrPO from C.officinalis is a homo-dodecameric protein (MW 740,000),measuring ∼150 Å in diameter with a single subunit dimensionof 85 Å × 56 Å × 55 Å [Fig. 3]. Both V-BrPOs are dominated byα-helical secondary structure with a few short β-strands. Themain tertiary structural motif in both enzymes is two four-helixbundles similar to the reported X-ray crystal structure ofV-ClPO from Curvularia inaequalis.23 Each V-BrPO buries ahigh percentage (e.g., 33% C. officinalis, 46% A. nodosum) of itssubunit surface in contacts with neighboring subunits givingrise to the high chemical and thermal stability reported forV-BrPOs.21,22,24–27

2.2 The vanadium site

The vanadium binding sites of V-BrPO (A. nodosum andC. officinalis) and V-ClPO (C. inaequalis) are remarkably simi-lar, but with certain notable differences. All of the structureshave a conserved protein scaffold for the vanadium-binding site.The amino acid residues that participate in vanadium binding

Fig. 2 Structure of the dimer of V-BrPO from Ascophyllum nodosum(PDB identification number, 1QI9). Figure was drawn in Swiss-PDBviewer and rendered with gl_render and POV-ray software. Vanadiumcofactors are represented as white/red stick and ball models.

Fig. 3 Structure of the dimer of V-BrPO from Corallina officinalis(PDB identification number, 1QHB). Figure was drawn in Swiss-PDBviewer and rendered with gl_render and POV-ray software. Phosphateanions are represented as gold/red stick and ball models.

181N a t . P r o d . R e p . , 2 0 0 4 , 2 1, 1 8 0 – 1 8 8

are conserved between V-BrPO and V-ClPO, and provide arigid oxyanion-binding site made up of an intricate hydrogen-bond network between the conserved residues and vanadate.16

The vanadium-binding site in V-BrPO and V-ClPO is locatedat the core of the four-helix bundle structural motif and ispositioned at the bottom of a deep funnel-shaped substratechannel, approximately 15–20 Å deep.21–23

The conserved vanadium binding site motif for V-BrPO andV-ClPO is P[S/A]YPSGHAT, where hydrogen vanadate isbound axially to a single histidine ligand in an overall trigonalbipyramidal coordination geometry.21–23 The three equatorialoxygens of vanadate are hydrogen bonded to two conservedarginines, a lysine, a serine and a glycine amide backbone. Theproposed apical hydroxide ligand is hydrogen bonded to a his-tidine [Fig. 4, shown for A. nodosum V-BrPO].21,23 Overlay ofthe X-ray coordinates for V-BrPO (A. nodosum) and V-ClPO(C. inaequalis), in addition to sequence alignments, shows thepresence of an additional histidine residue in proximity to thevanadate-binding site in V-BrPO. The additional histidineresidue in V-BrPO aligns with a phenylalanine in the V-ClPOstructure [Fig. 5]. The X-ray crystal structures of V-BrPO indi-cate that the extra histidine residue does not interact directlywith the vanadium cofactor, although it has been proposed thatthe histidine side chain participates as a proton donor andacceptor during the enzymatic reaction.21,28,29

Interestingly, comparison of the substrate channels leadingto the vanadium active-site in V-BrPO from A. nodosum andC. officinalis reveals only five residues (not including thosehydrogen bonded to the bound vanadate ion) out of the 17lining the substrate cavity that are conserved between the twostructures, and none of these five residues are conserved in theV-ClPO structure.21–23 The substrate channels leading to thevanadium binding sites are lined predominantly with hydro-phobic residues and a few hydrophilic residues.21,22 The hydro-phobic patches and few charged amino acids are proposed toinfluence substrate specificity as well as provide binding sites fororganic substrates.

2.3 Proposed catalytic cycle for vanadium bromoperoxidase

Extensive steady-state kinetic analyses of V-BrPO from A.nodosum and C. officinalis fit a substrate-inhibited bi-bi ping-pong kinetic mechanism, in which hydrogen peroxide binds firstfollowed by bromide oxidation.20,25,26,30,31 Based on these steadystate kinetic analyses and consistent with the X-ray crystalstructures of V-BrPO from A. nodosum and C. officinalis, and

Fig. 4 The vanadium site of V-BrPO (A. nodosum). Figure was drawnin Swiss-PDB viewer and rendered with gl_render and POV-raysoftware. Vanadium cofactor is represented as a gray/red stick and ballmodel.

the X-ray crystal structures of the monoperoxo-derivative ofV-ClPO, a minimum catalytic mechanism for V-BrPO has beenproposed [Fig. 6]. The bound vanadate first coordinates hydro-gen peroxide in a side-on bound fashion, producing a square-based pyramidal oxoperoxovanadium() intermediate.32 Theoxoperoxovanadium() intermediate is proposed to oxidize thesecond substrate in the cycle, bromide, by two electrons to gen-erate a bromonium ion-type intermediate.18 Detection or iso-lation of the oxidized bromine intermediate is hampered due toits reaction with organic substrates (RH, Scheme 1) or withexcess hydrogen peroxide to produce O2, preventing its build upin solution.19,20 As described below, the nature of the oxidizedbromine intermediate as either enzyme-trapped, or enzyme-bound, or a freely diffusible species seems to depend on thenature of the organic substrate.33–35

3 Selectivity of vanadium bromoperoxidase

Initial investigations into the reactivity of V-BrPOs with vari-ous substrates such as anisole or prochiral aromatic compoundsfailed to demonstrate any regio- or stereoselectivity uponbromination.36–38 The apparent lack of selectivity at that timewas interpreted to mean that V-BrPO produces a diffusibleoxidized bromine intermediate such as hypobromite, bromineor tribromide that would then carry out a molecular bromin-ation reaction.30 More recently, however, competitive kineticsstudies comparing the bromination of indole substrates by theV-BrPO/H2O2/Br� system to bromination by aqueous bromine(i.e., the equilibrium mixture of HOBr Br2 Br3

�) underidentical reaction conditions, showed that V-BrPO-catalyzedreactions were not consistent with a freely diffusible bromin-ating intermediate.33,39,40 Vanadium bromoperoxidase catalyzedthe preferential bromination and oxidation of indole derivativeswhen competed against equimolar mixtures of phenol red ormonochlorodimedone (MCD), whereas the same reactionsperformed with aqueous bromine showed the simultaneousbromination of all substrates present in solution.33,34,41

Diastereospecific bromohydrin formation by V-BrPO(C. officinalis) from (E )-4-phenyl-3-buten-2-ol has also beenbriefly reported.42 Bromination of (E )-4-phenyl-3-buten-2-olby V-BrPO, FeHeme ClPO (C. fumago) and N-bromo-acetamide (NBA) produced diastereomers of 2-bromo-1-phenylbutane-1,3-diol [Fig. 7]. The ratios of diastereomersproduced, i.e. S, R, S- (or R, S, R-) and R, S, S- (or S, R, R-),differed between the chemical and enzymatic reactions. Thecorresponding ratio for V-BrPO was 66% S, R, S- (or R, S, R-)

Fig. 5 Overlay of the X-ray coordinates of the vanadium site of wild-type V-ClPO (Curvularia inaequalis) shown in red and V-BrPO (A.nodosum) shown in yellow. Figure was drawn in Swiss-PDB viewer andrendered with gl_render and POV-ray software. Vanadium cofactor isrepresented as a purple stick and ball model. Overlap of V-ClPOphenylalanine and V-BrPO histidine is indicated.

182 N a t . P r o d . R e p . , 2 0 0 4 , 2 1, 1 8 0 – 1 8 8

Fig. 6 Proposed reaction scheme for V-BrPO catalysis.

and 34% R, S, S- (or S, R, R-), while NBA effected bromo-hydrin formation in a 75% to 25% ratio of the stereoisomers.The difference in diastereospecific product formation of theV-BrPO versus NBA is consistent with an influence from theprotein, such as substrate binding to V-BrPO and brominationtaking place at or near the active site.

3.1 Vanadium bromoperoxidase reactivity with indolesubstrates

Reactions with the indole derivative 1,3-di-tert-butylindole andV-BrPO resulted in regiospecific brominative oxidation to pro-duce 1,3-di-tert-butyl-2-indolinone as the sole product [Fig. 8].34

The formation of 1,3-di-tert-butyl-2-indolinone likely occurs byan electrophilic attack of the enzyme-generated brominatingspecies on the electron rich C2–C3 double bond of the pyrrolering producing first a bromoindolinium species, followed byhydration at the C2 position. On the other hand, reactions withaqueous bromine and 1,3-di-tert-butylindole were not regio-selective and produced many products including multiplybrominated species. Experiments with V-BrPO and 1,3-di-tert-butyl-2-indolinone as the substrate failed to produce any addi-tional brominated products, and did not compete with phenolred in competitive kinetic experiments. In contrast, bromin-

Fig. 7 V-BrPO-catalyzed diastereoselective bromohydrin formation.42

ation of 1,3-di-tert-butyl-2-indolinone with aqueous bromineproduced multiply brominated species, with brominationoccurring on the benzene ring of the indole derivative [Scheme2]. The regioselective reaction of 1,3-di-tert-butylindole withV-BrPO is consistent with organic substrate in the active sitecavity of V-BrPO. The selectivity of V-BrPO can be attributedto 1,3-di-tert-butylindole orienting specifically within the active

Fig. 8 V-BrPO-catalyzed regiospecific bromination of 1,3-di-tert-butylindole.34

Scheme 2 Bromination of 1,3-di-tert-butyl-2-indolinone by aqueousbromine and V-BrPO.34

183N a t . P r o d . R e p . , 2 0 0 4 , 2 1, 1 8 0 – 1 8 8

site cleft of V-BrPO, possibly along a hydrophobic patch ofresidues. V-BrPO-catalyzed formation of 1,3-di-tert-butyl-2-indolinone is the first example of regiospecific bromination byV-BrPO and suggests the presence of an enzyme-trapped orenzyme-bound brominating intermediate during catalysis.

3.2 Vanadium bromoperoxidase reactivity with organic sulfides

Given the tremendous reactivity of oxoperoxovanadium() com-pounds in organic oxidation reactions, it has been puzzling whyV-HPO enzymes do not carry out epoxidation, or hydroxylationreactions, in the absence of halides. Relatively recently the enan-tioselective oxidation of sulfides to sulfoxides in the absence ofhalides has been reported for V-BrPO.43–45 The concentration ofhydrogen peroxide and its rate of addition to the reaction solu-tions were found to affect both the yield and enantioselectivity ofsulfoxidation. For example, V-BrPO from A. nodosum convertsmethyl phenyl sulfide to the (R)-enantiomer (91% ee), whereasV-BrPO from C. pilulifera produces the (S )-enantiomer (55% ee)[Fig. 9].43,44,46 Kinetic resolution experiments using racemic non-aromatic cyclic thioethers with V-BrPO from A. nodosum showthat V-BrPO has a kinetic preference for conversion of the(R)-enantiomer to the corresponding sulfoxide.46 Also, V-BrPOisolated from C. officinalis catalyzes the selective sulfoxidation ofaromatic bicyclic sulfides to the (S )-sulfoxides in high (up to91%) ee [Fig. 9].45 However, the selectivity of sulfoxidation isexclusive to the marine algal V-BrPOs, since V-ClPO catalyzesthe formation of a racemic mixture of the sulfoxide. The reactiv-ity of V-BrPO (A. nodosum and C. pilulifera) to produce differingenantiomers suggests subtle differences exist in the substratechannel near the vanadium-binding site of these enzymes.

Mechanistic investigations of the source of oxygen in thesulfoxide products were established using 18O-labeled hydrogenperoxide. Experiments with V-BrPO from A. nodosum andmethyl phenyl sulfide with H2

18O2 established only 18O-labeled(R)-methyl phenyl sulfoxide produced in 63% ee.47 Thesesulfoxidation reactions demonstrated that V-BrPO catalyzesthe direct transfer of oxygen from the oxoperoxovanadium()intermediate to the sulfide, suggestive of aromatic sulfidebinding at or near the vanadium site.

Given the observed regioselective reactions in addition to thestereoselective sulfoxidation reactions catalyzed by V-BrPOs, a

Fig. 9 Sulfoxidation catalyzed by V-BrPO.43,44

picture is emerging in which subtle differences in the active sitesubstrate channel influence the outcome of regioselective andstereoselective products. The ability of V-BrPO to catalyzestereoselective reactions suggests a role for the enzyme in theproduction of halogenated marine natural products.

4 Biogenesis of halogenated marine natural products

Marine red macroalgae produce an abundance of halogenatedsecondary metabolites. Halogenated cyclic sesquiterpenes iso-lated from the red algal genus Laurencia and from herbivorousmollusks, such as Aplysia, that feed on Laurencia are one of themost prevalent structural classes of natural products isolatedfrom the algae.48,49 While their biosyntheses are still poorlyunderstood, biomimetic syntheses and enzyme-catalyzed halo-genation reactions have provided useful insight into the bio-genesis of halogenated metabolites isolated from marine algae.

4.1 Biosynthesis of halogenated cyclic terpenes

Among the halogenated natural products isolated from marinealgae, one of the largest and most varied groups are the halo-genated sesquiterpenes, which include halogenated chamigrenesisolated from the red algal genus Laurencia [Fig. 10].50 Biogene-sis studies suggest that most brominated cyclic sesquiterpenesare biosynthesized by a bromonium ion induced (or biologicalequivalent) cyclization of an acyclic terpene precursor [Scheme3].49,51 Initially biomimetic syntheses of halogenated sesqui-terpenes, carried out at a time when marine haloperoxidaseshad only been hypothesized, utilized direct bromine–carbonbond formation with successive ring closure of the terpene pre-cursor, or alternatively the indirect incorporation of bromine tothe cyclized terpene intermediate. Van Tamelen and Hesslerreported the first example of a bromonium ion induced cycliz-ation of methyl farnesate in 1966.52 In the reaction methylfarnesate was reacted with NBS in a mixture of water andtetrahydrofuran, to produce small quantities of a brominatedbicyclic ester despite the presence of competing nucleophiles.

A quarter of a century ago, Faulkner, Fenical, and Gonzalezproposed two biosynthetic routes to the basic halogenatedchamigrene skeleton from the sesquiterpene farnesyl pyro-phosphate by way of either a γ-bisabolene or a brominatedmonocyclofarnesyl intermediate [Scheme 4].4,51,53 Faulknerapplied the proposed biosynthetic scheme involving a bromin-ated monocyclofarnesol intermediate when he reported thebromonium ion induced cyclization of geranyl acetone inthe biomimetic synthesis of 10-bromo-α-chamigrene [Scheme5].7 Similary, Kato and coworkers used the brominatingagent 2,4,4,6-tetrabromocyclohexa-2,5-dienone (TBCO) in

Fig. 10 Examples of selected halogenated chamigrene naturalproducts.

Scheme 3 Hypothetical bromonium ion initiated cyclization of aterpene precursor.51

184 N a t . P r o d . R e p . , 2 0 0 4 , 2 1, 1 8 0 – 1 8 8

nitromethane to direct the cyclization of nerolidol to producethe brominated marine natural products α- and β-snyderol, and3β-bromo-8-epicaparrapi oxide [Scheme 6].54,55 In addition, thefirst enantiospecific synthesis of the brominated marine naturalproduct (�)-aplysistatin derived from (R)-(�)-malic acidinvolved using a bromonium ion induced cyclization reaction[Scheme 7].56 An important feature of the biomimetic reactionsis the lack of competing nucleophiles, such as water to quenchthe bromonium ion intermediate. While biomimetic synthesesproduced halogenated natural products in one or two steptransformations, the biomimetic reactions suffered from lowproduct yields and lack of specificity in bromine–carbon bondformation.

Based on biomimetic approaches, the bromination andcyclization of terpenes to produce cyclic halogenated naturalproducts appear to require a source of electrophilic brominein a rather non-nucleophilic environment. The active site cavityof V-BrPO possesses both of these qualities. We have begun toexplore the utility of V-BrPO in the proposed biosyntheticschemes with terpenes. The reactivity of V-BrPO was evaluatedwith geraniol, a monoterpene, and geraniol derivatives such asgeranyl acetate and geranyl acetone in mixtures of buffer andorganic cosolvents (i.e., ethanol, isopropanol, acetonitrile).35

V-BrPO reactivity with geraniol produced bromohydrins as themajor products, in addition to epoxide, dibrominated species,and two singly brominated monocyclic alcohols [Scheme 8]. The

Scheme 4 Proposed biosynthetic pathway from farnesyl pyro-phosphate to 10-bromo-α-chamigrene via a bromonium ion initiatedcyclization reaction.4,50,51

Scheme 6 Biomimetic synthesis of marine natural products α-, β-snyderol, and 3β-bromo-8-epicaparrapi oxide.54,55

brominated cyclized alcohols were produced in 1–2% yield andwere easily separated from other products. NMR spectroscopyof the cyclized geraniol products showed them to be analogousto α- and β-snyderol marine natural products.57 The mechan-ism of formation likely involves a bromonium ion attack at theC6–C7 terminal alkene of geraniol, generating a bromonium–geraniol intermediate. The proposed bromonium ion inter-mediate is subsequently attacked by the electron rich internalolefinic bond, followed by elimination reactions leading to

Scheme 5 Biomimetic synthesis of 10-bromo-α-chamigrene.7

Scheme 7 Biomimetic synthesis of the marine natural product (�)-aplysistatin.56

185N a t . P r o d . R e p . , 2 0 0 4 , 2 1, 1 8 0 – 1 8 8

products [Scheme 8].35 The brominated monocyclic prod-ucts were isolated as single diastereomers without enantio-selectivity: the NOE observed between H2 and H6 (geraniolnumbering) indicated the bromine substituent in the equatorialposition consistent with biomimetic Br�-initiated cyclizationreactions.7,55,56,58

Control reactions with aqueous bromine and geraniol, underidentical reaction conditions as V-BrPO, did not produce thebrominated monocyclic geraniol species, suggesting bromin-ation and cyclization of geraniol occurred at or near the activesite of V-BrPO.35 An enzyme–geraniol interaction to producethe brominated cyclized geraniol species was supported by theV-BrPO-catalyzed sequential bromination of first geraniolfollowed by phenol red in competitive kinetics studies [Fig. 11].In contrast, bromination of geraniol and phenol red by aque-ous bromine occurred simultaneously, suggesting that thebrominating intermediate in V-BrPO reaction with geraniol is

Scheme 8 V-BrPO-catalyzed bromination and cyclization of geraniolto the singly brominated α- and β-cyclized geraniol species.35

Fig. 11 Time course for the bromination of phenol red by (a) V-BrPO/Br�/H2O2 system and (b) aqueous bromine as a function of the geraniolconcentration. The reactions were carried out at 25 �C in the presenceof 50 µM phenol red and 40 mM KBr in 0.1 M phosphate buffer (pH5.7) with 30% v/v ethanol. Enzymatic reactions were initiated byaddition of 0.5 mM H2O2 and 5 nM V-BrPO. Nonenzymatic reactionswere initiated by addition of NaOBr (stock solution of 5 mM) in 10 µLaliquots at 30 s intervals. Production of bromophenol blue wasmonitored at 596 nm. Concentrations of geraniol: � 0 µM; � 50 µM; �100 µM.

enzyme-trapped or enzyme-bound and not a freely-diffusiblebrominating species. V-BrPO reactivity with derivatives ofgeraniol such as geranyl acetate [Scheme 9] and geranyl acetone[Scheme 10] also resulted in brominated and cyclized prod-ucts,35 similar to reactions performed by Faulkner in thebiomimetic synthesis of 10-bromo-α-chamigrene.7

4.2 Biosynthesis of halogenated cyclic ethers

In addition to the biosynthetic schemes for the production ofhalogenated sesquiterpenes, biomimetic syntheses of halogen-ated acetylenes or C15 acetogenins have also been explored.The halogenated cyclic acetogenins display features such asvaried oxane ring sizes, enyne or allenic side chains, and at leastone bromine atom.48 The lauroxocanes are the largest class ofLaurencia acetogenins and consist of the laurencin and laurea-tin structural types [Fig. 12]. The bromoether functionality inthe halogenated acetogenins has fascinated organic chemistswith respect to construction of the entropically unfavoredmedium ring ethers. Highly labile cis- and trans-laurediols iso-lated from Laurencia are assumed to be the common biogeneticprecursor of the brominated eight-membered cyclic ethers inLaurencia.59,60 Murai and colleagues in the last ten yearshave examined the cyclization of (3E )- and (3Z )-laurediolto (E )-deacetyllaurencin and (Z )-prelaureatin using a com-mercially available lactoperoxidase (LPO) and more recentlywith a crude preparation of a bromoperoxidase (BPO) isolatedfrom Laurencia nipponica [Scheme 11].61–63 Further reaction of(3E )-deacetyllaurencin with LPO or BPO produced marinenatural products laureoxanyne, laurefucin, and bromohydrinsas products, and further treatment of (Z )-prelaureatin withLPO or BPO produced laureatin, and isolaureatin [Scheme12].64 The yields of brominated cyclic products were extremelylow, but established a mechanism of bromonium ion initiatedcyclization of an acyclic precursor to the eight-memberedbromoether functionality.

Biosynthesis of the bromoether functionality utilizingV-BrPO was examined using the monoterpene nerol. V-BrPOreaction with nerol in buffered-cosolvent mixtures afforded anovel 8-membered cyclic bromoether structure, reminiscentof the oxane ring present in the halogenated acetogenins.35

Bromohydrins and epoxide products were also isolated from thereaction. The proposed mechanism for formation of the eight-membered bromoether likely involves a bromonium ion attackat on the C6–C7 terminal alkene generating a bromonium–nerol intermediate [Scheme 13]. Internal nucleophilic trappingof the bromonium-intermediate by the pendant alcohol ofnerol resulted in the eight-membered bromoether structure.Analogous to V-BrPO reaction with geraniol, the eight-membered bromoether was exclusive to the V-BrPO reactionand was not readily detected in control reactions with aqueousbromine under identical reaction conditions.

Fig. 12 Representative structures of halogenated C15 acetogeninsfrom the red algal genus Laurencia.

186 N a t . P r o d . R e p . , 2 0 0 4 , 2 1, 1 8 0 – 1 8 8

Scheme 9 V-BrPO-catalyzed bromination and cyclization of geranyl acetate.35

5 Summary

Vanadium bromoperoxidase is an enzyme present in marinealgae that catalyzes electrophilic halogenation reactions usinghydrogen peroxide to oxidize and thereby activate bromideion. The X-ray structures of V-BrPO and V-ClPO, along withinvestigations into the overall reactivity of haloperoxidases,are beginning to provide a molecular level understanding ofthe catalytic cycle of these enzymes. We have shown thatV-BrPO which has either been cloned or isolated from marine

Scheme 10 V-BrPO-catalyzed bromination and cyclization of geranylacetone.35

Scheme 11 Proposed biosynthetic scheme for the LPO- and BPO-catalyzed bromination and cyclization of laurediol to laurencin andlaureatin marine natural products.61–63

red algae can catalyze the bromination and cyclization ofterpene substrates that are likely biological precursors tohalogenated marine natural products. We have barely scratchedthe surface of the potential reactivity of V-BrPO with theseterpene substrates, as well as the role of V-BrPO in thebiogenesis of other halogenated natural products. Thus weare further investigating the features of vanadium halo-peroxidase that confer halide selectivity and organic substrateselectivity, as well as their reactivity with different terpenes

Scheme 12 Proposed biosynthetic scheme for the LPO- and BPO-catalyzed bromination of deacetyllaurencin and prelaureatin towardsthe production of marine natural products.64

187N a t . P r o d . R e p . , 2 0 0 4 , 2 1, 1 8 0 – 1 8 8

and other organic substrates, under a variety of differentconditions.

6 Acknowledgements

John Faulkner’s interest in our work on vanadium haloper-oxidase and his perceptive questions helped to focus ourinvestigations on vanadium bromoperoxidase. We have missedJohn over this past year, especially as new results emerge on ourV-BrPO-catalyzed reactions that we would wish to report tohim. We would love his feedback and especially the challenge toanswer every question he would surely ask about the productcharacterization, the reaction conditions and the mechanism ofthe reaction.

We are also grateful for funding from the NSF (CHE-0213523) and California Sea Grant (NA06RG0142, projectR/MP-94) which supports our research on the marinehaloperoxidase enzymes.

7 References

1 D. J. Faulkner, Nat. Prod. Rep., 2002, 19, 1–48.2 D. J. Faulkner, Nat. Prod. Rep., 2001, 18, 1–49.3 D. J. Faulkner, Nat. Prod. Rep., 2000, 17, 7–55.4 W. Fenical, J. Phycol., 1975, 11, 245–259.5 P. M. Gshwend, Science, 1989, 227, 1033–1035.6 G. W. Gribble, Acc. Chem. Res., 1998, 31, 141–152.7 L. E. Wolinksky and D. J. Faulkner, J. Org. Chem., 1976, 41, 597–

600.8 J. A. Manthey and L. P. Hager, Biochemistry, 1989, 28, 3052–3057.9 J. A. Manthey and L. P. Hager, J. Biol. Chem., 1985, 260, 9654–

9659.10 M. P. Roach, Y. P. Chen, S. A. Woodin, D. E. Lincoln, C. R. Lovell

and J. H. Dawson, Biochemistry, 1997, 36, 2197–2202.11 H. Vilter, Bot. Mar., 1983, 26, 451–455.12 H. Vilter, Bot. Mar., 1983, 26, 429–435.13 R. Wever and E. de Boer, Biochem. Biophys. Acta, 1985, 830, 181–

186.14 A. Butler and J. V. Walker, Chem. Rev., 1993, 93, 1937–1944.15 A. Butler, Chem. Eng. News, 2003, 81, 72.16 A. Butler, J. N. Carter and M. T. Simpson, in Handbook on

Metalloproteins; I. Bertini, A. Sigel, H. Sigel, Eds.; M. Dekker:New York, 2001, pp. 153–179.

17 D. O’Hagan, C. Schaffrath, S. L. Cobb, J. T. G. Hamilton andC. D. Murphy, Nature, 2002, 416, 279.

18 H. S. Soedjak, J. V. Walker and A. Butler, Biochemistry, 1995, 34,12689–12696.

19 R. Everett and A. Butler, Inorg. Chem., 1989, 28, 393–395.20 R. R. Everett, H. S. Soedjak and A. Butler, J. Biol. Chem., 1990, 265,

15671–15679.21 M. Weyand, H. Hecht, M. Kiesz, M. F. Liaud, H. Vitler and

D. Schomburg, J. Mol. Biol., 1999, 293, 595–611.22 M. N. Isupov, A. R. Dalby, A. A. Brindley, Y. Izumi, T. Tanabe,

G. N. Murshudov and J. A. Littlechild, J. Mol. Biol., 2000, 299,1035–1049.

23 A. Messerschmidt and R. Wever, Proc. Natl. Acad. Sci. USA, 1996,93, 392–396.

24 E. de Boer, H. Plat, M. Tromp, R. Wever, M. Franssen, H. van derPlas, E. Meijer and H. Schoemaker, Biotechnol. Bioeng., 1987, 30,607–610.

25 D. J. Sheffield, T. Harry, A. J. Smith and L. J. Rogers,Phytochemistry, 1993, 32, 21–26.

Scheme 13 V-BrPO-catalyzed bromination and cyclization of nerol tothe singly brominated eight-membered cyclic ether.35

26 J. Van Schijndel, P. Barnett, J. Roelse, E. Vollenbroek and R. Wever,Eur. J. Biochem., 1994, 225, 151–157.

27 J. V. Walker, University of California, Santa Barbara, 1998, PhDThesis.

28 G. J. Colpas, B. J. Hamstra, J. W. Kampf and V. L. Pecoraro, J. Am.Chem. Soc., 1996, 118, 3469–3478.

29 R. Renirie, W. Hemrika and R. Wever, J. Biol. Chem., 2000, 275,11650–11657.

30 E. de Boer and R. Wever, J. Biol. Chem., 1988, 236, 12326–12332.31 H. S. Soedjak and A. Butler, Biochim. Biophys. Acta, 1991, 1079,

1–7.32 A. Messerschmidt, L. Prade and R. Wever, Biol. Chem., 1997, 378,

309–315.33 R. A. Tschirret-Guth and A. Butler, J. Am. Chem. Soc., 1994, 116,

411–412.34 J. S. Martinez, G. L. Carroll, R. A. Tschirret-Guth, G. Altenhoff,

R. D. Little and A. Butler, J. Am. Chem. Soc., 2001, 123, 3289–3294.35 J. Carter-Franklin, J. D. Parrish, R. A. Tschirret-Guth, R. D. Little

and A. Butler, J. Am. Chem. Soc., 2003, 125, 3688–3689.36 N. Itoh, Y. Izumi and H. Yamada, J. Biol. Chem., 1987, 262, 11982–

11987.37 N. Itoh, A. Hasan, Y. Izumi and H. Yamada, Eur. J. Biochem., 1988,

172, 477–484.38 B. E. Krenn, Y. Izumi, H. Yamada and R. Wever, Biochem. Biophys.

Acta, 1989, 998, 63–68.39 A. Butler and R. A. Tschirret-Guth, in Mechanisms of Biohalogen-

ation and Dehalogenation; D. B. Janssen, K. Soda, R. Wever, Eds.;Royal Netherlands Academy of Arts and Sciences, Amsterdam:1996, pp. 55–68.

40 R. A. Tschirret-Guth, PhD thesis, University of California, SantaBarbara, 1996.

41 A. Butler, R. A. Tschirret-Guth and M. T. Simpson, in VanadiumCompounds: Chemistry, Biochemistry and Therapeutic Applications;A. S. Tracey, D. C. Crans, Eds.; American Chemical Society,Washington DC, ACS Symposium Series: 1998; Vol. 711, pp. 202–215.

42 P. Coughlin, S. Roberts, C. Rush and A. Willetts, Biotechnol. Lett.,1993, 15, 907–912.

43 M. Andersson, A. Willetts and S. Allenmark, J. Org. Chem., 1997,62, 8455–8458.

44 H. B. ten Brink, A. Tuynman, H. Dekker, W. Hemrika, Y. Izumi,T. Oshiro, H. E. Schoemaker and R. Wever, Inorg. Chem., 1998, 37,6780–6784.

45 M. Andersson and S. Allenmark, Tetrahedron, 1998, 54, 15293–15304.

46 H. B. ten Brink, H. L. Holland, H. E. Schoemaker, H. van Lingenand R. Wever, Tetrahedron: Asymmetry, 1999, 10, 4563–4572.

47 H. B. ten Brink, H. E. Schoemaker and R. Wever, Eur. J. Biochem.,2001, 268, 132–138.

48 K. L. Erickson, in Marine Natural Products; P. J. Scheuer, Ed.;Academic Press Inc., New York: 1983; Vol. V, pp. 131–257.

49 J. D. Martin, J. M. Palazon, C. Perez and J. L. Ravelo, Pure Appl.Chem., 1986, 58, 395–406.

50 J. D. Martin and J. Darias, in Marine Natural Products: Chemicaland Biological Perspectives; Academic Press, New York, 1978; Vol. 1,pp. 125–174.

51 D. J. Faulkner, Pure Appl. Chem., 1976, 48, 25–28.52 E. van Tamelen and E. J. Hessler, Chem. Commun., 1966, 411–413.53 A. G. Gonzalez, J. M. Aguiar, J. D. Martin and M. Norte,

Tetrahedron Lett., 1975, 2499.54 T. Kato, I. Ichinose, A. Kamoshida and Y. Kitahara, J. Chem. Soc.,

Chem. Commun., 1976, 518–519.55 T. Kato, K. Ishii, I. Ichinose, Y. Nakai and T. Kumagai, J. Chem.

Soc., Chem. Commun., 1980, 1106–1108.56 H. Shieh and G. D. Prestwich, Tetrahedron Lett., 1982, 23, 4643–

4646.57 B. M. Howard and W. Fenical, Tetrahedron Lett., 1976, 1, 41–44.58 T. R. Hoye and M. J. Kurth, J. Org. Chem., 1978, 43, 3693–3697.59 A. Fukuzawa, M. Aye and A. Murai, Chem. Lett., 1990, 1579–1580.60 A. Fukuzawa, M. Aye, M. Nakamura, M. Tamura and A. Murai,

Chem. Lett., 1990, 1287–1290.61 A. Fukuzawa, Y. Takasugi, A. Murai, M. Nakamura and

M. Tamura, Tetrahedron Lett., 1992, 33, 2017–2018.62 A. Fukuzawa, M. Aye, Y. Takasugi, M. Nakamura, M. Tamura and

A. Murai, Chem. Lett., 1994, 2307–2310.63 J. Ishihara, N. Kanoh and A. Murai, Tetrahedron Lett., 1995, 36,

737–740.64 J. Ishihara, Y. Shimada, N. Kanoh, Y. Takasugi, A. Fukuzawa and

A. Murai, Tetrahedron, 1997, 53, 8371–8382.65 A. Butler, Coord. Chem. Rev., 1999, 187, 17–35.

188 N a t . P r o d . R e p . , 2 0 0 4 , 2 1, 1 8 0 – 1 8 8

Acquisition and Utilization of Transition Metal Ions by ... · Recent advances in sampling and...

Documents

Transcript of Acquisition and Utilization of Transition Metal Ions by ... · Recent advances in sampling and...