Proc. Nat. Acad. Sci. USAVol. 72, No. 6, pp. 2217-2221, June 1975

Nuclear Magnetic Resonance Spectrum of Living Tunicate Blood Cells and theStructure of the Native Vanadium Chromogen

[contact shift/Ascidia/vanadium(III)/hemovanadin/vanadocyte]

ROBERT M. K. CARLSON

Department of Chemistry, Stanford University, Stanford, California 94305

Communicated by Carl Djerassi, March 12, 1976

ABSTRACT The 1H nuclear magnetic resonancespectrum of living tunicate blood cells was examined in anattempt to develop a biophysical assay for the nativevanadium chromogen. The living cell spectrum wasfound to exhibit a broad 21 ppm downfield Gaussian signalwhich, however, disappears immediately upon cell dis-ruption. Examination of the properties of this extremelylow field signal revealed that it corresponds to a labilevanadium(III) aquo complex contained in the cell vacu-oles, that vanadium(III) concentrations are rigidly regu-lated within these vacuoles, and that artifact formationdoes occur in the hemolysate. The living cell spectrumalso indicates the number of ligand-bound vanadium(HI)coordination sites in the native blood pigment. Resultsare discussed in relation to the possible functions of thevanadium chromogen.

The presence of a vanadium chromogen in the blood ofcertain tunicates (a group of strictly marine filter-feedinganimals related to the vertebrates) is one of the most fasci-nating and puzzling marine biochemical phenomena. Thebasic structure and function of other blood metallochromogens(hemoglobin, chlorocruorin, hemerythrin, and hemocyanin)have been established, and although substantial effort hasbeen expended in the study of the vanadium chromogen(hemovanadin), its structure and function remain contro-versial. More than 50 articles, including several reviews,concerning hemovanadin have appeared since Henze dis-covered the chromogen over 60 years ago (1-6).The uncertainty concerning the chemical structure of the

vanadium chromogen arises from its instability. The extraor-dinarily unstable green chromogen is enclosed in a 2-/Amdiameter vacuole (vanadophore) housed in an 8-pum diameterblood cell (vanadocyte) (7). Past studies indicate that thechromogen contains vanadium almost exclusively in the air-oxidizable, tripositive oxidation state and that the chromogenis maintained in a strongly acidic sulfuric acid solution whichis contained in the vanadophore (5, 6). Although livingvanadocytes possess a light green color, immediately uponlysis a dark red-brown solution is formed (Henze solution)(5). In Henze solution, vanadium(III) has been found asso-ciated with a protein and/or a small nitrogenous molecule ofunknown structure (5, 6). It seems certain that the vanado-phore encloses the vanadium chromogen-sulfuric acid solutionto protect the cell from the vacuole's strongly acidic andreducing contents and to maintain proper conditions in thevanadophore for the function of the chromogen. Opening thevanadophore in the presence of extravacuolar cytoplasmclearly constitutes a highly unphysiological condition andoffers opportunity for denaturation and artifact formation in

the hemolysate. The problem is further complicated by thefact that although tunicate blood contains numerous celltypes, only one contains the green vanadium chromogen (2, 3,7). No criteria for determining the absence of artifacts inmaterial isolated from the hemolysate have been devised.

Since past studies may have dealt with artifacts having nophysiological significance, some method of assay for thephysiologically active vanadium chromogen clearly wasneeded. Development of a functional biochemical assay forthe native vanadium chromogen would be extremely difficultbecause neither the biological function of the vanadocyte northe biochemical function of the vanadium chromogen has beenestablished. Since there is good evidence that the nativechromogen is a vanadium(III) complex and since vanadium-(III) complexes possess excellent nuclear magnetic resonance(NMR) contact shift properties (extremely short electron-spinrelaxation times and large hyperfine coupling constants), thepossible development of a biophysical assay for the physio-logically active vanadium complex was explored. The NMRsignals from protons of ligands bound to vanadium(III) ionscan be shifted many Hertz up or downfield without appre-ciable signal broadening (8-10). The NMR spectrum of livingvanadocytes, therefore, could exhibit paramagnetically shiftedsignals due to a vanadium(III) complex if the signals werenarrow enough to be observed and if the signal shifts weregreat enough to allow their resolution from the massivesignal due to intra- and extracellular water. The position andintensity of the paramagnetically shifted signals could thenbe used as an assay for the isolation and structural deter-mination of the organic ligand associated with vanadium(III)in vivo. The proton NMR spectrum of living vanadocytes wasfound to exhibit a paramagnetically shifted signal which,however, disappeared immediately upon cell lysis. The follow-ing is a report on the examination and interpretation of theproton NMR spectrum of living vanadocytes.

Living vanadocyte NMR spectrum

Healthy, uninjured specimens of the tunicate Ascidia ceratodes(Huntsman, 1912) were collected from either Monterey Bayor Tomales Bay, Calif. (locations 300 km apart) during allseasons over a 2-year period. The tubular heart was exposed,punctured, and cannulated near its anterior end and bloodwas withdrawn at a rate sufficiently slow to prevent collapseof the heart and damage to the blood cells. The proton NMRspectrum of eight packed cell preparations was taken over the2-year period. The NMA1R spectrum of each packed cellpreparation exhibited a broad signal approximately 21 ppmdownfield from sodium 2,2-dimethvl-2-silapentane-5-sulfonateand a massive water signal (Fig. ib). The mean value for the

2217

Abbreviation: NMR, nuclear magnetic resonance.

Dow

nloa

ded

by g

uest

on

Dec

embe

r 2,

202

0

Proc. Nat. Acad. Sci. USA 72 (1976)

a_~~~~~~~~~~.

b

C:--

25 24 23 22 21 20 19 18ppm

FIG. 2. NMR spectrum (measured at 60 MHz) of vanadocytepreparations undergoing time-dependent lysis. (a) 0 min, vanado-phores are intact and are light green in color; (b) 4 min, vanado-phores are beginning to lyse and the 21 ppm NMR resonanceshifts toward the water resonance and decreases in intensity;(c) 8 min, lysis continuing with concomitant NMR shifts;(d) 12 min, cell lysis complete and the lysed vanadophore solutionis dark red-brown and the 21 ppm NMR resonance is absent.

35 30 25 20 15 10 5 O ppn

FIG. 1. NMR spectra (measured at 100 MHz) of intactvanadocyte preparations of A. ceratodes and a standard vanadi-um(III) solution. (a) Spectrum of a living unpacked whole bloodsample (2.22 ml) contained in a 12 mm diameter NMR tube.Vertical arrow indicates position of weak 21 ppm vanadophoreNMR resonance. (b) Spectrum of a packed whole blood sampleprepared at 0o°4oC. Cells packed at 1000 X g in 5 mm diameterNMR tubes. Vanadophore 21 ppm peaks are numbered referringto the radio-frequency power at which the spectrum was re-

corded: (1) 119 db, (2) 113 db, (3) 107 db, (4) 77 db. The spec-

trum of the 5 mm diameter packed cell tube contained coaxiallyin a 12 mm diameter tube containing 2 ml of 98% D20 exhibiteda split 4.5 ppm water resonance due to bulk susceptibility effects,as discussed in the text. (c) Spectrum of a 0.72 molal vanadium-(III) chloride solution, 0.5 M in perchloric acid containing a

dioxane reference.

position of the low field peak for the-eight samples was foundto be 21.5 ppm, with a probable error of the mean of 0.1 ppm.The signal width at half-height was 1300-1400 Hz (measuredat 100 MHz). Line shape analysis showed that the 21 ppm

peak was Gaussian, whereas the 4.5 ppm peak was Lorentzian.To be certain that the 21 ppm peak was not an artifact of

cell packing, a spectrum of unpacked whole blood was run.

The spectrum of a sample of 2.22 ml of fresh blood was deter-mined immediately after withdrawal in a 12 mm diameter tubeusing the XL-100NMR spectrometer. The spectrum exhibiteda peak 21.7 ppm downfield from water, demonstrating thepeak's presence in the spectrum of living unpacked cells (Fig.la). A packed cell sample containing 4.8 times more cellsthan the unpacked cell sample had a 21 ppm signal intensity4.8 times greater, demonstrating that the intensity of the 21ppm signal is directly proportional to the number of cellspresent.The low field signal is observed in spectra run at both the

normal physiological temperature (1400) for A. ceratodesand the normal spectrometer operating temperature (330C).At the normal spectrometer operating temperature, the vana-

docytes in the NMR tube eventually hemolyzed, as indicatedby the contents of the NMR tube turning dark red-brownand by a drastic drop in plasma pH. Repetitive spectra of a

packed cell preparation that had been in the spectrometerfor an extended period of time revealed an interesting phe-nomenon: concurrent with lysis of the vanadophores, the 21

ppm peak shifts upfield and decreases in amplitude as itmoves (Fig. 2). The process is complete within 12 min, withthe 21 ppm signal disappearing into the water signal. The21 ppm peak is absent in the spectrum of the hemolysate.These data can be summarized as follows: (i) A 21 ppm signalcan be observed for any sample of healthy A. ceratode's blood,and the intensity of this signal is directly proportional to thenumber of blood cells present. (ii) The 21 ppm peak is quitebroad, 1300-1400 Hz wide, and is Gaussian in shape. (iii) The21 ppm signal can be observed only when intact vanadocytesare used. (iv) As the vanadophores lyse, the 21 ppm signalshifts upfield with a concurrent decrease in amplitude.

These data suggest that the 21 ppm signal in the spectrumof intact vanadocytes is due to vanadophore solvent waterprotons in rapid exchange with the protons of water in thecoordination sphere of vanadium(III) inside vanadophores.Since the 21 ppm signal is extremely broad yet of measurableamplitude, it must be due to a large number of protons in theenvironment characterized by the signal, and it is impossiblewith the quantity of vanadium present that protons of a

stable vanadium(III) complex could account for so massive a

signal. At 330C the water protons in the coordination sphereof vanadium(III) and bulk solvent water protons are in rapidexchange, and only a single weight averaged signal is observed(8, 11). Under these conditions the bulk water signal is shifteddownfield by the contact interaction of the water protonswith the unpaired electrons of the paramagnetic vanadium-(III) ion. The solution water shift, Aco, relative to pure waterfor this rapid exchange limit, is directly proportional to theconcentration of water molecules in the first coordinationsphere of vanadium(III), provided vanadium(III) is the

only paramagnetic ion present, and provided the vanadium-

(III) concentration is low compared to water. Aw is given byEq. [1]:

Ac4 = qpAM [1]

where AwM is the resonance frequency of water protons in the

first coordination sphere of the vanadium(III) ions relative to

pure water, q is the number of coordination sites free to inter-

act with solvent water molecules [maximum of six for the

hexaaquovanadium(III) ion], and p is the vanadium(III)molal concentration (12-14).As the vanadophores lyse, the vacuolar contents mix with

extravacuolar cytoplasm. This results in dilution of vana-

a

4

b 1 1-~

C

/

''j\\'

2218 Biochemistry: Carlson

Dow

nloa

ded

by g

uest

on

Dec

embe

r 2,

202

0

NMR of Living Vanadocytes 2219

dium(III), lowering p in Eq. [1], some oxidation vana-dium(III) to vanadium(IV), also lowering p, and possible com-bination of vanadium(III) with extravacuolar cytoplasmicconstituents, lowering q in Eq. [1]. Each of these processeswould result in an upfield shifting of the 21 ppm signal, as isobserved in the spectrum of packed cell preparations uponhemolysis (Fig. 2).

Since the 21 ppm peak is approximately Gaussian in shape,it possibly represents a distribution of vanadium(III) con-centrations in the vanadophores. Lysis of the vanadophoreswould result in a collapse of the vanadium(III) concentrationdistribution into a single concentration, resulting in a sig-nificant narrowing of the 21 ppm peak. Dilution of vanadium-(III) during vanadophore lysis would result in a smaller ratioof coordination sphere protons to solvent protons, while thetotal number of protons undergoing exchange would increase.Dilution, therefore, would produce a decrease in signal widthand an increase in total intensity of the 21 ppm signal as itshifted upfield. Oxidation of vanadium(III) released byhemolysis, however, would result in a broadening of the 21ppm signal, since vanadium(IV) greatly broadens water sig-nals (enhances relaxation) but does not cause appreciableparamagnetic signal shifts (15). In addition, formation ofhigh-molecular-weight vanadium(III) artifacts would alsoenhance solvent proton relaxation (16). The fact that the 21ppm signal decreases in amplitude as it shifts upfield indi-cates oxidation and/or high-molecular-weight artifact for-mation rather than a simple dilution effect as vanadophoreslyse.

Microscopic measurements compared with signal intensity

If the 21 ppm signal is attributable to water protons in rapidexchange with vanadium(III) enclosed in the vanadophores,then the intensity of the 21 ppm signal should be a measureof the amount of water in the vanadophores in the NMR tube.This proposition can be tested by comparing microscopicmeasurements of the volume of vanadophores with NMR in-tensity data. Microscopic measurements and NMR measure-ments were made with specimens of A. ceratodes taken at thesame time and location at Monterey Bay. The microscopicmeasurements are described and the results of the measure-ments are presented in the legend to Fig. 3. The NNIR packedcell sample was prepared from 10.7 ml of blood. From thedata presented in Fig. 3, the volume of the vanadophoresin the NMIR tube amounted to 0.028 ± 0.008 ml.NMIR spectral intensities were used to determine the quan-

tity of water corresponding to the intensity of the 21 ppmpeak. The volume of the NMR packed cell sample was 0.47ml. The intensity of the 21 ppm peak in the spectrum of the0.47-ml cell sample was correlated with the total spectralintensities of several 0.47-ml electrolyte samples of knownwater content. Saraf and Fatt (17), using a single coil NAIRspectrometer, reported that the NAIR signal intensity ofwater in electrolyte solutions can vary with solution con-ductivity, independent of proton content. When their ex-periments were duplicated with a cross coil NMR spec-trometer (Varian XL-100), it was found that for samples ofequal water content, the measured water signal intensitieswere equal to within 3% over the entire conductivity range.Since the 21 ppm peak was Gaussian in shape, its intensityequals [(7r/ln 2)'/2AX,/,Yo]/2 (where AX1/. is the peak widthat half-height and Y0 is the maximum peak amplitude).

_A\..l__~~~~1w .-.0w~~~~~~~~~~~~~~5

.".. 40

'uw**

#1..

FIG. 3. Vanadocyte blood cells of A. ceratodes containinggreen vacuoles (vanadophores). Phase contrast photomicrographof fresh vanadocytes, miagnification X1800. Cells (middle andright of picture) are compressed slightly under cover slip to allowaccurate determination of the number of vanadophores pervanadocyte. The average number of vanadophores per vanado-cyte was 14.0, with a probable error of the mean (Q) of 0.3 for 40measurements. The white circular area inside each vanadocyteis the cell nucleus. Vanadophore diameters were measured towithin 0.3 gsm, using a micrometer eyepiece and uncompressedvanadocytes containing approximately spherical vanadophores,and vanadophore volumes were calculated assuming sphericalshape. The average vanadophore volume was 5.0 X 10-12 cm3with a Q of 0.3 X 10-12 cm3 for 60 measurements. Total anddifferential cell counts were performed by the procedure ofVallee (26). The blood contained an average of 64 X 106 cells percm3 with a Q of 13 X 106 cells per cm3 for 25 measurements.Vanadocytes represented an average of 57% of the total numberof cell types with a Q of 1% for 10 measurements. (Photomicro-graph by Dr. K. J. Judy.)

The electrolyte water signals were Lorentzian in shape, theirintensities equaling [7rAXi. Yo]/2. Peak width measurementswere made from spectra taken using nonsaturation rf powers.Amplitude measurements of the signal were made at variousradio-frequency (rf) powers; the ratio of the slopes deter-mined by least squares analyses in the nonsaturation regionsof the plots of amplitude against square root of rf power wereused as the correct amplitude ratios. The intensity of the 21ppm signal was found to be 7.4% of the intensity of the watersignal of the electrolyte solutions containing 0.444 g of water.The 21 ppm signal in the spectrum of the packed cell sample,therefore, corresponds to 0.033 g of water.To determine the volume of vanadophores in the NAR

tube from the g of water corresponding to the 21 ppm signalwe assumed, on the basis of past studies, that most of thevanadium and sulfate present in the NMIR sample was con-tained in the vanadophores (5, 18). The NAIR sample con-tained 35.5 umol of vanadium, determined by phosphotung-state and atomic absorption analyses, and 52.5 Mmol of sul-fate, determined by volumetric analyses. The vanadophorewould then contain 0.040 g of a solution 1.08 molal in vana-dium and 1.59 molal in sulfate. A solution of these molalitieswas prepared as described in the legend of Fig. 4 and wasfound to have a density of 1.17 g/ml. If the vanadophore con-tains a solution of this density, as is supported by densitygradient measurements (unpublished data), then the volumeof the vanadophores in the NMIR tube must be 0.034 ml, avalue that correlates with the microscopic results.

Proc. Nat. Acad. Sci. USA 72 (1975)

Dow

nloa

ded

by g

uest

on

Dec

embe

r 2,

202

0

Proc. Nat. Acad. Sci. USA 72 (1975)

Ea

Ea

0

15.00

10.00 /

_.0- I,

0 0.25 0.50 0.75 1.00 1.25

[ V (m)

FIG. 4. Water NMR signal shifts in ppm as a function ofvanadium(III) molal concentrations; shifts measured relative to a

dioxane reference and referred to water. Data used in the de-termination of vanadium(III) concentration inside vanadophores:0, vanadium(III) chloride solutions in 0.3 M perchloric acid[hexaaquovanadium(III) ion] (8); A, vanadium(III) per-

chlorate solutions in 0.5 M perchloric acid, prepared fromvanadium(IIJ) sulfate solutions by precipitation of the sulfatewith a degassed barium perchlorate/perchloric acid solutionunder argon [hexaaquovanadium(III) ion]; El, vanadium(III)sulfuric acid solution (1.54 molal in sulfate) prepared by dis-solving pure vanadium metal in ultrapure concentrated sulfuricacid (ALFA) under argon, followed by filtration and dilutionwith degassed quartz-distilled water under argon; (3, vanadium(III) sulfuric acid solution (1.92 molal in sulfate) prepared as

above. The samples were prepared from measured volumes ofstock solutions of measured anion and vanadium molar concen-

trations which were diluted with appropriate acid solutions toknown volumes. Sample densities were measured and molalconcentrations calculated. Oxidation, formation of VO++, couldnot be detected from the visible absorption spectra of the samples(VO + + absorption maximum is 760 nm, VO + + molar extinctioncoefficient is 16.5) (27).

Vanadium content and the 21 ppm signal

As demonstrated above, the 21 ppm signal corresponds towater protons in rapid exchange between the solvent and thecoordination sphere of vanadium(IJJ) in the vanadophore.Unanswered, however, is how much of the vanadium presentin A. ceratodes blood is accounted for by the 21 ppm signal.Atomic absorption vanadium analysis, volumetric sulfateanalyses, and NMR vanadophore water content analysisindicate maximum vanadium(JII) and sulfate molal concen-

trations of 1.08 molal and 1.59 molal, respectively. Since thevanadophore signal shift magnitude (21.5 ppm) is related tothe molal concentration of vanadium(JII) in the vanado-phore (Eq. [1]), it should be possible to determine the va-

nadium(III) molal concentration in the vanadophore by com-

parison with the bulk resonance shifts of water in the spectraof suitable vanadium(III) standard solutions. Comparisonof the vanadium(IJJ) molal concentration determined fromatomic absorption and NMR intensity data with NMR shiftdata would indicate how much vanadium is accounted forby the 21 ppm vanadophore signal, and would give the maxi-mum number of vanadophore vanadium(III) coordinationsites interacting with ligands other than water in vivo.

Paramagnetic bulk susceptibility effects can cause sig-nificant contributions to resonance shifts measured relativeto an external reference (19). The position of the vanado-phore 21 ppm resonance was measured relative to the plasmawater resonance, the plasma water being an external refer-ence, i.e., not enclosed in the vanadophores. The spectrum of

coaxial 5 mm and 12 mm diameter tubes containing packedblood cells and 98% deuterium oxide, respectively, showed asplit water resonance (Fig. lb). The larger plasma watersignal is shifted upfield relative to the proton signal of waterin 98% deuterium oxide due to the presence of paramagneticvanadium(JII) in the vanadophores of the packed cell sample.The plasma water, therefore, is not a reference external to thevanadophores with respect to bulk susceptibility effectssince the plasma water resonance position is affected by theparamagnetic susceptibility of vanadium(JII) contained in thevanadocytes.

Since bulk susceptibility effects depend primarily on sampleand magnet geometries (20), spectra of identical packed cellpreparations were determined on spectrometers of differentmagnet geometries (100 MHz and 300 MHz spectrometers),and the resonance position of the vanadophore water signalwas found to be 21.5 ppm for both. The bulk susceptibilitycorrection that must be applied to the vanadophore waterresonance when measured relative to the plasma water reso-nance, therefore, is negligible because, although different bulksusceptibility shifts are expected for spectrometers of differ-ent magnet geometries, no shift difference was observed; theplasma water, therefore, functions as an internal reference.

Standard curves of water resonance shifts against vana-dium(III) concentrations in perchloric acid and sulfuric acidsolutions were prepared (Fig. 4). The shifts for the vanadium-(III) solutions were measured relative to an internal dioxanereference and referred to pure water. The difference in slopein Fig. 4 for vanadium(III) water resonances in sulfate andperchlorate media indicates complexation of sulfate (14).The molal concentration of vanadium(III) corresponding tothe 21 ppm vanadophore signal can be evaluated from Fig. 4.Several important assumptions were made in evaluating the21 ppm signal in terms of Fig. 4: (i) The solvent water pro-tons are in rapid exchange with the water protons in the firstcoordination sphere of vanadium(III) ions enclosed in thevanadophores. Vanadophore water resonance shifts measuredabove 200C follow the Curie law predicted for the fast ex-change conditions. (ii) Vanadium(III) is the only paramag-netic ion that occurs at a significant concentration in thevanadophore. Iron has been found to be the only paramagneticmetal ion other than vanadium that occurs in significantquantities in A. ceratodes blood (21). Atomic absorption ironand vanadium analyses of three packed cell digests of A.ceratodes prepared over a 6-month period, however, revealedan iron concentration less than 2% of the vanadium concen-tration for all three samples. There is no evidence that theiron is contained in vanadophores. (iii) AwM of the vanado-phore vanadium(JII) aquo-complex equals ACM of the vana-dium(III) sulfuric acid aquo-complex used in making themeasurements presented in Fig. 4. Similar assumptions con-cerning variations in AwM for other transition metal com-plexes have been made and discussed in past studies (14).,AM has been directly measured for the hexaaquovanadium-(III) complex (11) and can, in principle, be measured for thevanadium(III) aquo-complexes of both vanadium(III) sul-fate solutions and vanadophores. Preliminary low-tempera-ture (2100K), wide-line NMR studies of frozen vanadocytes(unpublished data) support this assumption.The 21.5 ppm vanadophore water resonance is 17.1 ppm

downfield from the plasma water resonance and 15.5 ppmdownfield from the proton resonance of a 1.54 molal sulfuric

Xii

2220 Biochemistry: Carlson

Dow

nloa

ded

by g

uest

on

Dec

embe

r 2,

202

0

NMR of Living Vanadocytes 2221

acid solution. A 15.5 ppm shift corresponds to a 0.65 molalhexaaquovanadium(III) ion solution or a 0.85 molal vana-dium(III) sulfate solution (Fig. 4). The weight of waterpresent in the vanadophores of the NMR sample was 0.033g, which corresponds to a total of 21 Mmol of vanadium for a0.65 molal solution and 28 Mmol of vanadium for a 0.85 molalsolution. The NMR sample contained a total of 35,mol ofvanadium. Twenty-one and 28 umol of vanadium are 60%and 80% of 35 Mmol, respectively. If all the vanadium in thepacked cell sample was present as vanadium(III) in thevanadophores and free to bind sulfate ions (5), then 80%corresponds to five of the six possible vanadium(III) coordi-nation sites, i.e., a maximum of one coordination site pervanadium(III) ion would be left available for binding an

organic ligand. Occurrence of unreduced vanadium in possiblevanadocyte stem cells, as demonstrated by electron micro-scopic studies (18), would further reduce the number ofavailable vanadium(III) coordination sites. The picture oflimited complexation of vanadium(III) in vanadophoresagrees. with the hypothesis developed by Bielig et al. (5).Limited complexation of vanadium(III) ions enclosed inliving vanadocytes does not indicate that the vanadophorerepresents an inactive store of vanadium because: (i) statisti-cal analyses of the 21 ppm peak shift variation interpretedusing Fig. 4 indicate that the vanadium(III) concentration isstrictly regulated in the vanadophore, the mean concentra-tion varying only 0.02 molal from animals of greatly varyingorigins, and (ii) 21 ppm peak widths indicate that within anysingle sample of blood the vanadium(III) concentrationrange of the vanadophores is sharply limited. Since it is im-probable that the tunicate would regulate the concentrationof a stored inactive compound so rigidly, it is reasonable toassume that the vanadophore is a functional unit.

Function of the vanadium chromogenTunicate blood hemolysate does not oxygenate reversibly,and it has long been thought that the vanadium chromogenis not involved in oxygen transport (22). Carlisle has shown,however, that intact living tunicate blood cells can act asoxygen storage and transport units (6). The present studyhas demonstrated that artifact formation does occur in thehemolysate, and it is reasonable to assume, therefore, thatvanadocytes possess properties that cannot be duplicated byconventional hemolysates. It has been alternatively proposedthat the vanadocyte (i) transports, stores, and processesnutrients (23); (ii) is involved in the production of the poly-saccharide outer covering of the animal (7), a theory con-tested by labeled glucose studies (24), or (iii) is involved inthe antimicrobial defense mechanism (25). These hypothesesare based primarily on histological observations and do notsuggest a biochemical reason for the presence of a concen-trated acidic vanadium(III) sulfate solution in the vanado-phore. It has been proposed that the vanadium chromogenfunctions in the production of acid for an unknown purpose(7, 21). However, the blood of certain tunicates containshighly acidic large vesicular cells that are devoid of the va-nadium chromogen (2, 3). It is generally felt that the reduc-ing properties of the vanadium chromogen are the key to itsfunction (6).

This study demonstrates that the native vanadium chromo-gen enclosed in living vanadocytes is essentially a vanadium-

(III) solution with a rigidly regulated concentration. Thesefindings suggest that unlike the metalloprotein macromole-cules of the other blood metallochromogens, it may be neces-sary to look at the vanadophore vacuole, and perhaps thevanadocyte as the smallest functional unit. The findings fur-ther suggest that conventional biochemical procedures usingthe disruption of whole blood cells by methods that resultin significant alterations in the vanadium(III) concentrationand that offer the opportunity for artifact formation withextravacuolar cytoplasm are likely to yield spurious resultsthat do not reflect the true physiological function of the nativechromogen in the intact cells.

I gratefully acknowledge the advice, support, and encourage-ment from Dr. David H. Rammler and the Molecular BiologyInstitute of Syntex Research. Sincere thanks go to Dr. Lois J.Durham for running the 100 MHz spectra and for advice andassistance in all phases of this study. I also thank Dr. Donald P.Abbott, Dr. M. L. Maddox, Dr. Kenneth J. Judy, Dr. Robert E.Linder, Dr. Norma G. Jardetzky, and Eric H. Meier who gavevaluable advice and assistance in the various phases of thestudy. I thank Dr. Stephen H. Smallcombe of Varian Associatesfor running the 300 MHz spectrum, Dr. Jules A. Tabris forcriticism of the manuscript, and Dr. Harden M. McConnell forvaluable discussions and for reviewing the final results. This workwas supported by Grant GM-06840 from the National Institutesof Health to Dr. Carl Djerassi. My warm appreciation goes toCarl Djerassi for making this study possible.1. Henze, M. (1911) Hoppe-Seyler's Z. Physiol. Chem. 72,

494-501.2. Webb, D. A. (1939) J. Exp. Biol. 16, 499-523.3. Webb, D. A. (1956) Pubbl. St. Zool. Napoli 28, 273-288.4. Bertrand, D. (1950) Bull. Am. Mus. Nat. Hist. 94, 403-455.5. Bielig, H. J., Bayer, E., Dell, H. D., Rohns, G., Mollinger,

H. & Rudiger, W. (1966) Protides Biol. Fluids 14, 197-204.6. Carlisle, D. B. (1968) Proc. Roy. Soc. Ser. B. 171, 31-42.7. Endean, R. (1960) Quart. J. Microsc. Sci. 101, 177-197.8. Rohrscheid, F., Ernest, R. E. & Holm, R. H. (1967) Inorg.

Chem. 6, 1315-1320.9. Rohrscheid, R., Ernst, R. E. & Holm, R. H. (1967) Inorg.

Chem. 6, 1607-1613.10. Rohrscheid, F., Ernst, R. E. & Holm, R. H. (1967) J.

Am. Chem. Soc. 89, 6472-6477.11. Chmelnick, A. M. & Fiat, D. (1972) J. Magn. Reson. 7,

325-331.12. Swift, T. J. (1973) in NMR of Paramagnetic Molecules:

Principles and Applications, eds. Lamar, G. N., Horrocks,W. DeW. & Holm, R. H. (Academic Press, New York andLondon), pp. 53-83.

13. Swift, T. J. & Connick, R. E. (1962) J. Chem. Phys. 37,307-320.

14. Luz, Z. & Shulman, R. G. (1965) J. Chem. Phys. 43, 3750-3756.

15. Reuben, J. & Fiat, D. (1969) J. Am. Chem. Soc. 91, 4652-4656.

16. Mildvan, A. S. & Cohn, M. (1970) Adv. Enzymol. 33, 1-70.17. Saraf, D. N. & Fatt, I. (1967) Nature 214, 1219-1220.18. Overton, J. (1966) J. Morphol. 119, 305-326.19. Dickinson, W. C. (1951) Phys. Rev. 81, 717-731.20. Live, D. H. & Chan, S. I. (1970) Anal. Chem. 42, 791-792.21. Swinehart, J. H., Biggs, W. R., Halko, D. J. & Schroeder,

N. C. (1974) Biol. Bull. (Woods Hole, Mass.) 146, 302-312.22. Winterstein, H. (1909) Biochem. Z. 19, 384-424.23. Andrew, W. (1965) Comparative Hematology (Grune and

Stratton, Inc., New York).24. Deck, J. D., Hay, E. D. & Revel, J. P. (1966) J. Morphol.

120, 267-280.25. Brown, A. C. & Davies, A. B. (1971) J. Invertebr. Pathol. 18,

276-279.26. Vallee, J. A. (1967) Bull. S.C. Acad. Sci. 66, 23-28.27. Baker, B. R., Sutin, N. & Welch, T. J. (1967) Inorg. Chem.

6, 1948-1951.

Proc. Nat. Acad. Sci. USA 72 (1975)

Dow

nloa

ded

by g

uest

on

Dec

embe

r 2,

202

0