Allosteric water and phosphate effects in Hoplosternum littorale
hemoglobins

Patricia Peres1, Walter F. de Azevedo Júnior1 and Gustavo O. Bonilla-Rodriguez2

1Departamento de Fı́sica, and 2Departamento de Quı́mica e Ciências Ambientais, IBILCE-UNESP, State University of São Paulo,

São José do Rio Preto SP, Brazil

This paper reports the results obtained using the osmotic
stress method applied to the purified cathodic and anodic
hemoglobins (Hbs) from the catfish Hoplosternum littorale,
a species that displays facultative accessorial air oxygen-
ation. We demonstrate that water potential affects the oxy-
gen affinity of H. littorale Hbs in the presence of an inert
solute (sucrose). Oxygen affinity increases when water
activity increases, indicating that water molecules stabilize
thehigh-affinity stateof theHb.This effect is the sameas that
observed in tetrameric vertebrate Hbs. We show that both
anodic and cathodic Hbs show conformational substrates

similar to other vertebrate Hbs. For both Hbs, addition of
anionic effectors, especially chloride, strongly increases the
number of water molecules bound, although anodic Hb did
not exhibit sensitivity to saturating levels of ATP. Accord-
ingly, for both Hbs, we propose that the deoxy conforma-
tions coexist in at least two anion-dependent allosteric states,
To and Tx, as occurs for human Hb. We found a single
phosphate binding site for the cathodic Hb.

Keywords: hemoglobin; osmotic-stress; catfish.

Water plays a unique and ubiquitous role in biomolecules
and biochemical reactions; folding, stability, and function
of protein molecules are all influenced by interaction with
water molecules [1]. A central role for water in determining
structure and regulating function of proteins is becoming
increasingly evident, as water molecules act as allosteric
effectors by preferentially binding to a specific protein
conformation [2].

Significant changes in protein hydration are conveniently
studied by the osmotic stress method, a simple method [3,4]
based on water activity of the solution, which is altered by
changing the concentrations of solutes (polyols, sugars and
amino acids).

Fish hemoglobins (Hbs), which display a wide range of
oxygen binding properties and allosteric effects, and are
characterized extensively both structural and functionally,
are excellent candidates for such an analysis. In several cases
fishes have iso-Hbs with marked functional differentiation
in terms of allosteric control and cooperation.

This study analyzed the effect of water on the cathodic
and anodic Hbs fromHoplosternum littorale, a catfish from
the Amazon basin that displays facultative accessorial air
oxygenation by air gulping and gas-exchange by using a

partially modified intestine when water oxygen falls below a
critical concentration [5]. The Hbs present different oxygen
affinities and responses to allosteric effectors. Its anodic Hb
displays a reverse Bohr effect in the stripped form, changing
to a normal response to protons in the presence of ATP.
The major component, named cathodic Hb, exhibits a
pronounced alkaline Bohr effect and, accordingly, a high
response to pH changes [5].

Generally, only one molecule of organic phosphate
(NTP) is bound per deoxy-Hb molecule, although addi-
tional binding sites for ATP have been proposed [5–8]. For
the cathodic Hb from the fish H. littorale, Weber et al. [5]
suggested the possible existence of one additional phosphate
binding site.

Materials and methods

Hemolysate preparation

Blood was collected by caudal vein puncture from adult
specimens at the Central Animal Facility of the State
University of São Paulo (IBILCE-UNESP) at São José do
Rio Preto SP (Brazil). The animals were anesthetized using
benzocaine (1 g per 15 L of water) and, after blood
collection, the specimens showed a fast recovery from
anesthesia. Subsequent Hb purification procedures were
carried out at low temperature (around 4 �C) using ultra-
pure water (Elga Sci.). Red blood cells (RBC) were washed
by centrifugation four times with buffered saline (containing
50 mM Tris/HCl pH 8.0 and 1 mM EDTA). RBC were
frozen in liquid N2 and hemolysis was accomplished by
adding buffer A (30 mM Tris/HCl pH 9.0), followed by
clarification by centrifugation (1000 g for 1 h). Using the
same buffer, but containing 0.2 M NaCl, initial purification
was performed by gel filtration on Sephacryl S-100 HR
(Sigma) on an equilibrated 2.6 · 30 cm column. The

Correspondence to G. O. Bonilla-Rodriguez, Departamento de

Quı́mica e Ciências Ambientais, IBILCE-UNESP, State University of

São Paulo, Rua Cristovão Colombo 2265, São José do Rio Preto SP,

CEP 15054–000. Fax: +5517 2212356, Tel.: +5517 2212361,

E-mail: bonilla@qca.ibilce.unesp.br

Abbreviations: 2,3-BPG, 2,3-biphosphoglycerate; Hbs, hemoglobins;

RBC, red blood cells.

Note: A website is available at http://www.qca.ibilce.unesp.br/

labbioq.html

(Received 23 July 2004, revised 1 September 2004,

accepted 14 September 2004)

Eur. J. Biochem. 271, 4270–4274 (2004) � FEBS 2004 doi:10.1111/j.1432-1033.2004.04366.x



fractions containingHbwere pooled and dialyzed overnight
against buffer A and subsequently purified on Q-Sepharose
usinga linear saline gradientbetween0and100 mMofNaCl.
The isolated components were concentrated by centrifuga-
tion on Amicon microconcentrators. Analytical isoelectric
focusing was performed in agarose gel. The Hb solutions
were stored in liquid N2 in aliquots that were thawed
immediately before oxygen binding studies were carried out.

Osmotic stress experiments

Water activity was varied by addition of pure sucrose
(Acros Organics). In the osmotic stress method changes in
the Hb–oxygen affinity are related to changes in water
activity that can be converted to changes in protein
hydration by use of linkage equations [2,9]. Oxygen binding
experiments were performed with 60 lM (heme) Hb solu-
tions in 30 mM Hepes buffer, pH 7.5 in the presence and
absence of ATP and NaCl, as described by Colombo and
Bonilla-Rodriguez [4]. All equilibrium measurements were
carried out at 20 �C by the tonometric method [10]. The
functional parameters P50 (O2 partial pressure at half
saturation) and cooperativity (n50) were calculated from
Hill plots by linear regression around half saturation.
Hemoglobin and methemoglobin concentrations were esti-
mated using the extinction coefficients for human Hb [11].
Data obtained from samples containing more than 5%
methemoglobin (final concentration) were discarded.

The Hb solution osmolalities (Osm) were determined
after binding experiments from freezing point depression
measurements using an Osmette A model 5002 osmometer
(Precision Systems Inc.). The osmolality was transformed to
the natural logarithm of water activity through the follow-
ing relationship [4]:

ln aw ¼ D
KfMw

¼ ��Osm

Mw
ð1Þ

where D is the freezing point depression, Kf ¼ 1.86
KÆkgÆmol)1 is the cryoscopic constant, and Mw is the
molarity of pure water (55.56 molÆL)1).

The effect of water as a single heterotropic ligand on
oxygenation is typically analyzed with the following linkage
equation [12,13]:

d lnðP50Þ
d lnðawÞ

¼ Dnw
4

¼ � noxyw � ndeoxyw

� �
ð2Þ

where aw is the water activity. The slope of the linkage plot
ln(P50) vs. ln(aw) gives the differential number of water
molecules bound in the conformational transition from the
deoxy to the oxy structures, Dnw.

The slopes were compared according to Zar [14] using
GRAPHPAD PRISM version 4.00 for Windows (GraphPad
Software, San Diego, CA, USA). We tested the null
hypothesis (no significant difference between slopes) for
paired experiments using a P threshold of 0.05.

Calculation of the association constants of ATP to the
forms oxygenated and deoxygenated of the cathodic Hb

The �x� number of molecules of ATP differentially bound
per heme between the deoxy- and oxy-Hb was calculated
using the linkage equation of Wyman [12]:

x ¼ D logP50=D log½ATP� ð3Þ

The association constants with ATP were calculated by a
nonlinear regression fitting using the program SIGMAPLOT

(Jandel Scientific, San Rafael, CA, USA), according to the
equation below [15]:

logðP50Þp ¼ logðP50Þa þ 1

4
log

1þ KDX

1þ KOX

� �
ð4Þ

where log(P50)p is the logarithm of P50 measured in the
presence of ATP, log(P50)a is measured in the absence of
ATP, KD and KO are the association constants to the
deoxygenated and oxygenated forms, respectively, and X is
the free molar concentration of ATP. O2 binding experi-
ments were performed at pH 7.5 and at 20 �C.

Results

O2 equilibria of Hb at various osmolalities

Cathodic Hb. We tested the oxygen affinity of the cathodic
Hb as a function of water activity in different experimental
conditions: for the strippedHb (in anATP and chloride-free
buffer solution), in the presence of 0.1 mM and 1 mM of
ATP, in a buffer containing 100 mM NaCl, and a last set
containing 100 mM NaCl + 1 mM ATP. The plots (Fig. 1)
show that ln(P50) varies linearly with changes in the water
activity (aw); this is in agreement with Colombo et al. [2]
and Hundahl et al. [16]. Oxygen-affinity decreased for all
the experimental sets containing ATP and/or chloride, in
comparison with the stripped Hb.

The analysis of the data according to the Wyman
equation (Table 1) shows that the cathodic Hb in the
stripped form, binds 41 ± 9 extra water molecules in the

Fig. 1. Relative shift in Hbct ln(P50) as a function of water activity (aw).

The different conditions were: stripped Hb, 0.1 mM and 1 mM of ATP,

100 mMNaCl and 100 mMNaCl + 1 mMATP. The straight lines are

a linear fit of the data using the integrated form of Wyman linkage

equation (Eqn 2). Experimental conditions: 30 mM Hepes buffer,

pH 7.5 and 20 �C.

� FEBS 2004 Allosteric effects in Hoplosternum littorale Hbs (Eur. J. Biochem. 271) 4271



T to R transition. In the presence of 0.1 mM and 1 mM of
ATP these numbers increase to 73 ± 8 and 65.6 ± 12,
respectively, and in the presence of 0.1 M chloride this rises
to 85 ± 12 water molecules. In the simultaneous presence

of 100 mM NaCl and 1 mM of ATP Dnw decreased
drastically to 4±16 water molecules, but oxygen affinity,
measured by P50, was higher than in the presence of 1 mM

ATP, a finding also reported byWeber et al. [5]. All theDnw
values obtained in the presence of ATP and/or Cl– were
significantly different than that from the stripped form.

Anodic Hb. The other fraction studied here has a similar
behavior concerning aw when compared with the cathodic
Hb and other fish Hbs [16], also indicating preferential
binding of water molecules to the R state. The allosteric
effectors significantly affect water and O2 binding (Fig. 2),
and both chloride and 0.1 mM ATP induced an increase of
O2 affinity, also described also by Weber et al. [5]. Linear
fitting of the data (Table 2) showed a Dnw of 58 ± 8 water
molecules for the stripped form, increasing to 68 ± 12 in
the presence of 0.1 mM and 1 mM of ATP. In the presence
of NaCl, Dnw rose to 116 ± 16 water molecules, the only
significant difference when compared to the stripped Hb. In
the presence of 1 mM of ATP and 100 mM of NaCl, Dnw
decreased to 28 ± 8. This value was not found to be
significant, probably due to the poor linearity of the data
with a higher aw. In contrast to the cathodic Hb, the
combined effect of Cl– and ATP induced the largest
decrease of O2 affinity.

Calculation of the association constants of ATP to the
oxygenated and deoxygenated forms of the cathodic Hb

Using Eqn (3) (Fig. 3), we calculated the slope, a Dx of
0.23 ± 8 · 10)5 ATP molecules/heme to Hb, which con-
firms the binding of a single ATPmolecule per Hb tetramer.

Table 1. Change in the number of water molecules (Dnw) ± SD bound to the cathodic Hb in the transition from fully deoxy to fully oxy forms,

measured by tetramer in different experimental conditions. Experiments for the cathodic Hb were performed in 30 mM Hepes buffer pH 7.5 and

20 �C. The slopes were compared using the stripped condition as a reference.

Sample Experimental condition

Dnw ± SD

Wyman Statistical analysis

Correlation

coefficient

Hbct Sucrose, stripped Hb 41 ± 09 Reference 0.972

Sucrose + ATP 1 mM 66 ± 12 * 0.988

Sucrose + ATP 0.1 mM 73 ± 08 ** 0.905

Sucrose + NaCl 100 mM 85 ± 12 * 0.911

Sucrose + ATP 1 mM + NaCl 100 mM 4 ± 16 ** 0.893

*P ¼ 0.001 < P < 0.01, **P < 0.001.

Fig. 2. Relative shift in Hb
an
ln(P50) as a function of water activity (aw).

The different conditions were (strippedHb, 0.1 mM and 1 mM of ATP,

100 mM NaCl and 100 mM NaCl+1 mM ATP). The straight lines are

a linear fit of the data using the integrated form of Wyman linkage

equation (Eqn 2). Experimental conditions: 30 mM Hepes buffer,

pH 7.5 and 20 �C.

Table 2. Change in the number of water molecules (Dnw) ± SD bound to the anodicHb in the transition from fully deoxy to fully ‘oxy’ forms, measured

by tetramer in different experimental conditions. Experiments for the anodic Hb were performed in 30 mM Hepes buffer pH 7.5 and 20 �C. The slopes
were compared using the stripped condition as a reference.

Sample Experimental condition

Dnw ± SD

Wyman Statistical analysis

Correlation

coefficient

Hban Sucrose, stripped Hb 58 ± 08 Reference 0.965

Sucrose + ATP 1 mM 68 ± 12 ns 0.891

Sucrose + ATP 0.1 mM 68 ± 12 ns 0.888

Sucrose + NaCl 100 mM 116 ± 16 * 0.993

Sucrose + ATP 1 mM + NaCl 100 mM 28 ± 08 ns 0.873

ns ¼ P > 0.05, *P < 0.001.

4272 P. Peres et al. (Eur. J. Biochem. 271) � FEBS 2004



It was possible to calculate the ATP association constants
to the oxygenated and deoxygenated Hb according to
Eqn (4). The value of the binding constant in the deoxy-
genated form (KD) was 2.2 · 105 ± 1.3 · 104 M

)1, and for
the oxygenated form (KO) was 2.6 · 102 ± 3.3 · 101 M

)1.

Discussion

Hemoglobin O2 equilibria as a function of water activity

The analysis of conformational changes by the osmotic
stress approach [2] has proven to be reliable, despite its
experimental simplicity, as direct measurements of water
binding by a crystal quartz microbalance [17] showed
agreement with the calculated Dnw. Using water activity as
a probe allows, accordingly, to analyze conformational
changes induced by allosteric effectors that would be
difficult or expensive to follow by other methods, and this
possibility has been used by other authors to study Hbs
[16,19]. BecauseH. littorale’s anodic and cathodic Hbs have
been functionally well described by Weber et al. [5], we
decided to focus our analysis on their conformational
transitions, and secondarily on phosphate binding.

Although having very different oxygen-binding proper-
ties, bothHbs respond toan increase inwater activitywithan
increase onoxygenaffinity, indicatingpreferential bindingof
water molecules to the R state, also reported for other
vertebrate Hbs [16,19], despite their functional differences.

Concerning the values found forDnw during oxygenation,
for the cathodic Hb, in the stripped condition the value is
smaller than in the presence of saturating levels of Cl– or
ATP, suggesting that in the absence of anions, the Hb
assumes a new conformational state, different from the
classical T state (Tx), adopting the intermediary state,
denominated T0, more hydrated than the Tx. This fact is in
agreement with the findings reported by Colombo and
Seixas [3] for humanHb, and it shows that thisHb, although
showing a significant reverse Bohr effect, follows a pattern
that has already been described for human Hb, which has a
normal response to proton binding. Hemoglobins with a
reverse Bohr effect appear to have some relationshipwith air

breathing, as they appear in fishes and amphibians with
adaptations, as pointed out byWeber et al. [5]. Interestingly,
the anodic Hb showed a distinct behavior, with chloride
exerting the only significant effect on its conformation. This
high value is close to that reported for the anodic eel Hb in
the presence ofKCl andGTP (� 118), although the authors
did not test chloride alone [16].

The fact that the O2 affinity from the anodicHb increased
in the presence of chloride or low phosphate concentrations
was first reported by Weber et al. [5], using data gathered
in the presence of low concentrations of NaCl and 2,3-
biphosphoglycerate (2,3-BPG) at pH 7.5. The unexpected
increase in the O2 affinity could be interpreted as a result of
binding to the R state, as proposed by the previous authors,
but could also suggest an excess of negative charges in the
Hb central cavity (a1–b2 interface), and anion binding to
this region would destabilize the interdimeric interface.
The large Dnw obtained for this Hb is probably related to
the role exerted by chloride binding, but its explanation
would require primary sequence determination and crystal-
lographic analysis or at least molecular modeling, using a
crystallized Hb as a template.

When we compareDnw values obtained in the presence of
ATP and chloride, however, for both Hbs, the last anion
induces larger conformational changes, evidence that phos-
phate at high concentrations can lock the Hb structure in
a T-like conformation, similar to previous findings from
Caôn [18].

Calculation of the ATP association constants to the
oxygenated and deoxygenated forms of the cathodic Hb

The value obtained for the association constant of ATP to
the deoxygenated form (KD) of the cathodic Hb is about 10
times larger regarding 2,3-BPG binding to human Hb
(KD ¼ 3.6 · 104 M

)1) [20], showing that ATP binds to the
cathodic Hb more strongly than 2,3-BPG to human Hb.
The obtained value is similar to found for Hb-II of the fish
Piaractus mesopotamicus (3.1 · 105 M

)1) [21]. Concerning
the estimative for the ATP association constant to the
oxygenated form (KO), this agrees with that reported for
oxygenated Hb human (Ko ¼ 3.5 · 102 M

)1), and greater
than for P. mesopotamicus Hb (2.7 · 101 M

)1). This strong
phosphate binding, combined with a reverse Bohr effect,
would ensure effective control of O2 uptake with high
affinity, as well as its delivery by the interplay of the pH and
phosphate concentration within the red blood cells.

In conclusion, we showed that bothHbs investigated here
respond to an increase in water activity by stabilizing the
R state conformation, and that the presence of an inter-
mediate conformational state controlled by anion binding in
the oxygenation process is similar amongst Hbs, similarly as
found by other authors [3,16]. We did not observe evidences
of the presence of additional phosphate binding sites to the
cathodic Hb, as suggested by Weber et al. [5].

Acknowledgements

We thank Dr Márcio F. Colombo, who supplied the Osmometer for

osmolality measurements. This work was supported by grants from

FAPESP (01/11553–3 and 03/00085–4), FUNDUNESP (474/04) and

CNPq.

Fig. 3. Variation of Hbct oxygen affinity vs. free ATP concentration at

pH 7.5. The concentration of ATP, varied from 0 to 35 mM. The

symbol o represents value of logP50 in absence of ATP.

� FEBS 2004 Allosteric effects in Hoplosternum littorale Hbs (Eur. J. Biochem. 271) 4273



References

1. Royer, W.E. Jr, Pardanani, A., Gibson, Q.H. & Peterson, E.S.

(1996) Ordered water molecules as key allosteric mediators in a

cooperative dimeric hemoglobin. Proc. Natl Acad. Sci. USA 93,

14526–14531.

2. Colombo, M.F., Rau, D.C. & Parsegian, V.A. (1992) Protein

solvation in allosteric regulation: a water effect on hemoglobin.

Science 256, 655–659.

3. Colombo, M.F. & Seixas, F.A.V. (1999) A novel allosteric con-

formation of human Hb revealed by the hydration and anion

effects on O2 binding. Biochemistry 38, 11741–11748.

4. Colombo,M.F.&Bonilla-Rodriguez,G.O. (1996)The water effect

on allosteric regulation of hemoglobin probed in water/glucose

and water/glycine solutions. J. Biol. Chem. 271, 4895–4899.

5. Weber, R.E., Fago, A., Val, A.L., Bang, A., VanHauwaert, M.L.,

Dewilde, S., Zal, F. & Moens, L. (2000) Isohemoglobin differ-

entiation in the bimodal-breathing Amazon catfish Hoplosternum

littorale. J. Biol. Chem. 275, 17297–17305.

6. Amiconi, G., Bertollini, A., Bellelli, A., Coletta, M., Condo, S.G.

& Brunori, M. (1985) Evidence for two oxygen-linked binding

sites for polyanions in dromedary hemoglobin. Eur. J. Biochem.

150, 387–393.

7. Tamburrini, M., Riccio, V., Romano, M., Giardina, B. & di

Prisco, G. (2000) Structural and functional analysis of the two

haemoglobins of the antarctic seabird Catharacta maccormicki.

Characterization of an additional phosphate binding site by

molecular modeling. Eur. J. Biochem. 267, 6089–6098.

8. Riccio, A., Tamburrini, M., Giardina, B. & di Prisco, G. (2001)

Molecular dynamics analysis of a second phosphate site in the

hemoglobins of the seabird, south polar skua: is there a site–site

migratory mechanism along the central cavity? Biophys. J. 81,

1938–1946.

9. Feuerlein, R.J. & Weber, R.E. (1996) Oxygen equilibria of

cathodic eel hemoglobin analysed in terms of the MWC model

and Adair’s successive oxygenation theory. J. Comp. Physiol. B

165, 597–606.

10. Giardina, B. & Amiconi, G. (1981) Measurement of binding

of gaseous and nongaseous ligands to hemoglobin by

conventional spectrophotometric procedures. Methods Enzymol.

76, 417–427.

11. Benesch, R., Macduff, G. & Benesch, R.E. (1985) Determination

of oxygen equilibria with a versatile new tonometer. Anal. Bio-

chem. 11, 81–87.

12. Wyman, J. (1964) Linked functions and reciprocal effects in

hemoglobin: a second look. Adv. Prot. Chem. 19, 223–286.

13. Wyman, J. & Gill, S.J. (1990) Binding and Linkage: Functional

Chemistry of Biological Macromolecules. University Science

Books, Mill Valley, California.

14. Zar, J.H. (1999) Biostatistical Analysis, 4th edn. Prentice Hall,

Upper Saddle River, NJ, USA.

15. Bonilla, G.O., Nagatomo, C.L.,Matsuura,M.S.A., Focesi, J.R. &

A. (1994) Interactions of adenosine triphosphate with snake

hemoglobin studies in Liophis miliaris, Boa constrictor and

Bothrops alternatus. Compar. Biochem. Physiol. B. 109, 701–707.

16. Hundahl, C., Fago, A., Malte, H. &Weber, R.E. (2003) Allosteric

effect of water in fish and human hemoglobins. J. Biol. Chem. 278,

42769–42773.

17. Salvay, A.G., Grigera, J.R. & Colombo, M.F. (2003) The role of

hydration on the mechanism of allosteric regulation: in situ mea-

surements of the oxygen-linked kinetics of water binding to

hemoglobin. Biophys. J. 84, 564–570.

18. Caôn, O. (2003) Contribuição Da Solvatação Á Entalpia E Entr-

opia de Oxigenação de Hemoglobinas. PhD Thesis, Universidade

Estadual Paulista Júlio deMesquita Filho, São José do Rio Preto,

SP, Brazil.

19. Müller, G., Fago, A. & Weber, R.E. (2003) Water regulates

oxygen binding in hagfish (Myxine glutinosa) hemoglobin. J. Exp.

Biol. 206, 1389–1395.

20. De Rosa, C.M., Sanna, T.M., Messana, I., Castagnola, M., Gal-

tieri, V., Tellone, V., Scatena, R., Botta, B., Botta,M. &Giardina,

B. (1998) Glycated human hemoglobin (HbAlc): functional char-

acteristics and molecular modeling studies. Biophys. Chem. 72,

323–335.

21. Poy, C.D. (2001) Purificação e caracterização parcial das hemo-

globinas de Pacu (Piaractus Mesopotamicus, Pisces). Dissertation,

Universidade Estadual Paulista Júlio de Mesquita Filho, São José

do Rio Preto, SP, Brazil.

4274 P. Peres et al. (Eur. J. Biochem. 271) � FEBS 2004