Eur. J. Biochem.

243, 85-92 (1997)

0 FEBS 1997

Quaternary structure of the extracellular haemoglobin

of the lugworm Arenicola marina
A multi-angle-laser-light-scatteringand electrospray-ionisation-mass-spectrometry analysis
Franck ZAL', Brian N GREEN', Franqois H LALLIER', Serge N VINOGRADOV' and Andrk TOULMOND'
' Equipe Ecophysiologie, Station Biologique, UPMC - CNRS - INSU, Roscoff, France
* Micromass UK Ltd, Altrincham, Cheshire, UK
? Biochemistry Department, School of Medicine, Wayne State University, Detroit, USA

(Received 24 July19 October 1996) - EJB 96 1109/3

To elucidate the quaternary structure of the extracellular haemoglobin (Hb) of the marine polychaete
Arenicola marina (lugworm) it was subjected to multi-angle laser-light scattering (MALLS) and to
electrospray-ionisation mass spectrometry (ESI-MS). It was also subjected to SDS/PAGE analysis for
comparative purposes. MALLS analysis gave a molecular mass of 3648 % 24 kDa and a gyration radius
of 11.3 rt 1.7 nm. Maximum entropy analysis of the multiply charged electrospray spectra of the native,
dehaemed, reduced and carbamidomethylated Hb forms, provided its complete polypeptide chain and
subunit composition. We found, in the reduced condition, eight globin chains of molecular masses
15952.5 Da ( a l ) , 15974.8 Da (a2), 15920.9 Da (bl),16020.1 Da (b2), 16036.2 Da (b3),16664.8 Da (c),
16983.2 Da ( d l ) , 17033.1 Da (d2) and two linker chains Ll, 25 174.1 Da, and L2, 26829.7 Da. In the
native Hb, chains b, c, d occur as five disulphide-bonded trimer subunits T with masses of 49560.4 Da
(TI), 49613.9 Da (T2), 49658.6 Da (T3), 49706.8 Da (T4), 49724.5 Da (T5). Linker chains LI and L2
occur as one disulphide-bonded homodimer 2LI ( D I ) of 50323.1 Da and one disulphide-bonded hetero-
dimer LI-L2 (02) of 51 981.5 Da. Polypeptide chains a and d possess one free cysteine residue and chains
d possess an unusual total of five cysteine residues. Semi-quantitative analysis of ESI-MS data allowed
us to propose the following model for the one-twelfth protomer: [(3aI)(3a2),T'l (Tcorresponding to either
2'3, T4 or T5). From electron micrograph data TI and T2 are probably located at the centre of the molecule
as mentioned in previous studies. The Hb would thus be composed of 198 polypeptide chains with 156
globin chains and 42 linker chains, each twelfth being in contact with 3.5 linker subunits, providing a
total mass of 3682 kDa including haems in agreement with the experimental molecular mass determined
by MALLS. From ESI-MS relative intensities and the model proposed above, the globidlinker ratio gave
0.71 :0.29 and 0.73 :0.27, respectively. The estimation of haem content by pyridine haemochromogen and
by cyanmethaemoglobin (HiCN) methods also support the globin chain number provided by ESI-MS.
Keywords: Arenicola ; haemoglobin ; quaternary structure ; multi-angle laser-light scattering; electrospray
mass spectrometry.

The giant extracellular haemoglobins (Hb) and chlorocruor-
ins found in annelids [1] and vestimentiferans [2, 31 are charac-
terised by an acidic isoelectric point and by an hexagonal sym-
metry in electron micrographs consisting of two superimposed
hexagonal arrays of twelve spherical subunits, the hexagonal bi-
layer (HBL). These HBL Hb also contain a low haem and iron
contents (about 67% of those observed in other Hbs) and consist
of two types of chains, globin chains ( ~ 1 6 - 1 8kDa), account-
ing for approximately 70% of the total mass, and haem-deficient
linker chains ( ~ 2 4 - 2 8kDa) necessary for the assemblage into
the HBL structure [1, 4-61. The intertidal marine polychaete
Arenicola marina (L.) was one of the first 60s ( ~ 3 6 0 kDa)
0 spectrometry (ESI-MS), has been successfully applied to some
Correspondence to F. Zal, Equipe d'Ecophysiologie, Station Biolo-
gique, B. P. 74, F-29682 Roscoff cedex, France
Fax: +33 98 29 23 24.
E-mail: zal@sb-roscoff.fr
Abbreviations. HBL, hexagonal bilayer ; MALLS, multi-angle laser-
light scattering; TEM, transmission electron microscopy ; STEM, scan-
ning transmission electron microscopy; ESI-MS, electrospray ionisation
mass spectrometry ; MaxEnt, maximum entropy.
HBL Hb to be investigated by ultracentrifugation [7] and by
transmission electron microscopy (TEM) [8]. Although the func-
tional respiratory properties of A. marina Hb have been well
studied over many years [9-18], its structure remains poorly
understood despite several electrophoretic studies [19 -241. Re-
duced annelid Hb exhibit a variety of SDS/PAGE patterns and
the establishment of a correlation between unreduced and re-
duced Hb subunits is often difficult [25]. This difficulty is prin-
cipally due to the limited resolution and mass accuracy (5-
10%) of SDS/PAGE. Recently, a higher resolution (>1500) and
more accurate technique (0.01 %), electrospray ionisation mass
HBL Hb allowing the determination of their complete polypep-
tide chain composition [26-291. These HBL Hb were shown to
consist of 4-6 globin chains in the range 16-18 kDa and 3-4
linker chains in the range 24-27 kDa. Previous studies by SDS/
PAGE on A. marina HBL Hb have shown that the reduced Hb
consists of four types of polypeptide chains of 14000 ( I or A ) ,
16000 (ZZ or B), 31 000 (IZZ or C) and 35000 Da (ZV or D ) [21,
241. However, Pionetti and Pouyet [20] found five different
86 Zal et al. (Eur: J. Biochem. 243)
polypeptide chains, three around the molecular masses 14000-
16000 Da and two around 30000 Da. Moreover, El Idrissi Slit-
ine and Toulmond 1241, using two-dimensional electrophoresis,
revealed that components A and B could be further separated
into four polypeptide chains having the same relative molecular
mass but different isoelectric points (i.e. A I , A2 and B I , B2).
Although SDS/PAGE does not seem well suited for the precise
determination of the quaternary structure of HBL Hb, we have
subjected A. marina Hb to this analysis for comparison with the
literature data.
In the present work we provide detailed data on the masses
and complete polypeptide chain composition of the native HBL
Hb of the lugworm A . rnarincr as determined by ESI-MS coupled
with maximum entropy 1301. In addition, using the stoichiome-
try of the polypeptide chains and subunits constituting this Hb,
and an accurate mass measurement of native HBL Hb obtained
by multi-angle laser-light scattering (MALLS), we propose a co-
herent model of the quaternary structure of this HBL Hb.
EXPERIMENTAL PROCEDURES
Animal collection. Individuals of A. marina were collected
at low tide from a sandy flat near Roscoff (Penpoull beach),
Nord Finistkre, France, and kept in local running sea water for
24 h. Blood samples were withdrawn from the lugworm's
ventral vessel into glass capillary micropipettes and centrifuged
for 10 min at 10000 g at 4°C to remove insoluble material. The
resulting samples were either purified immediately or kept
frozen (-45°C) until use.
Purification technique. Hb solutions were purified by gel
filtration on a 1X 30 cm Superose 6-C column (Pharmacia LKB
Biotechnology, Inc.) using a low-pressure FPLC system (Phar-
macia). The column was equilibrated with a saline buffer as de-
scribed earlier [3]. Flow rate was typically 0.5 ml/min, and the
eluate was monitored with an ultraviolet detector (Pharmacia).
The elution product corresponding to the HBL Hb was collected
and concentrated with a 10-kDa cut-off microconcentrator, Cen-
tricon-10 (Amicon). The column was calibrated with the following
marker-proteins (HMW, Pharmacia) : aldolase (158 kDa), cata-
lase (232 kDa), ferritin (440 kDa), thyroglobulin (669 kDa).
Multi-angle laser-light scattering. MALLS measurements
were performed with a DAWN DSP system (Wyatt Technology
Corp.) directly on-line with the FPLC system described above.
The eluate was simultaneously monitored with an ultraviolet de-
tector and a refractometer (Erma, 7512). The Debye fit method
was used for molecular mass and gyration radius determinations
[31]. In this method the variation rate of the refractive index as
a function of concentration, dnldc, was set as 0.190 ml/g, typical
for proteins.
SDSEAGE. SDS/PAGE was carried out on slab gels using
the discontinuous buffer system of Laemmli [32]. The stacking
gel was constituted of 4% and 2.7% of acrylamide and bis-
acrylamide, respectively, pH 6.8, and the resolving gel by 10%
and 2.7%, respectively, pH 8.8. HBL Hb (1 mg/ml) were incu-
bated at 100°C for 90 s in the presence or absence of 5 % 2-
mercaptoethanol. Sample migration was performed at a constant
current of 150 mA using a Tris/glycine buffer, pH 8.3, for about
2 h. A two-dimensional electrophoresis method was used to es-
tablish the relationships between the subunits and their constitu-
ent polypeptide chains after 2-mercaptoethanol treatment as
follows. A gel slice containing the observed bands after SDS
migration was cut out and incubated under reducing conditions.
It was then placed at the top of a new gel and subjected to
electrophoresis as described above. The gels were stained over-
night on an agitator with Coomassie Brillant Blue R-250 (Bio-
rad) and destained by diffusion in a 25 % (by vol.) methanol and
7.5 % (by vol.) acetic acid solution. The relative molecular mass
of each constituent polypeptide chain was determined from a
plot of the log molecular mass versus the mobility of low-molec-
ular mass protein standards (LMW, Pharmacia).
Reverse-phase chromatography. HPLC was performed
using a Synchropack RP 1000 C,, column (100-nm diameter
pore size), 4.6 mmX250 mm (SynChrom, Inc.) and a gradient
HPLC system (Waters Chromatography Div., Millipore Corp.)
as previously described [28]. Linear gradients of water- aceto-
nitrile in 0.1 % aqueous trifluoroacetic acid were used with a
flow rate of 1 ml/min. The absorbance of the eluate was moni-
tored at either 220 nm or 400 nm. The chromatographic solvents
were reagent grade (J. T. Baker, Inc.). Each fraction was iden-
tified by SDS/PAGE and the peak areas were estimated by cut-
ting out the chart paper corresponding to the peaks and deter-
mining its mass.
Electrospray-ionisation mass spectrometry. Electrospray
data were acquired on a Quattro 11 (Micromass UK Ltd.) scan-
ning over the mlz range 600-2500 in 10 s/scan and about 45
scans were averaged to produce the final spectrum. Sample con-
centrations of 0.5 pg/pl in 1:1 (by vol.) acetonitrile/water con-
taining 0.2% formic acid were introduced into the electrospray
source at 5 pllmin. To establish the relationships between the
subunits and their constituent polypeptide chains, a solution of
the native Hb (5 pg/pl) was reduced with 10 mM dithiothreitol
at room temperature (20°C) and pH 8-9 (adjusted with ammo-
nium bicarbonate). Incubation was allowed to proceed for typi-
cally 1, 5 , 10 and 20 min, after which 10-pl aliquots of the Hb
solution were diluted with 90 pl of 50% aqueous acetonitrile
containing 0.2% formic acid and analysed by ESI-MS. Car-
bamidomethylation with iodoacetamide was carried out on the
native and reduced Hb for free and total cysteine determination,
respectively, under the same conditions as carboxymethylation
[33]. Deglycosylation was carried out using recombinant N-Gly-
cosidase F (Boehringer Mannheim Corp.) [34]. Mass scale cali-
bration employed the multiply charged series from horse heart
myoglobin (16951.5 Da; Sigma Cat. no. M-1882). Molecular
masses are based on the atomic masses of the elements given
by IUPAC [35]. The raw ESI-MS spectra were processed using a
maximum-entropy(MaxEnt)-based approach [36, 371 employing
the MemSysS program (MaxEnt Solutions Ltd.) incorporated as
part of the VG MassLynx software suite on a 60-MHz Pentium
PC .
Haem content and protein determination. Haem content
was determined by two different approaches, pyridine haemo-
chromogen [38] and HiCN [39] methods with the help of an
appropriate reagent solution. The reagent was made up as fol-
lows: 0.60 mM K,Fe(CN),, 0.76 mM KCN, 1.62 mM KH,PO,,
pH 7.0. The millimolar absorption coefficient is 32.0 at 557 nm
for reduced pyridine haemochromogen and 11.0 at 540 for
HiCN. For protein determination, we used the ultraviolet absor-
bance as described by Layne [40], using the following equation:
protein (mg/ml) = 1.55 A,,,)-0.76 A,,, with A,,, and A,,,, corre-
sponding to the absorbances at 280 nm and 260 nm, respectively.
Transmission electron microscopy. Arenicola Hb was di-
luted (1:900) with saline buffer [3] and applied to a very thin
carbon substrate supported on a microgrid, stained with 2%
(mass/vol.) uranyl acetate solution as described earlier [41]. The
specimens were examined with an electron energy of 80 keV,
using a Jeol JEM-1200EX microscope.
RESULTS
Molecular mass of Arenicola marina Hb by FPLC and
MALLS. Samples of the lugworm HBL Hb showed only one
 Zal et al. (Eur J. Biochern. 243) 87
107 , r

Fig. 2. SDSPAGE profiles of A. marina HBL Hb. Lane A, unreduced

haemoglobin ; lane B, after reduction with mercaptoethanol. The approx-
imate relative molecular mass were determined by molecular mass pro-
tein calibrates (LMW, Pharmacia).

8 5 13 11 12
Elution Volume i m l )

Fig. 1. MALLS analysis of Arenicola marina HBL Hb during its elu-

tion on a gel-exclusion column. Concentration of eluate shown as
(right axis). (A) Molecular mass (Da) versus elution volume (ml). (B)
Gyration radius (nm) versus elution volume (ml).
LP ,
10
, . . ; o , . , I
30
. .

symmetrical peak in gel filtration, indicating that the procedure
described was sufficient to obtain an homogeneous preparation.
The apparent relative molecular mass of A. marina Hb corres-
ponded to about 3600 kDa. MALLS analysis yielded profiles as
shown in Fig. 1 A with molecular masses estimated during the
whole elution of the peak. Mean mass of A. marina HBL Hb
was 3648 2 24 kDa (n = 1024) with a slight polydispersity as
indicated by the slope of the individual estimates. The gyration
radius corresponding to the mean of all dimensions of the mole-
cule was 11.3 ? 1.7 nm (Fig. 1B).
Polypeptide chains composition determined by SDSIPAGE.
The nomenclature used to name the different subunits and poly-
peptide chains is the same as that used by El Idrissi Slitine and
Toulmond [24]. Fig. 2 presents the SDYPAGE profiles of A.
murina Hb in the presence or absence of reducing agent. The
unreduced Hb dissociated in SDS into four subunits (Fig. 2, lane
A): I, fI, III and IV with apparent molecular masses of 14, 31,
33 and 42 kDa, respectively. After reduction (Fig. 2, lane B),
two groups of constituents were observed with molecular mass
of 14.1-16.4 kDa ( A l , B I , A2, B2) and 32-37 kDa (C and D).
Subunit I corresponds to the monomeric globin chain A l , sub-
unit fI and IfI would be disulphide-bonded dimers of non-haem
containing chains C and D, and subunit IV corresponds to a
disulphide-bonded trimer of globin chain ( B I , A2, B2).
Tine, z n
. I . . . 50, , . . 60
I 70
I
-&....&A
Fig.3. HPLC profile of A. marina purified HBL Hb on a Syn-
chropack c,,RP-1000 column using an acetonitrile-water gradient
in 0.1 % aqueous trifluoroacetic acid. Flow rate was 1 ml/min.
Arenicola HBL Hb profile obtained by HPLC. HPLC of
native A. marina Hb gave the chromatogram shown in Fig. 3.
The conditions used allowed us to resolve all the components
identified in the SDS/PAGE experiment including the com-
pletely dissociated haem group. The linker chains LI and L2 are
completely separated from the monomeric globin chains M I and
M2 and from the disulphide-linked trimer complex and linker
chain dimers. The globin/linker ratio determined by measure-
ment of mas7 was estimated to be approximately 0.72:0.28
(n = 6).
Polypeptide chains composition determined by ESI-MS.
ESI-MS analyses were made on the native Hb ( n = 4), on the
reduced Hb after 10, 30 and 60 min exposure to dithiothreitol
(n = 2 of each), on the carbamidomethylated Hb (n = 2), and
on the reduced and carbamidomethylated Hb ( u = 5). We will
employ the nomenclature previously used to name the subunits
and polypeptide chains of Lumbricus terrestris HBL Hb [42]
and Rijtia pachyptila [29]. The ESI-MS MaxEnt deconvoluted
spectrum of native A. marina Hb (Fig. 4A) shows two monomer
chains a1 and a2 of relative intensity 0.5 :Z .O, a complex mixture
of trimers (TI - T5) occurring with the relative intensities
0.5:0.6:1.0:1.0:0.8, and two linker dimers DI and 0 2 of rela-
88 Zal et al. (EUKJ. Biochem. 243)

DI
02 T3
15914.Y 49659.5
I I 'O7"
T4
49709.3

49600 4YRUU 52016.6

16000 162OU

tro-c . . , .-4

T4tCam DI
49767.8 5n322.9

1 D2
51983.4

0
4Y600 49800

C
SUOOO
K
i 2003.4

16664.4

15184.9
15953.7 17U32.6
"h a2
15976.0 *' L2
26828.0 16981.1

16000 I6250 16500 16750 170UO 17250

c+3Cam
16835.7 c+3Cnrn
1007
I 16835.1

cm
D 2+3
16141.9
mI+JCnni
n2t3Cam
16147.9 b2+3Cmm 1 d24 5C"Olt
17118.6

3i
1% - d2+SCam
17318.6
16125., L1+12Cnm
25856.1
LI+I OCm L2+ 12Cm 16000 16250 16500 16150 17(1UU 17250
25743.;) ,.27516.0
16091.1

L . . ., . niass
is000 zoo00 zso00 3oow 3sooo 40000 4sooo sooon (Da)

Fig. 4. ESL-MS spectra of A. marina HBL Hb. (A) MaxEnt-processed spectra of native Hb. Insets show details of monomeric chains a1 and a2
and trimeric complex. (B) MaxEnt processed spectra of carbamidomethylated Hb. Insets show details of carbamidomethylated a chains and trimers.
( C ) MaxEnt-processed data for Hb reduced with 5 mM dithiothreitol for 30 min. Inset shows details of monomeric chains forming the trimeric
complex. (D) MaxEnt-processed data for reduced and carbamidomethylated Hb. Inset shows details of carbamidomethylated globin-like chains.

tive intensity 1.0:0.6. Their masses are summarised in Table 1.
Fig. 4 B shows the spectrum of the Hb after carbamidomethyl-
ation whereupon the masses of the monomer chains (a1 and a2)
and the trimer complex increased by 57 Da, inferring that these
components each contain one free cysteine (Cys). The masses
of D l and D2 remained essentially the same after carbamido-
methylation, indicating that they have no free Cys.
Upon dithiothreitol treatment, the trimer complex and the
two linker dimers D l and 0 2 disappeared and eight new compo-
nents ( b l , h2, b3, c, d l , d2, L l and L2) appeared (Fig. 4C). The
sums of the masses of the b, c and d chains in various combina-
tions [ ( b l or b2 or b3)+c+(dl or d2)] imply that these chains
compose the complex mixture of the trimers. Similarly, these
data clearly indicate that DI is a homodimer of chain L1 and
0 2 is a heterodimer of chains LI and L2 (Table 2). By compar-
ing the carbamidomethylated (Fig. 4 B) and the reduced/
carbamidomethylated (Fig. 4D) spectra, it is possible to deter-
mine the total number of cysteine and their status (i.e. free or
involved in a disulphide bridge) for each chain. Thus, chains a1
and a2 possess three Cys, two of which are disulphide bonded
in the native Hb and one is free. Also, the b, c and d chains
contain three, three and five Cys, respectively, of which two Cys
in each chain form an intrachain disulphide bond in the native
Hb. In order to form the trimers, chain b and c each provide one
Cys for coupling to the d chain, leaving chain d with one free
Cys as observed for the native trimer complex after carbamido-
methylation. The linker chains LI and L2 were each found to
possess 12 Cys, all of which are involved in disulphide bonds.
All the ESI-MS results are summarised in Tables 1 and 2.
After deglycosylation, w e observed no change in the masses
of the chains and subunits in A. marina Hb (data not shown),
which implies that the Hb is not glycosylated.
 Zal et al. (EUKJ. Biochem. 243) 89
Table 1. ESI-MS-determined masses of A. marina Hb polypeptide chains and subunits. Native masses are means of five determinations on
native Hb (2estimated error). Free Cys represents the number of free Cys in chains and subunits in native Hb. Chain indicates the chains composing
the subunits in native Hb (see also Table 2). Reduced masses are means of four determinations on reduced Hb (i- estimated error). Red/ Cam masses
are means of five determinations on reduced and carbamidomethylated (Redcam) Hb (? estimated error). Corrected mass is the Red/Cam m d S S
corrected for carbamidomethylation (57.052 DdCys). Values represent measured reduced masses (iestimated error). Total number of Cys in chain
determined from differences between Red/Cam and reduced masses divided by 57.052 and rounded to nearest integer.
~ ~~

Chain/Subuni t Native Mass Free Cys Chain Reduced Mass Red/Cham Mass Corrected Mass Total Cys

Da Da

Monomer
a1 15952.5 2 1.0 15954.1 5 1.0 16125.621.0 15 954.4 5 1.O 3
a2 15974.82 1.0 15 976.8 t 1.0 I6 148.0 2 1.O 15 976.8 2 1.O 3
Trimer bl 15922.85 1.0 16092.12 1.0 15920.9 t 1.0 3
T1 49 560.4 i- 4.0 b2 16019.6f 1.O 16191.351.0 16020.1 f 1.0 3
T2 49613.9 5 4 . 0 b3 16036.0 5 1.0 16207.4 2 1.O 16036.22 1.0 3
T3 49658.6 t 4.0 C 16664.4k 1.0 16836.021.0 16664.82 1.0 3
T4 49706.8 i- 4.0 dl 16981.5 ? 1.0 17268.5 t 1.0 16983.2 2 1.O 5
T5 49 724.5 t 4.0 d2 17032.451.0 17318.451.0 17033.1 2 1 . 0 5
Dimer
DI 50323.1 5 4 . 0 LI 25 170.0 -+ 2.0 25 858.7 5 2.,0 25 174.1 2 2.0 12
02 51 981.5 5 4.0 L2 26827.3 t 2.0 27514.3 t 2.0 26829.7 2 2.0 12

Table 2. Correlation between disulphide-bonded subunit masses in 3.5 linker chains (e.g. 3 L1 and 0.5 L2) corresponding to a total
A. marina Hb and their constituent polypeptide chain masses. Calcu- linker mass of 88937 Da. In addition, the molecule contains two
lated masses were derived from w m of reduced chain masses (corrected extra copies of each trimer TI and T2 for a mass of 205746 Da
mass column, Table 1) Medsured masses were measured valuea for including haem. Finally the molecular mass of the native HBL
native HBL Hb (Table 1)
Hb obtained from the sum of individual masses corrected
Chains Calculated Measured Mass for haem content, gives a value of 12X(289738) + 205746
Mass (Da) Mass Differ- -- 3682 kDa for the whole molecule and a value of 23 602 Da of
ence proteidhaem group.

Number of globin chains estimated by haem and protein

content. The Hb concentration was estimated to be 145 2 3 mg/
TI = bl + L + dl-10H" 49558.8 49560.4 -1.6 ml (iz = 3) by protein concentration measurement, assuming that
T2 = bl +
c + d2-IOH" 49608.7 49613.9 -5.2 Hb was the major protein component in the vascular blood [43,
+
T3 = b2 c + dl-10H" 49 658.0 49658.6 -0.6
441. Using the pyridine haemochromogen method, the number
T4 = b2 + c + d2-10H" 49 707.9 49706.8 +1.1
T5 = b3 + c + d2-10H" 49 724.0 49724.5 -0.5 of haem groups was determined to be 1.53 [6.08 20.50 mM
D1 = L1 +L1-24Hh 50 324.0 50 323.1 +0.9 haem (n = 3), 23848.68 Da protein/mol haem] and by HiCN
0 2 = LI + L2-24Hh 51 979.6 51981.5 -1.9 method to be, 156 16.20 2 0.26 mM haem (11 = 3), 23 387.09 Da
protein/mole of haem]. These estimations were made using mass
" 10H to account for three intra-chain and two interchain disulphide of 3648 kDa for the whole HBL Hb as determined by MALLS.
bonds.
24H to account for ten intrachain and two interchain disulphide
bonds.
DISCUSSION
There is an appreciable scattering in the published values for
the molecular mass of Arenicola HBL Hh obtained mainly by
Quaternary structure model of A. marina Hb. MaxEnt analy- equilibrium sedimentation (range 2850-3850 kDa) 17, 8, 19-
sis of ESI-MS spectra produces semi-quantitative relative inten- 21, 231. This is probably due to the experimental conditions
sity data. It provides a zero charge spectrum in which the areas used, in particular to the divalent cation concentrations in the
under the peaks are proportional to the sum of the intensities of various buffers used for the experiments. It is now well known
the peaks in the original raw multiply charged spectrum. Using that the stability of HBL Hb is strongly dependent on the pres-
mass measurements of the native Hb obtained by MALLS and ence of group IIA cations (i.e. Ca2+,Mg" and Srz') [3, 45, 461.
the exact masses of polypeptide chains determined by ESI-MS, At present, there are several techniques available for the deter-
we propose a coherent model of the quaternary structure of A. mination of the molecular masses of huge molecules like the
marina Hb. According to these results this Hb would consist of HBL Hh, i.e. gel filtration, ultracentrifugation, small-angle X-
198 polypeptide chains divided into 156 globin-like chains (108 ray scattering, STEM mass mapping and MALLS. The two for-
monomers and 16 trimers) and 42 linker chains (Table 3). In this mer have accuracies limited to 5-lo%, the third needs ideal
model, the relative intensities of the two linker chains were conditions which are very difficult to obtain. STEM mass map-
LllL2 = 1.O: 0.2, the globin/linker ratio was approximately ping is not possible with all HBL Hb as reported recently [28],
0.73 :0.27. The one-twelfth subunit of Arenicola HBL Hb corres- and this is the case for Arenicola HBL Hb (Vinogradov, S. N.,
ponds to a dodecamer 1 ( 3 ~ 1 ) ( 3 a Z ) ~( TT corresponding
] to either unpublished results). MALLS, however, is an absolute method
T3, T4 or T5) with a calculated mean molecular mass of available for the determination of molecular mass and structure
200 800 2 27 Da. This basic subunit would be associated with without the need for universal calibration or other erroneous as-
90 Zal et al. (Eur: J. Biochern. 243)

Table 3. Model for the quaternary structure of A. marina HBL Hb. Masses are taken from Table 1. BPI, percentage of base peak intensity from
MaxEnt-processed ESI-MS spectra. The number of copies for each chain was calculated using the mass of Hb determined by MALLS. The number
of copies proposed in the model, was calculated assuming a D, point-group of symmetry of the molecule.

Subunit Chain Mass BPI 2 SD Copies -+ SD using Copies Total mass

experimental mass in the model

Da % Da
Monomers a1 15 952.5 16.872 3.57 38.62 8.2 36 514 290
a2 15974.8 29.64 2 5.41 67.7 t 12.5 72 1150186
Trirner5 TI bl + c f dl 49560.4 1.42 5 0.50 1.1 -+ 0.8 2 99121
T2 hl + L' + d2 49613.9 1.532 1.33 1.2? 1.0 2 99 228
T3 h2 + c f dl 49 658.6 6.222 2.35 4 . 6 2 1.7 4 198634
T4 h2 + c + d2 49706.8 5.98 2 2.37 4.4-t 1.7 4 198 827
T5 63 + c + d3 49724.5 5.79 i 1.80 4.22 1.3 4 198 898
Sub-total globins 156 2519 184
linkers L1 25 174.1 28.42? 12.35 26.5 i 8.85 35 88 1 093
L2 26829.7 5.11 i 3.60 4.6i 2.55 7 187 808
Sub-total linkers 42 1 068 901
Sub-total haem ' 96 174
Total mass of Hb 3 684 259
Experimental m a s h 3 648 000

156X616.5 Da.
t' Value obtained by MALLS.

sumptions and is a very accurate technique [31]. Furthermore,
MALLS analysis can be performed directly on-line with an
FPLC system using physiological buffers, thus avoiding sample
deterioration and allowing sample recovery. The molecular mass
found for Arenicola HBL Hb by MALLS (3648 -C 24 kDa) is
in good agreement with the most recently published values of
3740 t- 120 kDa and 3560 ? 40 kDa, respectively [21, 201.
On the whole, our SDSIPAGE results on reduced Arenicola
Hb are in good agreement with previous studies (20, 21, 241.
However, we obtained a high resolution thus clearly revealing
four components around 14100-16400 Da, whereas two of the
earlier works found only two [20] and three [21] components.
El Idrissi Slitine and Toulmond [24] also found four polypeptide
chains by IEFIPAGE, and they concluded that the polypeptide
chains A and B were of the same molecular mass but different
isoelectric points. Until recently, these results would have been
judged sufficient to model the quaternary structure of this Hb.
However, the recent success of ESI-MS in characterising the
primary composition of these molecules [26-291 illustrates that
considerably more detail can be provided by this technique than
by electrophoresis. This is due to the higher resolving power
available from ESI-MS (= 6 Da at 16 kDa) and the ability of
the technique to give unambiguous and more precise values for
the molecular masses (< ? 0.01 %). In the case of Areiiicola, our
ESI-MS analysis revealed a total of eight different polypeptide
chains in the 16-17-kDa region, by exhibiting two a, and three
b. one c and two d chains. This complexity was also found in
the trimer, with five plausible combinations of the six chains b,
c and d, despite the less clearly defined group of peaks compos-
ing the trimer (Fig. 4A).
Arenicoh Hb does not possess glycosylated chains in con-
trast to Lunzbricus terrestris where both the trimer and one of
the linker chains are glycosylated [28]. Furthermore, the linker
chains appear to occur as disulphide-bonded dimers in the native
Hb, similar to those found in Qlorrhynchus heterochaetus [27],
another polychaete annelid. An even number of interchain disul-
phide bonds appear to be involved in forming the linker dimers
in Arenicola, since the carbamidomethylated Hb data suggest
that there is no free Cys present in the dimers and the reduced/
carbamidomethylated data suggest that the linker monomer
chains each contain an even number (12) of Cys.
Two of the first analyses of HBL Hb by ESI-MS 126, 281
gave globinllinker ratios that agreed with the results obtained by
HPLC. Our results are in total agreement with these authors,
since we found globinllinker ratios of 0.72:0.28 and 0.73:0.27
by HPLC and ESI-MS, respectively. These observations are very
important and demonstrate that the ionisation behaviour of glo-
bin and linker chains is similar and, therefore, that ESl-MS can
be a fully quantitative method for HBL Hb studies. Using the
EST-MS stoichiometry we propose a model composed of 156
globin chains in good agreement with the results obtained by
pyridine haemochromogen (153 haem groups) and by HiCN
method (156 haem groups). In addition, Toulmond [47] deter-
mined a number of 1 5 2 + 2 Fe atoms assuming a mass of
3600 kDa for the native molecule and Vinogradov and his col-
laborators [21] proposed 1 6 3 2 1 2 Fe atoms for a mass of
3700 kDa. These authors have determined a minimum molecular
mass of 23 640 Da by atomic absorption spectroscopy [47] and
22900 Da by the method of Cameron [21]. These two values
are very close to our results, 23 848 Da and 23 387 Da, found by
pyridine haemochromogen and HiCN methods. These experi-
mental values are also in good agreement with the value derived
from our model (23 602 Da).
Now, if we consider that the one-twelfth subunit is made up
of 12 polypeptide chains there would be only 144 globin chains
in the whole molecule. These results suggest that, in contrast to
Macrohdella [26], Tvlorrhynchus [27], Lunibricus [28] and R$
tia 1291 HBL Hb investigated by ESl-MS, Arenicola HBL Hb
would contain 13 haem groupslone-twelfth subunit or, more
probably, that there would be a thirteenth subunit at the center
of the molecule. This latter hypothesis is substantiated by a close
examination of electron micrographs (Fig. 5). However, the
composition of this central subunit would be different from that
of the other one-twelfth subunits as suggested by ESI-MS stoi-
chiometry and it could be made up of four trimers (2 T1 and
2 T2) although alternative compositions may be possible. The
presence of a thirteenth subunit located at the center of Arenicola
HBL Hb has never been observed before, but two papers men-
 Zal et al. (Eur: J. Biockem. 24.3)
Fig.5. Electron micrographs of A. marina HBL Hh, negatively
stained with 2 % uranyl acetate showing the two frequent orienta-
tions, face view and lateral view (arrows). (A) View of A. nmrina
HBL Hb; Scale bar = 60nm. (B and C) Details (X5) of the face and
lateral views pointed in (A), showing the central subunit.
Fig. 6. Model of A. manna HBL Hb, adapted from [59]. (A) Drawing
of typical HBL Hb on top view displaying the polypeptide chains for
half a Hb molecule. Each one-twelfth being constituted of one trimer
(T3, T4 or T5) as mentioned by Waxman (1975) 1.591, with nine mono-
meric chains. We have represented the two trimers (T), e.g. T1 and T1
or TI and T2, at the center of the HBL structure as viewed on electron
micrograph. (B) Detail of a one-twelfth showing the possible arrange-
ment of monomeric chains.
tioned this eventuality [23, 481. Several annelid HBL Hb possess
a central subunit like, for example, Neplzthys iizcisa [49-511,
Ophelia hicornis [52-55], Glossoscolex paulistus [48], Eophila
tellinii 1.561, Oenoize ,fulgida [57] and Euzonus mucronatcz [%I.
Ghiretti-Magaldi and co-workers [S3] showed that the central
subunit for 0. hicornis appeared to differ from the external one-
twelfth as suggested here for Arenicolu. However, the presence
of a central subunit ion HBL Hb is not always clear since, for
the same species, some TEM preparations could be composed
of molecules with or without the central subunit [S2]. No expla-
nation exists for this phenomenon, but it is probable that chang-
ing the experimental conditions induces the dissociation of this
unstable subunit. For Arenicolu preparations we think that the
saline buffer used (containing among other things, 32 mM Mg”
and 11 mM Ca2+)maintains the integrity of the molecular archi-
tecture of this HBL Hb, since in all preparations made with Tris/
HCI buffer no central subunit could be observed (1571 and Zal,
F., unpublished result).
The top view of HBL Hb resembles a regular hexagon ap-
proximately 27 nm between parallel sides. Using the ESI-MS
results, we propose the model as presented in Fig. 6. Each corner
of the hexagon would be composed of three smaller units, com-
posed of the monomeric globin chains a, forming an equilateral
triangle as proposed elsewhere [59], with a trimer subunit at the
center of the triangle. This complex appears to have a mean
mass of around 200 kDa. However, to improve this theoretical
model, a three-dimensional reconstruction from cryoelectron mi-
croscopic images would be necessary as performed recently on
91
Lurnhricus [60], Mcicrobdellu (611, Rifia [62] HBL Hb and the
chlorocruorin of Eudistylia 1631.
Other interesting information revealed by ESI-MS concerns
the presence of an unusual number of cysteine residues on glo-
bin-like d chains and the presence of a free cysteine residue
on the monomeric globin-like u chains. The Annelid globin-like
chains sequenced can be classified into four groups according to
the location ofcysteine residues [64]. In all groups, an intrachain
disulphide-bridge is formed and the remaining cysteine residues
all participate in interchain disulphide-bridges except in Lurnelli-
hrachia sp. [65] and Riftia pachyptila [29] which belong to ves-
timentiferans. These possess some globin-like chains with a free
cysteine residue that can be involved in sulphide-binding ([65 I
and unpublished results). However, the polypeptide d chains,
with five cysteine residues including the one free, and the pres-
ence of free cysteines on a1 and (12 make Arenicola HBL Hb
atypical when compared with the four groups mentioned above.
Sulphide concentration in the pore water of the sediment can
reach several hundred pmol/l-‘ 1661 and may therefore be con-
sidered an important environmental factor for a sediment-dwell-
ing animal such as Arenicola [67, 681. In addition, Patel and
Spencer [69] have reported that the chemistry of A. tnarirzu Hb
differed in several respects from vertebrate Hb. In the presence
of sulphide, vertebrate oxyhaemoglobin forms sulfhaemoglobin.
The failure of Arenicola HBL Hb to form this compound might
be indicative of a biochemical adaptation appropriate to an ani-
mal which may at times be exposed to high concentrations of
sulphide (701. The free cysteine on the polypeptide d chains
could intervene in this mechanism by avoiding the formation
of sulfhaemoglobin and allowing sulphide oxidation [70]. The
oxidation product, defined as brown pigment, and which pos-
sesses a higher sulphide oxidation activity than HBL Hb. could
be due to some modification or breakage of the prosthetic-
group-protein linkage during the reaction leading to its forma-
tion [69]. However, for Arenicola, this would be a detoxification
mechanism (i.e. sulphide immobilisation by the Hb), and not
an adaptation to symbiotic life as is the binding of sulphide by
vestimentiferan Hb [71, 721.
We are indebted to T. Azoulay of the Sopares Society for MALLS
analysis. We also thank J. Sourimant and S. Hourdez for the excellent
preparation of materials and transmission electron micrographs. This
work was supported by research grants from Centre Nationnl de la Re-
cherche Scientifiiyue (UPR 9042), l’lnstitut National des Sciences de
I’Univers, l’lnstitut Franpis de Recherche et d’Exploitation de la Mer
(Unite‘ de Recherche Marine no 7 , the National Sciences Foundation
and the National Institutes of Health (DK38674).
REFERENCES
1, Vinogradov, S. N. (1985) in Respirutor). pigments in animals (Laniy.
J. N.. Truchot, J. P. & Gilles, R., eds) pp. 9-20, Springer-Verlag.
Berlin.
2. Suruki, T., Takagi, T. & Ohta, S. (1988) Biochem. J. 255,541-545.
3. Zal, F., Lallier, F. H., Wall, J. S., Vinogradov, S. N. & Toulmond,
A. (1996) J. Biol. Chem. 271, 8869-8874.
4. Vinogradov, S. N., Kapp, 0. H. & Ohtsuki, M. (1982) in Electron
niicroscopy of proteins (Harris, J., ed.) pp. 135- 263, Academic
press, New York.
5. Vinogradov, S. N. (1985) Comp. Biochern. Physiol. B82, 1-15.
6. Vinogradov, S. N., Walz, D. A., Pohajdak, B., Moens, L., Kapp, 0.
H.. Suzuki, T. & Trotman, C. N. A. (1993) Comp. Biochem. Phy-
siol. BI 06, 1- 26.
7. Svedberg, T. & Eriksson, I. B. (1933) J. Am. Cheni. Soc. 55,2834-
2841.
8. Breton-Gorius, J. (1963) A m . Sci. Natl Zool. Paris 5, 211 -272.
9. Toulmond. A. (1970) C. R. Acud. Sci. Paris 270, 1368-1371.
10. Toulmond, A. (1970) C. R. A c ~ t l Sci.
. Paris 270, 1487-1490.
92 Zal et al. (Eur: J. Biochem. 243)

11. Toulmond, A. (1971) C. R. Acad. Sci. Paris 271, 1368-1371. 44. Wells, R. M. G. (1973) Comp. Biochem. Physiol. 46, 325-332.
12. Toulmond, A. (1971) C. R. Acad. Sci. Paris 272, 3184-3187. 45. Sharma, P. W., Kuchumov, A. R., Chottard, G., Martin, P. D., Wall,
13. Toulmond, A. (1973) Respir: Physiol. 19, 130-144. J. S. & Vinogradov, S. N. (1996) J. B i d . Chem. 271, 8754-8762.
14. Toulmond, A. (1975) J. Exp. B i d . 63, 647-660. 46. Lamy, J. N., Green, B. N., Toulmond, A., Wall, J. S., Weber, R. E.,
15. Weber, R. E. (1970) Comnp. Biorhem. Physiol. 35, 3 79 - 189. Vinogradov, S. N. (1996) Chem. Rev., in the press.
16. Everaarts, J. M. & Weber, R. E. (1974) Comp. Biocliem. Physiol. A 47. Toulmond, A. (1979) C. R. Acnd. Sci. Paris 0288, 651 -654.
48, 507-520. 48. El Idrissi Slitine, F., Torriani, 1. & Vachette, P. (1990) in Invertebrate
17. Rasmussen. K. K. & Webert, R. E. (1979) Ophelia 18, 151-170. diowygen carriers (PrCaux, G. & Lontie, R., eds) pp. 225-228,
18. Weber, R. E. (1981) Nature 292, 386-387. Leuven Univ. Press, Leuven.
19. Waxman, L. (1971) J. B i d . Chem. 246, 7318-7327. 49. Wells, R. M. G. & Dales, R. P. (1976) Comp. Biochem. Physiol.
20. Pionetti, J. M. & Pouyet, J. (1980) Eur: J. Biochem. 105, 131-138. A54, 387-394.
21. Vinogradov, S. N., Kosinski, T. F. & Kapp, 0. H. (1980) Biochinz. 50. Messerschmidt, U., Wilhelm, P., Pilz, I., Kapp, 0. H. & Vinogradov,
Biophys. Acta 621, 3 15- 323. S. N. (1983) Biochim. Biophys. Acta. 742, 366-373.
22. Vinogradov, S. N., Shlom, J. M. & Doyle, M. (1980) Comp. Bio- 51. Vinogradov, S. N., Kapp, 0. H., Messerschmidt, U., Wilhelm, P. &
chrm. Physiol. R65, 145- 150. Pilz, 1. (1983j in Structure andfunction of invertebrute r e p i r a t q
23. Wilhelm, P. P., Pilz, 1. & Vinogradov. S. N. (1980) Ifit. J. B i d . proteins (Wood, E. J., ed.) pp. 197-201, Harwood Acad. Publish-
Mncromol. 2. 383-384. ers, London.
24. El Idrissi Slitine. F. & Toulmond, A. (1991) Comp. Biochem. Phy- 52. Mezzasalma, V., di Stefano, L., Piazzese, S., Zagra, M., Salvato, B.,
siol. B100, 631-634. Tognon, G. & Ghiretti-Magaldi, A. (1985) Biochim. Biophys.
25. Vinogradov, S. N., Shlom, J. M. & Frossard, P. (1980) Conip. Bio- Acta. 829, 135-143.
chern. Physiol. B67, 1-16, 53. Ghiretti-Magaldi, A., Zanotti, G., Tognon, G. & Mezzasalma, V.
26. Weber, R. E., Make, H., Braswell, E. H., Oliver, R. W. A,, Green, (1 985) Biochim. Biopliys. Acta. 829, 144- 149.
B. N., Sharma, P. K., Kuchumov, A. & Vinogradov, S. N. (1995) 54. Cejka, Z., Santini, C., Tognon, G. & Ghiretti-Magaldi, A. (1991) J .
J . Mol. Rid. 251, 703-720. Struct. B i d . 107, 259 -267.
27. Green, B. N., Suzuki. T., Gotoh, T., Kuchumov, A. R. & Vinogradov, 55. Cejka, Z., Kleinz, J., Santini, C., Hegerl, R. & Ghiretti-Magaldi, A.
S. N. (1995) J . B i d . Chem. 270, 18209-18211. (1992) J. Struct. Biol. 109, 52-60.
28. Martin, P. D., Kuchumov, A . R., Green, B. N., Oliver, R. W. A,, 56. Cejka, Z., Ghiretti-Magaldi, A,, Tognon, G., Zanotti, G. & Baumeis-
Braswell, E. H., Wall, J. S. & Vinogradov, S. N. (1996) J. Mol. ter, W. (1989) J. Ultrastruct. Mol. Struct. Res. 102, 71 -81.
B i d . 255, 154-169. 57. Van Bruggen, E. F. J. & Weber, R. (1974) Biochim. Biophys. Acta
29. Zal, F., Lallier, F. H., Green. B. N., Vinogradov, S. N. & Toulmond, 359, 210-212.
A. (1996) J . B i d . Cheni. 271. 8875-8881. 58. Terwilliger, R. C., Terwilliger. N. B., Schabtach, E. & Dangott. L.
(1977) Comp. Biochem. Physiol. A57, 143-149.
30. Ashton, D. S., Beddell, C. R., Green, B. N. & Oliver, R. W. A.
59. Waxman, L. (1975) J. B i d . Cheni. 250, 3790-3795.
(1994) FEBS Lett. 342, 1-6.
60. Schatz, M., Orlova, E. V., Dube, P., Jager. J. & van Heel. M. (1995)
31. Wyatt, P. J. (1993) Anal. Chin?. Acta 272, 1-40.
.I. Struct. Biol. 114, 28-40.
32. Laemmli, U. K. (1970) Nature 227, 680-685.
61. De Haas, F., Boisset, N., Taveau, J. C., Lambert, O., Vinogradov, S.
33. Crestfield, A. M., Moore, S. & Stein, W. H. (1963) J. Biol. Chem.
N. & Lamy, J. N. (1996) Biophys. J. 70, 1973-1984.
239, 622-627.
62. De Haas, F., Zal. F., Boisset, N., Taveau. J. C., Lambert, O., Lallier,
34. Tarentino, A. L., Gomez, C. M. & Hummer, T. H. J r (1985) Bio-
F. H., Toulmond, A. & Lamy, J. N. (1996) Proteinx, in the press.
chemist?> 24, 4665-4671. 63. De Haas, F., Taveau, J. C., Boisset, N., L,ambert, O., Vinogradov, S.
35. IUPAC Commission on atomic inasses and isotopic abundances N. & Lamy, J. N. (1995) J . Mol. Biol. 255, 140-163.
(1993) .I. Phys. Chetn. Ref: Data 22, 1571-1584. 64. Suzuki, T., Takagi. T. & Ohta, S. (1990) Biochem. J. 266, 221-225.
36. Ferrige, A. G., Seddon, M. J. & Jarvis. S. A. (1991) Rapid Comnzun. 65. Suzuki, T., Takagi, T., Okuda, K., Furukohri & Ohta, S. (1989) Zool.
Mass Spectrorn. 5, 374-379. sci. 6 , 915-926.
37. Ferrige, A. G., Seddon. M. J., Green, B. N., Jarvis, S. A. & Skilling, 66. Volkel, S. & Grieshaber, M. K. (1992) J. Comp. Physiol. B162,
J. (1992) Rapid Commun. Mass Spectrorn. 6, 707-711. 469 -477.
38. De Duve, C. (1948) Acta. Chem. Scnud 2, 264. 67. Groenendaal, M. (1979) Neth. J. Sea Res. 13, 562-570.
39. Van Assendelft, 0. W. (1970) Spectrophotometry of haemoglobin 68. Groenendaal, M. (1980) Neth. J. Seu Res. 14, 200-207.
derivatives, Royal Vangorcuni, Ltd., Assen. 69. Patel, S. & Spencer, C. P. (1963) Comp. Biochem. Ph
40. Layne, E. (1957) Methods Enzymol. 3, 447. 82.
41. Valentine, R . G., Shapiro, B. M. & Stadtman, E. R. (1968) Biochen- 70. Patel, S. & Spencer, C. P. (1963) J. Mar: B i d . Assoc. U. K. 43,
istry 7, 2143-2152. 167- 175.
42. Fushitani, K. & Riggs, A . F. (1988) Proc. Nut/ Acad. Sci. USA 85, 71. Arp, A. J., Childress, J. J. & Vetter, R. D. (1987) J. Exp. Biol. 128,
9461 -9463. 139-158.
43. Florkin, M. (1935) A r h . Int. Physiol. 40, 283-290. 72. Volkel. S . & Grieshaber, M. K. (1994) Mar: B i d . 118. 137-174.