An ether-linked tetrafunctional acylating reagent and its cross-linking reactions with hemoglobin

Article abstract of DOI:10.1139/v98-236

A new type of tetrafunctional reagent for cross-linking proteins has been prepared and used to modify human hemoglobin A. DPEE (1,2-bis{2-[3,5-bis(3,5-dibromosalicyloxycarbonyl) phenoxy]ethoxy}ethane)) has two separate pairs of reacting sites connected by

Full text of DOI:10.1139/v98-236

271
An ether-linked tetrafunctional acylating reagent
and its cross-linking reactions with hemoglobin
Ronald Kluger, Krisztina Paal, and J. Gordon Adamson
Abstract: A new type of tetrafunctional reagent for cross-linking proteins has been prepared and used to modify
human hemoglobin A. DPEE (1,2-bis{2-[3,5-bis(3,5-dibromosalicyloxycarbonyl) phenoxy]ethoxy}ethane)) has two
separate pairs of reacting sites connected by a flexible tetraether chain. DPEE is capable of connecting a cross-link
within a protein to another cross-link, either within the same protein molecule or between molecules. DPEE was
readily prepared by esterification of a tetraether-linked bisphthalate (prepared by coupling of 1,2-bis(2-
iodoethoxy)ethane and 5-hydroxyisophthalic acid). DPEE reacts with deoxy hemoglobin to produce a mixture of
modified proteins. Ion-exchange HPLC was used to separate the modified proteins in the mixture. The most abundant
products were selected for structural analysis, which used data from reverse-phase chromatography and tryptic peptide
mapping. To prevent dissociation of the modified proteins during analysis, the products were further reacted with the
bifunctional reagent, bis(3,5-dibromosalicyl) fumarate, which produces fumaryl cross-links between α-subunits. From
peptide analysis of the separated products, the major modified protein from DPEE was identified as a novel species
with four links within the same α2β2 tetramer. In addition, a minor product that involves cross-links in two different
proteins was observed. These results imply that the reagent reacts primarily in a folded state within the protein.
Key words: acylation: cross-linking, multifunctional, hemoglobin, reaction pattern.
Résumé : On a préparé un nouveau type de réactif tétrafonctionnel pour effectuer la réticulation des protéines et
on l’a utilisé pour modifier de l’hémoglobine humaine A. Le DPEE (1,2-bis{2-[3,5-bis(3,5-dibromosalicyloxycarbonyl)
phénoxy]éthoxy}éthane) possède deux paires séparées de sites réactionnels liés par une chaîne tétraéther flexible. Le
DPEE peut relier une réticulation à l’intérieur d’une protéine avec une autre réticulation de la même molécule de pro-
téine ou entre des molécules. On peut facilement préparer le DPEE par estérification d’un bisphtalate relié à un tétraé-
ther (préparé par couplage du 1,2-bis(2-iodoéthoxy)éthane et de l’acide 5-hydroxy-isophtalique). Le DPEE réagit avec
la désoxyhémoglobine pour produire un mélange de protéines modifiées. On a fait appel à la CLHP par échange d’ions
pour séparer les protéines modifiées présentes dans le mélange. On a choisi les produits les plus abondants pour réali-
ser une analyse structurale en utilisant des données obtenues par chromatographie en phase inversée et par cartographie
d’un peptide tryptique. Dans le but d’empêcher la dissociation des protéines modifiées au cours de l’analyse, on a fait
réagir les produits avec le réactif bifonctionnel, fumarate de bis(3,5-dibromosalicyle) qui produit des réticulations fuma-
ryles entre les sous-unités α. Sur la base de l’analyse peptidique des produits séparés, on a déterminé que la protéine
principale découlant de la modification par le DPEE est une nouvelle espèce comportant quatre liaisons à l’intérieur du
même tétramère α2β2. On a aussi observé la présence d’un produit mineur qui comporte des réticulations dans deux
protéines différentes. Ces résultats impliquent que le réactif agit principalement dans un état replié de la protéine.
Mots clés : acylation : réticulation, multifonctionnel, hémoglobine, patron de réaction.
[Traduit par la Rédaction] Kluger et al. 279
moglobin from dissociating into dimers contains two reac-
tive functional groups that create a cross-link (12–14). The
Human hemoglobin A is an α 2β2 tetramer that reversibly
dissociates into αβ dimers (1, 2). Cross-linking prevents dis-
sociation into dimers while providing the possibility for sys-
tematic alteration of the protein’s properties (3-8). Cross-
linked hemoglobins are of special interest since they may
have clinical significance as red cell substitutes in transfu-
sion medicine (9–11). A minimal reagent that prevents he-
properties of the resulting altered protein depend on the sites
that are connected and the structure of the link (7, 14–18).
Reagents with three reactive functional groups can produce
a cross-link between two sites within a protein, leaving a
functional group available for bioconjugation (19). Alterna-
tively, all three groups can react within the protein, creating
a novel cross-linked structure (7, 20, 21).
Received October 13, 1998.
K. Paal and R. Kluger.1 Lash Miller Laboratoires, Department of Chemistry, University of Toronto, Toronto, ON M5S 3H6,
Canada.
J.G. Adamson. Hemosol, Inc., 115 Skyway Avenue, Toronto, ON M9W 4Z4, Canada.
1Author to whom correspondence may be addressed. Telephone: (416) 978-3582. e-mail: rkluger@chem.utoronto.ca
Can. J. Chem. 77: 271–279 (1999)
© 1999 NRC Canada

272
Can. J. Chem. Vol. 77, 1999
Reagents with four reactive centers present additional
results from size-exclusion chromatography and tryptic pep-
tide mapping (26) enables us to propose structures for the
major modified proteins that result from reaction of DPEE
and hemoglobin.
possibilities for creating novel modified proteins (22, 23). A
reagent that has two pairs of connected reactive sites can
produce two cross-links that are connected by the core of the
reagent. Depending on the geometry of the core, the two
cross-links can form within one protein molecule or in two
separate molecules, resulting in modified-connected pro-
teins. If one or two functional groups of the reagent remain
intact in the reaction with the protein these sites can be uti-
lized for further conjungation reactions.
We have reported that a relatively small reagent with four
reactive groups, BTDS (3,5,3′,5′-biphenyltetracarbonyl
tetrakis(3,5-dibromosalicylate)), reacts with deoxy hemoglo-
bin to give principally a product that has a stable cross-link
between the β-subunits. The two additional reactive groups
of the reagent do not combine with the protein (22). It is
likely that once reaction has occurred at two of the groups,
the geometric relationship of the remaining groups prevents
them from being in proximity to nucleophiles in the same
protein or in a second protein.
Commercial reagents and chemicals were used without
further purification. Solvents were dried prior to use. Re-
agents for preparation of buffers were analytical grade or
better. Purified human hemoglobin A was from Hemosol,
Inc. Proton NMR spectra were obtained at 400 MHz or
200 MHz and 13C NMR spectra at 100 MHz.
The structures of the materials associated with the synthe-
sis of the DPEE are shown in Schemes 1 and 2 in the Re-
sults section. The synthesis of the linked esters follows the
general procedure of Collman et al. (4).
Synthesis of 1,2-bis{2-[3,5-bis(ethoxycarbonyl) phenoxy]-
ethoxy}ethane (2)
Br
HO2C
HO2C
Br
Diethyl 5-hydroxyisophthalate (0.95 g, 0.0040 mol) and
potassium tert-butoxide (0.47 g, 0.0042 mol) were stirred in
20 mL anhydrous tetrahydrofuran at room temperature for
30 min. A solution of 0.74 g (0.0020 mol) of 1,2-bis(2-
iodoethoxy)ethane (1) in 20 mL anhydrous tetrahydrofuran
was added and the reaction mixture was refluxed for 48 h
under nitrogen. Water was added and the product was ex-
tracted with ether. The solution was dried over magnesium
sulfate, filtered, and the solvent removed. Separation on a
silica column (3 cm × 30 cm, hexanes – ethyl acetate 1:1)
Br
Br
O
O
O
O
O
O
CO2H
CO2H
Br
O
O
Br
Br
1
produced a white powder (0.60 g, 51%). H NMR (CDCl3)
Br
δ: 8.25 (2 H, t, J = 1.2 Hz, ArH), 7.74 (4 H, d, J = 1.2 Hz,
ArH), 4.36 (8 H, q, J = 6.9 Hz, CH2), 4.20 (4 H, t, J =
5.6 Hz, CH2), 3.87 (4 H, t, J = 5.6 Hz, CH2), 3.75 (4 H, s,
CH2), 1.36 (12 H, t, J = 6.9 Hz, CH3).
BTDS
To arrive at new types of altered proteins from a reagent
with four reactive groups, we therefore assumed that a greater
separation of the groups is necessary. In this paper we report
the synthesis of DPEE (1,2-bis{2-[3,5-bis(3,5-dibromosal-
icyloxycarbonyl) phenoxy]ethoxy}ethane). DPEE contains
isophthalyl dibromosalicylate reaction sites, as in BTDS, but
they are connected in a longer, more flexible, and more hy-
drophilic array.
1,2-Bis[2-(3,5-dicarboxyphenoxy) ethoxy]ethane (3)
The tetraethyl ester (0.60 g, 0.0010 mol, 2) was dissolved
in 30 mL ethanol. The solution was heated to 65°C. Then
0.25 g (0.0060 mol) sodium hydroxide dissolved in 1 mL
water was added and the reaction mixture kept at 65°C for
22 h. The resulting white precipitate was filtered and dis-
solved in 10 mL of water. The tetraacid (3) was precipitated
by addition of concentrated hydrochloric acid and filtered.
The product was dried under vacuum (0.46 g, 95%). 1H
NMR (DMSO-d6) δ: 8.05 (2 H, s, ArH), 7.62 (4 H, s, ArH),
4.19 (4 H, s, CH2), 3.75 (4 H, s, CH2), 3.60 (4 H, s, CH2); a
broad band is also present under the three high-field signals.
Br
Br
Br
Br
O
O
C
O
O
HO
HO
O
C OH
O
O
C
O
O
O
O
O
O
OH
1,2-Bis{2-[3,5-bis(chlorocarbonyl) phenoxy]ethoxy}-
ethane (4)
C
O
O
O
Br
The tetraacid (0.23 g, 0.0047 mol, 3) was dissolved in
15 mL thionyl chloride containing three drops of dimethyl-
formamide. The solution was refluxed under nitrogen for 5 h
and then the solvent was evaporated. The product (yellow
oil) was dried under vacuum and solidified overnight (0.26 g,
Br
Br
Br
DPEE
1
100%). H NMR (CDCl3) δ: 8.43 (2 H, t, J = 1.6 Hz, ArH),
Chemically modified hemoglobins are readily amenable
to ion spray mass spectral analysis (21, 24, 25). Combining
the results of such an analysis for separated materials with
7.90 (4 H, d, J = 1.6 Hz, ArH), 4.24 (4 H, t, J = 5 Hz, CH2),
3.89 (4 H, t, J = 5 Hz, CH2), 3.74 (4 H, s, CH2). This
tetraacid tetrachloride was used without further purification.
© 1999 NRC Canada

Kluger et al.
273
C-4 reverse-phase HPLC column. In addition to known
peaks corresponding to heme and native globin chains, new
peaks appeared that correspond to cross-linked globin
chains. The molecular weights of these materials were deter-
mined by ion spray mass spectrometry (30, 31). The sites of
modification were deduced by comparison of peptide pat-
terns (reverse-phase HPLC) from digestion of native and
modified chains with trypsin and endoproteinase Glu-C (14,
22, 32–34) along with ion spray mass spectral results as has
been described in detail for other cross-linked hemoglobins
(24, 27). Reference data for weights and mass spectra of na-
tive globin chains are conveniently summarized in the work
of Adamczyk and Gebler (25). Size exclusion chromatogra-
phy was used to distinguish materials by their approximate
molecular weights (35). Columns were calibrated with cross-
linked and oligomerized hemoglobin.
The formation of a cross-link between α-subunits of
DPEE-Hb with fumaryl bis(3,5-dibromosalicylate) was ac-
complished by the procedure described for production of
doubly cross-linked hemoglobin (14). After the 20 h incuba-
tion period of deoxy hemoglobin with DPEE (described
above), a deoxygenated solution of fumaryl bis(3,5-dibro-
mosalicylate) (0.0035 g, 5.0 × 10–6 mol) in 4.0 mL 50 mM
sodium borate buffer (pH 8.0) was added to the hemoglobin
reaction solution under nitrogen. (This reagent should intro-
duce a cross-link between the ε-amino groups of the α-99
lysyl residues of hemoglobin (14, 16). It is effective if the
sites in the β-subunits that make up the DPG-binding site (3,
36) are blocked.) The mixture was kept at 37ЊC for 2 h and
then the flask was flushed with carbon monoxide. The solu-
tion of twice-modified carbonmonoxy hemoglobin was
passed through a Sephadex G-25 column (2.5 × 15 cm)
equilibrated with 50 mM sodium phosphate buffer (pH 7.5)
in order to remove residual reagent. (The resulting material
is referred to as XL2-Hb.) C-4 reverse-phase HPLC columns
were used to separate the hemes and the globin chains. In
another set of analyses of XL2-Hb, gel filtration chromatog-
raphy was used to separate modified hemoglobins by their
gross molecular weight (size). We used size-exclusion col-
umns under conditions that dissociate the αβ dimers from
hemoglobin that is not cross-linked between these dimers,
and also used conditions under which hemoglobin remains
tetrameric (“nondissociating conditions”). Elution was under
isocratic conditions. The nondissociating eluent was
0.050 M pH 7.0 sodium phosphate with 0.15 M sodium
chloride flowing through a PSEC-12 column at 0.5 mL min–1.
The dissociating eluent was 0.025 M pH 7.2 Tris with 0.5 M
magnesium chloride flowing through a TSK G2000 swxl
column at 0.2 mL min–1.
1,2-Bis{2-[3,5-bis(3,5-dibromosalicyloxycarbonyl)
phenoxy]ethoxy}ethane (DPEE, 6)
tert-Butyl 3,5-dibromosalicylate (0.66 g, 0.0019 mol) (27,
28) and potassium tert-butoxide (0.21 g, 0.0019 mol) were
stirred in 15 mL anhydrous tetrahydrofuran at room temper-
ature for 30 min. Then a solution of 0.26 g (0.00047 mol) of
4 in 15 mL anhydrous tetrahydrofuran was added and the re-
action mixture was stirred at room temperature overnight.
Water was added and the product was extracted with ether.
The extract was dried over magnesium sulfate, filtered, and
the solvent removed, leaving the tetra tert-butyl ester (5).
The white solid was dissolved in 15 mL anhydrous trifluoro-
acetic acid (to convert the tert-butyl ester to the free acid)
and left at room temperature for 2.5 h. Dry ether was added
(15 mL) and a white precipitate formed. The product (6),
DPEE, was filtered and dried in vacuum (white powder,
0.55 g, 74%). IR (KBr): 3177 (br s, νOH), 1736 (s, νCϭO),
1195 (s, νC—O) cm–1; 1H NMR (DMSO-d6) δ: 8.36 (2 H, t,
J = 1.4 Hz, ArH), 8.33 (4 H, d, J = 2.5 Hz, ArH), 8.09 (4 H,
d, J = 2.5 Hz, ArH), 7.98 (4 H, d, J = 1.4 Hz, ArH), 4.31 (4
H, br s, CH2), 3.79 (4 H, br s, CH2), 3.63 (4 H, s, CH2); a
broad band is also present under the three high-field signals;
13C NMR (acetone-d6) δ: 163.57, 163.25, 160.59, 148.36,
140.13, 134.68, 131.79, 128.06, 124.54, 122.12, 120.29, 120.06,
71.50, 70.29, 69.47; MS (negative FAB), M – 1 calcd. for
C50H30Br8O20: 1589; found: 1589.
Reactions of human hemoglobin A with DPEE were car-
ried out as described previously (22). A typical example is
described here. A solution of carbonmonoxy hemoglobin
(2.0 mL, 1.25 mM, 0.000025 M) in 50 mM Bis-Tris buffer
(pH 6.5) was passed through a Sephadex G-25 column (2.5
× 15 cm) equilibrated with 50 mM sodium borate buffer (pH
8.0). The carbonmonoxy hemoglobin was converted to oxy
hemoglobin by irradiation under flowing oxygen at 0ЊC for
2 h. The solution was then kept under flowing nitrogen at
37ЊC for 2 h to convert the hemoglobin to the deoxy form.
DPEE (0.0080 g, 5.0 × 10–6 mol) was dissolved in 1.0 mL of
dioxane, and 3.0 mL of 50 mM sodium borate buffer (pH
8.0) was added. Oxygen was removed from the solution by
evacuation followed by addition of nitrogen. This was re-
peated three times. The material was then added under nitro-
gen to the deoxy hemoglobin solution. The reaction mixture
was kept at 37ЊC for 20 h with humidified nitrogen flowing
through the rotating flask. The flask was then flushed with
carbon monoxide. The solution of modified carbonmonoxy
hemoglobin was passed through a Sephadex G-25 column
(2.5 × 15 cm) equilibrated with 50 mM sodium phosphate
buffer (pH 7.5) to remove residual reagent. (The resulting
material is referred to as DPEE-Hb.) C-4 reverse-phase
HPLC columns were used to separate heme and the various
modified and native globin chains under denaturing condi-
tions. Hemoglobins of different molecular weights were re-
solved using gel filtration chromatography, both with high
concentrations of magnesium salt that lead to dissociation
into αβ dimers (29) as well as under nondissociating condi-
tions where the tetramers remain intact.
The synthesis of DPEE efficiently introduces the neces-
sary paired reactive functional groups for cross-linking onto
a flexible linking chain. The route could be extended to pro-
duce a reagent library by varying the length of the polyether
linker. The synthesis of the ether is based on Collman’s
method for producing a tetraether-linked tetraester (4). The
reactive groups are conveniently introduced by Klotz’s
procedure for forming dibromosalicylates (28) (Schemes 1
and 2).
General procedures for structural analysis were as previ-
ously reported (22). The product mixture was separated on a
© 1999 NRC Canada

274
Can. J. Chem. Vol. 77, 1999
Scheme 1.
CO2Et
CO2Et
KO
I
O
O
I
THF
1
O
O
EtO
OEt
O
O
O
O
NaOH, H2O
EtOH
O
O
O
O
O
O
EtO
OEt
O
O
O
O
2
HO
HO
OH
OH
O
SOCl2/DMF
O
O
O
O
3
Cl
Cl
Cl
Cl
O
O
O
4
DPEE reacts with deoxy hemoglobin but not with carbon-
monoxy hemoglobin, based on analysis of the resulting glo-
bin chains by C-4 reverse-phase HPLC. This is consistent
with reaction occurring at the binding site for DPG in hemo-
globin, which is more accessible in the deoxy form (37). The
observed product pattern of the reaction remained constant
as the ratio of the reagent to hemoglobin was varied from
1:2 to 2:1 and temperature was changed from 37 to 50°C.
The reaction mixture appears to contain a considerable vari-
ety of altered protein products. The relative amounts were
estimated based on peak areas in the chromatograms. We
chose the materials in the largest peaks for structural analy-
sis.
from other peaks retain one or more 3,5-dibromosalicyl
(DBS) esters from DPEE or are partially hydrolyzed. Struc-
tures consistent with the observed masses are indicated as
substituted derivatives of the core structure below:
O
R
1
O
R
3
O
O
O
O
R
2
R
4
O
O
7–13
The heme and globin chains were separated on a C-4 re-
verse-phase HPLC column (see Fig. 1). The products in the
largest peaks were detected and their molecular weights
were determined from the parent peaks of ion spray mass
spectra. Table 1 summarizes the ion spray mass spectral as-
signments associated with major peaks that were observed
on reverse-phase HPLC analysis. The product from the larg-
est chromatographic peak (7) is a tetraamide formed by
acylation at four amino groups of two β-subunits. Products
The mass of the linked globin chain that is proposed as
the structure of the major product (7) is consistent with the
electrospray mass spectral peak. This is a dimer of two β-
subunits linked by four amide connections derived from the
core of DPEE (14), which has a molecular mass of 410. The
molecular mass of a β-subunit is 15 868 and that of the α-
subunit is 15 126. Combining the masses of two β-subunits
and 14, less the mass of hydrogens that are dissociated from
© 1999 NRC Canada

Kluger et al.
275
O
Scheme 2.
Br
O
O
O
KO
tBuO2C
THF
Br
Cl
Cl
Cl
O
_
(OCH2CH2)3
O
O
Cl
O
4
O
Br
Br
tBuO2C
tBuO2C
Br
Br
O
O
O
O
CO2tBu
CO2tBu
O
O
TFA
_
(OCH2CH2)3
O
O
O
O
O
Br
Br
O
Br
Br
Br
Br
5
O
Br
Br
O
14
O
O
O
CO2H
CO2H
Br
HO2C
HO2C
Br
O
O
_
(OCH2CH2)3
O
O
O
Br
Br
O
NH
O
6
Val-β'-1
HN
To determine the residues within hemoglobin that are
modified by DPEE in 7, the material from the peak corre-
sponding to the cross-linked β-subunits was subject to
tryptic digestion and peptide mapping. The products of di-
gestion were separated by C-18 reverse-phase HPLC and the
peptides were compared to those from digestion of the un-
modified β chain (26). Three fragments of the native β digest
were absent from that of the modified β-subunits: β-T1, β-
T9, and β-T10a′. β-T1 is the amino terminus (Val-β-1) and β-
T9 is connected to β-T10a′ through the β-Lys-82 residue, in-
dicating that the amino termini of both β-subunits as well as
the ε-amino groups of both β-Lys-82 residues are acylated
by DPEE (Fig. 2). (The other peaks listed in Table 1 repre-
sent smaller quantities and the structural assignments are
based on masses, with attachment sites based on the consis-
tent patterns of modification seen with other reagents with
similar functionality. Digests were not performed for these.)
The results for 7 indicate that there are two sets of cross-
links connecting positions 1 and 82 of β-subunits introduced
by a single tetrafunctional reagent (15).
HN
O
O
HN
O
Lys−β -82
O
O
O
NH
O
O
NH
O
NH
Lys−β'-82
HN
O
O
Val−β -1
15
Formation of bis tetramers
The four functional groups of DPEE are arranged so that
it is possible for them to produce a connection between
tetramers. If we assume that reaction occurs at positions 1 or
each resulting amide, gives a predicted mass of 32 142 for 7.
This is consistent with what is observed.
© 1999 NRC Canada

276
Can. J. Chem. Vol. 77, 1999
Table 1. Ion spray mass spectral assignments for altered globin chains derived from the reaction of DPEE with hemoglobin. The
structures are indicated based on the core structure above connected to hemoglobin β-subunits at the indicated amino group. A
hydrolyzed site is indicated as “OH” and an intact ester as “DBS”. Mass assignments for the native peaks are based on literature
summations (25). The structure 12 is listed as two possibilities where either ester group from DPEE has undergone hydrolysis.
Mass (found)
Mass (calcd.)
Label
R1
R2
R3
R4
32 140
32 163
16 421
30 971
31 250
31 713
31 713
31 444
32 142
32 160
16 420
30 972
31 250
31 714
31 714
31 436
7
8
9
10
11
12a
12b
13
NH(ε)-β-Lys-82
NH(ε)-β-Lys-82
NH(ε)-α-Lys-99
NH(ε)-α-Lys-99
NH(ε)-α-Lys-99
NH(ε)-β-Lys-82
NH(ε)-β-Lys-82
NH(ε)-β-Lys-82
NH(α)-β′-Val-1
NH(α)-β′-Val-1
DBS
NH(ε)-α′-Lys-99
NH(ε)-α′-Lys-99
DBS
NH(ε)-β′-Lys-82
NH(ε)-β′-Lys-82
DBS
DBS
DBS
NH(ε)-α′-Lys-99
NH(ε)-α′-Lys-99
NH(ε)-α′-Lys-99
NH(α)-β-Val-1
OH
DBS
OH
DBS
OH
OH
OH
DBS
OH
Table 2. Expected masses (kDa) of peaks of modified hemo-
globins. Addition of 0.5 M magnesium chloride causes dimers
that are not cross-linked to separate from tetramer. See text for
details (indicated as “dissociating”).
Fig. 1. Globin chain separation after reaction of deoxy
hemoglobin with DPEE: 20 h at pH 8.0 and 37°C
Mass
(dissociating) after
Mass
formation of α–α
Structure
Mass
(dissociating)
cross-link
16
17
18
19
128
128
64
64, 32
128
64
128
128
64
64
64
64
will give material that does not dissociate into dimers.
Therefore, introducing the α–α cross-link should not change
the effect of adding magnesium chloride. These possibilities
are summarized in Table 2.
We note that α–α cross-linking permits separations that
1
α β1
β α
82
82
'
'
'
'
'
β α
α β
82 of β-subunits, and the product is symmetrical, schematic
structures 16 or 17 are possible. To determine whether any
bis tetramer formed, we analyzed the reaction solutions on
size exclusion columns that give an indication of mass based
on calibrated globular size. Since hemoglobin tetramers
(64 000 Da) are in equilibrium with αβ dimers (32 000 Da)
(38), cross-linking between two β-subunits or between two
α-subunits prevents dissociation while a link within a sub-
unit does not. Addition of magnesium chloride to the solu-
tion influences the equilibrium toward dissociation into
dimers (29). Thus, 16 will dissociate into dimers while 17
will not.
16
1
1
α β
β α
82
82
1
' ' 1
α β
'
β α
82
82
17
Hemoglobins that are doubly cross-linked, with a link in
between α- as well as β-subunits, can be prepared by addi-
tion of the reagent bis(3,5-dibromosalicyl)fumarate (DBBF),
to a ββ-cross-linked hemoglobin (14). Thus when the α-sub-
units of 16 are subsequently cross-linked by reaction with
DBBF, the tetrameric protein will no longer dissociate into
dimers. In the possible cross-linked tetrameric structures,
shown as 18 and 19, reaction with all four sites of DPEE
distinguish 16 from 17. Structure 17 has cross-links between
β-subunits, preventing dissociation into dimers. Cross-
linking between the α-subunits does not change the results
for dissociation. Therefore, formation of 17 should not change
the apparent product–mass distribution. However, two of the
four αβ dimers of 16 can dissociate. If an α–α cross-link is
© 1999 NRC Canada

Kluger et al.
277
Fig. 2. Tryptic-Glu-C peptides of the β-subunits linked at four sites by DPEE. Comparison is noted against native protein digest.
introduced, there will be a decrease in the amount of 32K
material present.
Structures 18 and 19 are both capable of generating the
major product. Both are fully cross-linked and will not dis-
sociate into dimers. Both have a molecular mass of 64K.
Both have all the β-1 and β-82 amino groups converted to
amides.
on reverse-phase HPLC analysis. The rates of the reactions
between hemoglobin and DPEE are considerably slower
than with other reagents that have the same dibromosalicyl
leaving groups on aromatic hydrocarbon cores (trimesyl and
biphenyl) (22, 27). These cross-linkers readily react with he-
moglobin and reactions go to completion within a few hours.
It is very likely that the slow reaction of DPEE would be the
result of the hydrophobic end groups of DPEE stacking on
one another, blocking its accessibility, and controlling its re-
action patterns.
Our data show that the major product is a β–β′ cross-
linked tetraamide containing one tetraacyl unit derived from
DPEE. Several different modified hemoglobins could give
data consistent with our results. From the mass and tryptic
digestion analyses we know that cross-links are only be-
tween β-subunits, and all four amino groups from β-Val-1
(N-terminal α-amino group) and β-Lys-82 (ε-amino group)
of both chains must be acylated. From the reaction patterns
of other reagents, we know that the acyl groups on meta po-
sitions of a substituted benzene can connect ε-amino groups
of the lysyl residues at position 82 of each β-subunit. They
also can connect the α-amino group of the N-terminal valine
of one β-subunit and the ε-amino group of Lys-82 of the
other β-subunit. The same amino groups can also be con-
nected within the same subunit. The β-Val-1 amino groups
are too far apart to be connected by the meta-substituent
groups.
1
1
β α
β α
82
1
82
1
'
'
'
'
β α
β α
82
82
18
19
Separation of the globin chains by C-4 reverse-phase HPLC,
before and after treatment of the reaction product with
DBBF, shows that the reagent completely converts all native
α-subunits in the product solution. The results of analysis by
size exclusion chromatography are given in Table 3.
DPEE appears to react with amino groups in the β-sub-
units of deoxyhemoglobin within its DPG-binding site as do
other 3,5-dibromosalicylate esters. However, unlike BTDS,
which reacts at only two of its four ester sites, major prod-
ucts include species in which all four esters are converted to
amide derivatives of hemoglobin. The reaction of DPEE
with deoxy hemoglobin gives one major product (>50%)
and several minor products, including β–β, α–β, and α–α
cross-linked products. Even after a long reaction period,
considerable amounts of native protein subunits are found
If one end of DPEE has reacted between amino groups of
Val-1 and Lys-82 on different β-subunits, then the other end
must react with the alternative Val-1 and Lys-82 amino groups
(18). The schematic diagrams for 20 and 21 (for 18 and 19)
are shown in terms of functional groups (following page).
The alternative products (19 and 21)) have the meta-sub-
stituted dicarbonyls from one end of DPEE connected as
amides on the same β-subunit to Val-1 and Lys-82. The other
end of DPEE is then connected to the other β-subunit in the
© 1999 NRC Canada

278
Can. J. Chem. Vol. 77, 1999
Table 3. Size exclusion chromatography analysis of relative amounts of various modified hemoglobins (DPEE-Hb) and
those that have been reacted further to cross-link their α-subunits (HBXL2). Results are for distribution of relative
amounts, according to their mass. Dissociating conditions refer to the presence of 0.5 M magnesium chloride.
Hemoglobin
Conditions
% ≥128 kDa
% 64 kDa
% 32 kDa
Native (human)
DPEE-Hb
DPEE-Hb
HBXL2
Dissociating
Dissociating
Nondissociating
Dissociating
Nondissociating
0.7
2.5
5.7
5.2
6.1
15.3
89.3
87.7
82.4
86.3
82.0
8.2
6.6
12.4
7.6
HBXL2
O
β
β
Val-1
Val-1
O
O
NH
O
NH
Lys-82
Lys-82
O
NH
NH
O
O
O
O
O
NH
NH
Lys-82
NH
O
Lys-82
NH
Val-1
β
β
O
20
21
While these assignments are self-consistent, we know that
same manner. Based on the reaction patterns of other iso-
phthalyl-DBS esters, we know that the ε-amino group of a β-
Lys-82 is most likely to be converted to an amide. Since the
α -amino groups of each N-terminal β-Val-1 are also modi-
fied in the product, and the meta-substituted reagent cannot
span the distance between the two N-termini, these must re-
act more quickly than does the other β-Lys-82 ε-amino group
with the same end of DPEE. The reagent must arrange itself
within the DPG binding site in order to achieve this selectiv-
ity, perhaps directed by hydrogen-bonding interactions.
The data in Table 3 show that a portion of the reaction
products consists of two connected β-subunits, derived from
different tetramers, producing a bis tetramer (mass: 128 kDa).
Such a species of connected tetramers should be of interest
for studies of cooperativity (1) as well as for analysis of
physiological effects. Approximately half of the 128 kDa
material found in DPEE-Hb under nondissociating condi-
tions is converted to 64 and 32 kDa species under dissociat-
ing conditions (Table 3). This scenario is consistent with
structure 16 (Table 2). The portion of 128 kDa species that
is present under dissociating conditions could be explained
by structure 17, which should be stable under such conditions.
The introduction of cross-links between α-subunits holds
the tetrameric structure together regardless of whether cross-
linking has occurred between the β-subunits (14, 18). Stabi-
lization of DPEE-Hb by α–α cross-linking, giving HBXL2,
makes the 128 kDa species stable under dissociating condi-
tions. These findings are also consistent with structure 16. In
such a case, the reagent would have had to join Val-1 and
Lys-82 on the same subunit within each tetramer. Structure
17 remains intact under dissociating conditions, requiring
that the cross-linking of each tetramer be between different
β-subunits. The results from size exclusion chromatography
confirm that at least one of the minor but significant prod-
ucts is a bis tetramer.
other difunctional meta-substituted DBS reagents react be-
tween the ε-amino groups of each of the β-Lys-82 residues
only (14, 22, 27, 33). The formation of such a product within
one tetramer would require that the two β-Val-1 amino
groups be connected at meta substituents. This is most un-
likely since the distance between the amino termini of the
two β chains of hemoglobin is too great (18.4 Å in the deoxy
and 19.9 Å in the oxy conformation) (37) to be spanned by
the isophthalyl group. No reagent that we have examined,
even with a larger span (32), connects the N-termini of two
β-subunits.
The slow rate of the reaction of DPEE with deoxyhemo-
globin is consistent with either of the four-way cross-linked
tetramers (18 and 19) being the product. Since movement of
the bulky reagent inside the DPG binding site is likely to be
restricted, it should be in a folded conformation as it enters
the cavity. In addition, there is probably a limited orientation
by which DPEE can get into the DPG site, once it is in an
appropriate overall conformation.
The reaction of DPEE with deoxy hemoglobin is slower
than that of smaller reagents, suggesting that the reagent is
less accessible to the protein, presumably because it exists in
a folded state. This has consequences in the reaction pattern
as well. The major product from the reaction of DPEE with
hemoglobin has four amino groups of the α2β2 tetramer, two
from each β-subunit, modified and cross-linked by a single
molecule of reagent. In contrast, the rigid reagent derived
from a symmetrical biphenyl tetracarboxylate (BTDS),
which is unable to fold onto itself, reacts with hemoglobin
primarily at only two sites in any modification process. The
“knotted” species from the DPEE reaction may present an
© 1999 NRC Canada

Kluger et al.
279
15. E. Bucci, A. Razynska, H. Kwansa, B. Matheson-Urbaitis, M.
O’Hearne, J.A. Ulatowski, and R.C. Koehler. J. Lab. Clin.
Med. 128, 146 (1996).
16. J.A. Walder, R.H. Zaugg, R.Y. Walder, J.M. Steele, and I.M.
Klotz. Biochemistry, 18, 4265 (1979).
interesting basis for studies of the protein’s structural dy-
namics. Our results show that DPEE is sufficiently long to
produce bis tetramers but the product is not present to a
large extent, consistent with a predominant folded state of
the reagent. A long but rigid reagent will give such products
in greater proportion.
17. Y. Zheng and K.W. Olsen. Artif. Cells Blood Substitutes Im-
mobilization Biotechnol. 24, 587 (1996).
18. S.R. Snyder, E.V. Welty, R.Y. Walder, L.A. Williams, and J.A.
Walder. Proc. Natl. Acad. Sci. U.S.A. 84, 7280 (1987).
19. R. Kluger and Y. Song. J. Org. Chem. 59, 733 (1994).
20. K.W. Olsen, Q.Y. Zhang, H. Huang, G.K. Sabaliauskas, and T.
Yang. Biomater. Artif. Cells Immobilization Biotechnol. 20,
283 (1992).
This work was supported by the Natural Sciences and En-
gineering Research Council of Canada (Strategic Project).
21. R. Kluger, J. Wodzinska, R.T. Jones, C. Head, T.S. Fujita, and
D.T. Shih. Biochemistry, 31, 7551 (1992).
22. K. Paal, R.T. Jones, and R. Kluger, J. Am. Chem. Soc. 118,
10380 (1996).
23. P.A. Pavlik, M.K. Boyd, and K.W. Olsen. Biopolymers, 39,
615 (1996).
24. Z. Yu, G. Friso, J.J. Miranda, M.J. Patel, T. Lo-Tseng, E.G.
Moore, and A.L. Burlingame. Protein Sci. 6, 2568 (1997).
25. M. Adamczyk and J.C. Gebler. Bioconjugate Chem. 8, 400
(1997).
26. R.T. Jones. Methods Enzymol. 231, 322 (1994).
27. R. Kluger, Y. Song, J. Wodzinska, C. Head, T.S. Fujita, and
R.T. Jones. J. Am. Chem. Soc. 114, 9275 (1992).
28. E.J. Delaney, S.E. Massil, G.Y. Shi, and I.M. Klotz. Arch.
Biochem. Biophys. 228, 627 (1984).
29. G. Guidotti. J. Biol. Chem. 242, 3685 (1967).
30. J.B. Fenn, M. Mann, C.K. Meng, and S.F. Wong. Mass
Spectrom. Rev. 9, 37 (1990).
1. M. Perutz. Nature (London), 228, 1 (1970).
2. R.E. Dickerson and I. Geis. Hemoglobin: structure, function,
evolution, and pathology. The Benjamin/Cummings Publishing
Co., Menlo Park, Calif. 1983. p. 1.
3. A. Arnone, R.E. Benesch, and R. Benesch. J. Mol. Biol. 115,
627 (1977).
4. J.P. Collman, J.I. Brauman, J.P. Fitzgerald, P.D. Hampton, Y.
Naruta, J.W. Sparapany, and J.A. Ibers. J. Am. Chem. Soc.
110, 3477 (1988).
5. S. Miura and C. Ho. Biochemistry, 23, 2492 (1984).
6. R. Kluger, R.T. Jones, and D.T. Shih. Artif. Cells Blood Sub-
stitutes Immobilization Biotechnol. 22, 415 (1994).
7. M.A. Schumacher, M.M. Dixon, R. Kluger, R.T. Jones, and
R.G. Brennan. Nature (London), 375, 84 (1995).
8. M.A. Schumacher, E.E. Zheleznova, K.S. Poundstone, R. Kluger,
R.T. Jones, and R.G. Brennan. Proc. Natl. Acad. Sci. U.S.A.
94, 7841 (1997).
31. J.B. Fenn, M. Mann, C.K. Meng, S.F. Wong, and C.M.
Whitehouse. Science (Washington, D.C.), 246, 64 (1989).
32. R.T. Jones, C.G. Head, T.S. Fujita, D.T. Shih, J. Wodzinska,
and R. Kluger. Biochemistry, 32, 215 (1993).
9. R.M. Winslow. Hemoglobin-based red cell substitutes. Johns
Hopkins University Press, Baltimore, Md. 1995.
10. M. Jing, F.G. Panico, J.L. Panico, M.A. Ledvina, S. Bina, and
S.M. Muldoon. Artif. Cells Blood Substitutes Immobilization
Biotechnol. 24, 621 (1996).
11. S.M. Cohn and T.J. Farrell. J. Trauma, 39, 210 (1995).
12. K. Kikugawa, K. Adachi, H. Kosugi, and T. Asakura. Hemo-
globin, 7, 533 (1983).
13. I.M. Klotz, D.N. Haney, and L.E. Wood. J. Biol. Chem. 260,
16215 (1985).
14. R.T. Jones, D.T. Shih, T.S. Fujita, Y. Song, H. Xiao, C. Head,
and R. Kluger. J. Biol. Chem. 271, 675 (1996).
33. R. Kluger, L. Shen, H. Xiao, and R.T. Jones. J. Am. Chem.
Soc. 118, 8782 (1996).
34. M.P. Kavanaugh, D.T. Shih, and R.T. Jones. Biochemistry, 27,
1804 (1988).
35. H. Pötzschke, W.K.R. Barnikol, R.G. Kirste, and M. Rosenbaum.
Macromol. Chem. Phys. 197, 1419 (1996).
36. R. Benesch and R.E. Benesch. Biochem. Biophys. Res. Commun.
26, 162 (1967).
37. G. Fermi and M.F. Perutz. J. Mol. Biol. 175, 159 (1984).
38. J.M. Manning, L.R. Manning, S. Morgan, R.C. Beavis, and B.
Chait. Proc. Natl. Acad. Sci. U.S.A. 88, 3329 (1991).
© 1999 NRC Canada