Core–shell protein clusters comprising haemoglobin and recombinant feline serum albumin as an artificial O2 carrier for cats

Kyoko Yokomaku a, Motofusa Akiyama a, Yoshitsugu Morita a, Kiyohito Kihira b and Teruyuki Komatsu *a
aDepartment of Applied Chemistry, Faculty of Science and Engineering, Chuo University, 1-13-27 Kasuga, Bunkyo-ku, Tokyo 112-8551, Japan. E-mail:
bJEM Utilization Centre, Human Spaceflight Technology Directorate, Japan Aerospace Exploration Agency (JAXA), 2-1-1 Sengen Tsukuba-shi, Ibaraki 305-8505, Japan

Received 24th January 2018 , Accepted 13th March 2018

First published on 20th March 2018

This report describes the synthesis and structure of core–shell protein clusters comprising haemoglobin (Hb) at the centre and recombinant feline serum albumin (rFSA) at the exterior, named as haemoglobin–albumin clusters (Hb–rFSA3). Specifically, we highlight their capability as an artificial O2 carrier that can be used as a red blood cell (RBC) substitute for cats, the most populous pet animal in the world. First, rFSA was expressed by genetic engineering using Pichia yeast. The proteins show identical features to the native FSA derived from feline plasma. Single crystals of rFSA were prepared under a microgravity environment on the international space station (ISS), from which the structure was first revealed at 3.4 Å resolution. Subsequently, bovine Hb was wrapped covalently by rFSA using an α-succinimidyl-ε-maleimide crosslinker, yielding Hb–rFSA3 clusters. Three rFSA entities enfolded the Hb nuclei satisfactorily, giving the protein clusters a negative surface net charge (pI = 4.7) and preventing an immunological response against anti-Hb antibodies. The O2 affinity was higher (P50 = 9 Torr) than that of the native Hb. The Hb–rFSA3 clusters are anticipated for use as an alternative material for RBC transfusion, and as an O2 therapeutic reagent that can be exploited in various veterinary medicine scenarios.


“Are you a cat person or a dog person?” It is said that there are people of two types in the world. Based on official estimates, domestic cats outnumber dogs in the U.S., Canada, most of Western Europe, Russia, and Japan.1 Cats are the most populous pet animals worldwide (625 million cats, 425 million dogs).2 As the pet population increases, the demand for pet medical care continues to grow. Thereby, the expected frequency of blood transfusions is bound to increase. Among blood components, red blood cells (RBCs) are the most important to save lives. Nevertheless, blood banks for pet animals have not been approved. Therefore, veterinarians themselves must obtain blood for transfusion therapy. Not surprisingly, cats possess different blood types: type A (73–99%), type B (<33%), and type AB (the rarest).3 Mismatched blood transfusions cause mortal anaphylactic and haemolytic reactions because cats have naturally occurring alloantibodies.4 If an RBC substitute for cats were realized, veterinarians would not need to find donors, much less matched donors.

Development of RBC substitutes for human beings has continued for the last few decades.5–8 The promising compounds are Hb-based O2 carriers,7,8 for example, poly(ethylene glycol)-conjugated Hb9–11 and polymerized Hb.12,13 These artificial O2 carriers are designed with increased molecular size to avoid renal excretion. The transfusion of such heterogeneous molecules, however, often engenders a mild increase in blood pressure, which is inferred to be attributable to (i) nitric oxide (NO) scavenging by Hb diffused into the extravascular space, and (ii) O2 oversupply to vascular walls by diffusion.6,14–17 For that reason, no product has satisfied all requirements for practical use.18–20 Polymerized bovine Hb (Oxyglobin®) was commercially available in the U.S. and U.K. for the treatment of anaemic dogs.21 Alayash et al. reported that this formulation contains unpolymerized Hb at an amount of 37.8%.22 Consequently, some adverse effects have been reported, such as vomiting, anorexia, and pyrexia, as well as discoloration of the skin, mucous membranes, and urine.23

We earlier reported core–shell structured dual-protein clusters of Hb and HSA, named haemoglobin–albumin clusters (Hb–HSA3), as an artificial O2 carrier for human beings.24–28 In the mammalian bloodstream, serum albumin is the major plasma protein component (typical concentration of 2–5 g dL−1), which plays two crucial roles: (i) transporting a diverse range of metabolites and drugs and (ii) maintaining colloid osmotic pressure.29,30 Because the Hb nucleus is wrapped covalently by three HSAs, the Hb–HSA3 clusters do not induce hypertensive action.27 Moreover, they have a long lifetime in the intravascular space. These benefits are probably attributable to electrostatic repulsion between the negatively charged surface of the clusters and the basement membrane around the endothelial cells. To develop this product for use with domestic cats, the Hb must be covered with feline serum albumin (FSA) to escape unfavourable immunological reactions. However, it is difficult to acquire feline plasma (blood) to obtain FSA in adequate amounts. Our first challenge was the preparation of recombinant feline serum albumin (rFSA) using a genetic engineering procedure. To our surprise, little information exists in relation to FSA, which rather contrasts the fact that canine serum albumin has been widely studied.31–33 If it were possible to produce rFSA, then Hb–rFSA3 clusters could be manufactured in sufficient quantities. Of course, rFSA itself is expected to be beneficial as a plasma expander for several pharmaceutical situations. We have secreted rFSA using Pichia pastoris, which is known as the highest expression host cell, and have defined the physicochemical properties of the protein. Single crystals of rFSA were prepared aboard the International Space Station (ISS). X-ray crystallography revealed the three-dimensional (3D) structure of the protein for the first time. Using this pure rFSA, we synthesized Hb–rFSA3 clusters, which are anticipated as a promising RBC substitute as well as an O2 therapeutic reagent for use in numerous veterinary medicine scenarios.

Results and discussion

Expression and physicochemical properties of rFSA

We first expressed rFSA using Pichia yeast (Pichia pastoris GS115), which secretes the protein into the culture medium in gram quantities. The accumulated rFSA was collected by centrifugation and was purified using Cibacron blue-based affinity chromatography, followed by anion exchange chromatography. The rFSA showed identical features to native FSA derived from feline plasma. SDS-PAGE of rFSA depicted a single band with the same mobility as that of FSA (Fig. 1A inset). The elution curves of rFSA and FSA in size-exclusion chromatography (SEC) demonstrated single and sharp peaks with equivalent retention times (Fig. 1A). The protein purity was found to be 99.99%. The MALDI-TOF mass spectrum of rFSA exhibited a molecular-related ion peak at 65748.19 Da, which was almost identical to the observed mass of native FSA (65737.20 Da) and the simulated mass of the amino acid sequence (65843.48 Da) (the experimental range <0.16%) (Fig. S1, ESI). FSA comprises a single polypeptide with 584 amino acids. The pairwise sequence homology of FSA relative to HSA is 81.9% (Table S1, ESI).34
image file: c8tb00211h-f1.tif
Fig. 1 (A) SEC chromatograms of rFSA and FSA (inset: SDS-PAGE of these proteins). (B) CD spectra of rFSA and FSA in PBS solution (pH 7.4) at 25 °C.

The CD spectrum of rFSA in phosphate-buffered saline (PBS) solution (pH 7.4) coincided with that observed for FSA (Fig. 1B), suggesting that their secondary structures are equivalent. The isoelectric focusing (IEF) pattern demonstrated one band for both rFSA and FSA with a pI value of 4.5 (vide infra), which was somewhat lower than that of HSA (pI = 5.0).24,26 FSA possesses more glutamic acid residues (69 per molecule) and fewer lysine residues (52 per molecule) than HSA (62 glutamic acid and 59 lysine).34 As a consequence, the total charge of rFSA becomes rather negative. The UV-visible absorption spectrum of rFSA in PBS showed a broad band at 280 nm (data not shown). The molar absorption coefficient (ε280) of rFSA was ascertained as 4.40 × 104 M−1 cm−1, which is identical to the value of FSA. The high ε280(FSA) compared to ε280(HSA) (3.50 × 104 M−1 cm−1)29 might be derived from the existence of many tyrosine residues (20/molecule in FSA, 18/molecule in HSA). All of the physicochemical observations proved that rFSA and native FSA are fundamentally the same in terms of their general features.

Crystal structure of rFSA

Several researchers have reported the structures of HSA in various forms (defatted and liganded).35–39 The crystallographic co-ordinates of bovine, equine, leporine, and canine serum albumins are included in the Protein Data Bank (PDB).33,40 In contrast, the 3D structure of FSA has never been published in the relevant literature. Since rFSA is a very soluble protein, it was difficult to produce a single crystal with good quality. We therefore conducted the crystallization of rFSA under the microgravity environment in the Japanese Experimental Module “Kibo” of the ISS. The crystals of rFSA belong to the monoclinic space group C 1 2 1 (Table S2, ESI). The protein comprises a single polypeptide containing three homologous domains I, II, and III (I, Glu-1–Glu-196; II, Arg-197–Glu-383; III, Pro-384–Ala-584), each of which is divided into two subdomains A and B (Fig. 2). Despite the symmetrically repeating conformations of domains I–III, rFSA forms a unique heart-shaped structure due to the asymmetric spatial alignment of the domains. There are 35 cysteine residues, in which 34 sulfhydryl groups participate in 17 disulphide bonds, with the exception of Cys-34. The overall structure of rFSA resembles that of rHSA and other animal albumins.33,35–40
image file: c8tb00211h-f2.tif
Fig. 2 Crystal structure of rFSA. The rFSA comprises three morphologically equivalent domains I (red region), II (green region) and III (blue region), each of which is divided into two subdomains (A and B). Cys-34 in subdomain IA is the only Cys residue with a sulfhydryl group.

It is noteworthy that the superposition image of rFSA and defatted rHSA demonstrates clearly that the configurations of the three subdomains are identical (Fig. S2, ESI). This result reflects that rFSA was in defatted form, although our purification protocol did not include a fatty acid removal process, such as a powdered activated-carbon treatment under an acidic solution. We used anion exchange chromatography at the last step of the purification. Anionic ligands bound to the protein are likely to be eliminated during column separation, thus yielding a defatted rFSA. As described above, rFSA has more glutamic acid residues and fewer lysine residues than HSA has. The acidic property of rFSA was apparent by its electrostatic potential representation (Fig. S3, ESI).

In terms of its general structure, the environment around Cys-34 of rFSA resembles that of rHSA. Careful inspection revealed, however, three apparent differences in the amino acid sequence in a flexible loop surrounding the S atom of Cys-34 (Val-77–Tyr-84, sequence VA[S with combining low line]LR[D with combining low line][K with combining low line]Y) compared to rHSA (VA[T with combining low line]LR[E with combining low line][T with combining low line]Y). We inferred that the molecular space around Cys-34 in rFSA with shorter serine, aspartic acid, and positively charged lysine, is more polar than that of rHSA. This finding might be related closely with the observations that the mercapto-ratios of Cys-34 in rFSA (approximately 30–50%) were lower than the value of rHSA (approximately 50–60%).

Moreover, we performed molecular dynamics (MD) simulations of rFSA in an aqueous medium at 300 K. The root mean square deviation (RSMD), which designates the conformational stability, reached a plateau within 4 ns in both the rFSA and rHSA structures (Fig. S4, ESI), which indicates that the protein structures retain their conformations and stabilities during the simulations. Root mean square fluctuation (RMSF) is an index of the degree of fluctuation in the molecule. As anticipated, the turn (loop) regions at the corner (Ser-58, Ala-60, Ala-561–Ala-566; RMSF > 3 Å) showed high fluctuation (red-coloured moieties) (Fig. 3A). The RMSF values with respect to the amino acid residue number completely overlapped with those of rHSA (Fig. 3B). In the course of the MD simulations, it became apparent that rFSA has similar dynamic properties to rHSA at ambient temperature.

image file: c8tb00211h-f3.tif
Fig. 3 Molecular dynamics simulations of rFSA and rHSA (PDB ID: 1AO6).37 (A) Variations in root mean square fluctuation (RMSF) at Cα for 10 ns MD trajectories represented as the width of the backbone trace of rFSA and rHSA. (B) Comparison of the RMSF values of the Cα atoms for rFSA and rHSA.

Synthesis of Hb–rFSA3 clusters

To wrap Hb with rFSAs, we used α-succinimidyl-ε-maleimide (SMCC) as a coupling reagent (Fig. 4). Initially, the crosslinker was reacted with carbonyl Hb to prepare maleimide-activated Hb (MA-Hb). The free sulfhydryl group of Cys-34 in rFSA is responsible for covalent binding to the surface maleimide groups on Hb. Site-specific binding of Cys-34 of rFSA to the MA-Hb yielded core–shell protein clusters with Hb at the centre and rFSAs at the exterior. Because bovine Hb also possesses a free sulfhydryl group at Cys-93(β),41 the MA-Hbs are likely to produce Hb polymers. Such intermolecular linkage, however, did not occur because of (i) the short length of the crosslinker arm (10 Å length), and (ii) the masking of Cys-93(β) in Hb by the maleimide-terminus of the excess SMCC. The SMCC is an appropriate coupling reagent to create structurally defined core–shell Hb–rFSA3 clusters.
image file: c8tb00211h-f4.tif
Fig. 4 Schematic illustration of the synthetic route of the Hb–rFSA3 clusters. Cys-34 of rFSA and the surface Lys group of Hb were coupled with a crosslinker (SMCC). The product is designated as Hb–rFSA3, in which the central Hb is covered with an average of three rFSA molecules.

The SEC chromatogram of the reaction mixture demonstrated three new peaks in the high molecular weight region (Fig. 5A, elution time 13–17 min). Native-PAGE also exhibited three bands above rFSA (Fig. S5A, ESI). All high-molecular-weight components were collected using gel filtration chromatography (GFC) (Fig. 5A). Based on the total protein assay and Hb assay, the average rFSA/Hb ratio of the product was determined as 3.0 ± 0.2 (72% yield). The majority of the clusters are Hb–rFSA3 heterotetramers, but Hb–rFSA2 trimers and Hb–rFSA4 pentamers are also included. The mixture of these components is designated as Hb–rFSA3 with an italicized subscript 3.

image file: c8tb00211h-f5.tif
Fig. 5 (A) SEC chromatograms of the reaction mixture (maleimide-activated Hb + rFSA) and isolated Hb–rFSA3 clusters. (B) CD spectra of the Hb–rFSA3 clusters (0.2 μM), Hb (0.2 μM), and rFSA (0.2 μM) in PBS solution (pH 7.4) at 25 °C.

The CD spectrum of the Hb–rFSA3 clusters coincided perfectly with the sum of the Hb spectrum and a three-fold enlarged rFSA spectrum (Fig. 5B). This result proved that (i) the secondary structure of the individual protein entities was unaltered by the protein coupling and (ii) the average rFSA/Hb composition ratio is 3/1 (mol/mol). The dynamic light scattering (DLS) measurements showed that the hydrodynamic diameter of the Hb–rFSA3 clusters (13.3 nm) was significantly larger than that of Hb (6.4 nm). The isoelectric point of the Hb–rFSA3 clusters (pI: 4.7) was close to the value of rFSA (pI: 4.5) (Fig. S5B, ESI), which supports the wrapping of Hb by rFSAs. Remarkably, the molecular characteristics of the Hb–FSA3 clusters, which were synthesized using native FSA, are exactly the same as those of the Hb–rFSA3 clusters. We inferred that the Hb–rFSA3 clusters with negative surface net charges and large molecular size can circulate in the bloodstream for a long time and that they would not elicit an undesirable vasopressor response through NO depletion.


An immuno-chromatographic quick occult blood test kit enabled us to visualize the immunological reactivity of the Hb–rFSA3 clusters against anti-Hb antibodies. For this experiment, clusters bearing a human HbA core have been made: HbA–rFSA3 clusters. The HbA-antibody immune complexes migrated to the test area and bound to the rabbit anti-HbA antibody, forming sandwich complexes, which showed a red test-line signal (Fig. S6A, ESI). In contrast, the red line was undetectable in the cases of rFSA and HbA–rFSA3 clusters. These results suggest that the central HbA in the Hb–rFSA3 clusters is covered satisfactorily with rFSAs, and that the anti-HbA antibody cannot recognize the epitope of the inner HbA.

Furthermore, the immunological reactivity of the Hb–rFSA3 clusters against anti-FSA antibody was investigated using a latex agglutination immunoassay. The absorption intensity of the reactant with the Hb–rFSA3 clusters exhibited linear concentration dependence, which coincided with the calibration line of the FSA standard (Fig. S6B, ESI). The antigen epitopes of the rFSA exterior are preserved even after coupling with Hb. We have inferred that the immunochemical response of the Hb–rFSA3 clusters was quite low against anti-Hb antibodies, but that it was sufficiently high against anti-FSA antibodies.

O2 affinity and stability of the oxy form

The visible absorption spectra of the Hb–rFSA3 clusters in PBS solution (pH 7.4) equilibrated with O2, N2, and CO gases (oxy, deoxy, and carbonyl forms) were fundamentally the same as those observed for Hb–HSA3, Hb–FSA3, and native Hb (Fig. 6A).25,42 The same absorption maxima of these clusters demonstrated that the wrapping of Hb with albumins did not affect the electronic states of the haem groups (Table S3, ESI). The P50 value of the Hb–rFSA3 clusters was found to be 9 Torr (37 °C), whereas native Hb showed a P50 of 23 Torr (Fig. 6B and Table 1).24,25 The Hill coefficient (n value) also decreased from 2.6 to 1.5. The covering of Hb with rFSA reduced the P50, thereby increasing the O2 affinity by 2.6-fold compared to native Hb. Equivalent P50 and n reductions were also observed in the Hb–FSA3 clusters (P50 = 9 Torr) (Table 1). The high O2 affinity and low cooperativity of the Hb–rFSA3 clusters were explained by the two structural variations of the Hb nuclei: (i) the capping of the sulfhydryl group of Cys-93(β), and (ii) the chemical modification of the surface amino groups of Lys residues by SMCC. Modification of Cys-93(β) is known to raise the O2 affinity10,11,43 and to cause a marked decrease in the possible motion of the α/β subunit interface of Hb.43 Actually, the number of cysteinyl thiol groups of MA-Hb was only 0.2 per molecule, signifying that two Cys-93(β) were masked by the maleimide of SMCC, which did not participate in the crosslinking between Hb and rFSA. Modification of surface lysine residues by the SMCC's succinimide is necessary for cluster preparation, but it generally provokes measurable changes in the O2 affinity and Hill coefficient.9,44,45 In particular, Lys-82(β) plays a key role in the quaternary structure transition from the R (relaxed) state to a T (tense) state. We recently found that Lys-82(β) is a predominant binding partner of Cys-34 of HSA based on the 3D reconstruction of a Hb–HSA3 cluster.28
image file: c8tb00211h-f6.tif
Fig. 6 (A) UV-vis absorption spectra of Hb–rFSA3 clusters in oxy, deoxy, and carbonyl forms in PBS solution (pH 7.4) at 25 °C. (B) O2 dissociation curves of Hb–rFSA3 and Hb–FSA3 clusters in PBS solution (pH 7.4) at 37 °C.
Table 1 O2 binding parameters of Hb–rFSA3 and Hb–FSA3 clusters in PBS solution (pH 7.4) at 37 °C
Haemoproteins P 50 (Torr) n (−) k ox (h−1)
a Ref. 25.
Hba 23 2.6 0.037
Hb–rFSA3 9 1.5 0.041
Hb–FSA3 9 1.5 0.039
Hb–HSA3a 9 1.5 0.035

The high O2 affinity might be favourable as an RBC substitute. Winslow et al. demonstrated that a Hb-based O2 carrier with a low O2 affinity dissociates excess O2 in the arterioles, inducing autoregulatory vasoconstriction.14,46 Intaglietta et al. reported that RBCs with a lower P50 (high O2 affinity) improve microvascular function in a haemorrhagic-shocked hamster model.47 The increase of O2 affinity might be beneficial to reduce arteriole O2 transport, and to prevent cardiovascular effects.

The autoxidation rate constant (kox) of the oxygenated Hb core in the Hb–rFSA3 clusters was measured in PBS solution (pH 7.4) at 37 °C. The kox value of the Hb–rFSA3 clusters was ascertained as 0.041 h−1, which was similar to the data of Hb–FSA3, Hb–HSA3, and native Hb (Table 1).25 The oxy form of the internal Hb maintains good stability even after coverage with rFSA.


The Hb–rFSA3 clusters shown in this work were designed as an artificial O2 carrier for exclusive use with cats. rFSA has been produced in a gram-scale quantity using Pichia yeast expression. The protein exhibited identical properties to those observed for native FSA. We first ascertained the crystal structure of rFSA, which revealed that the protein morphology closely approximates that of rHSA. The covalent coupling of Hb with rFSA yielded the core–shell structured Hb–rFSA3 clusters. The Hb centre becomes a stealth protein covered with three rFSA molecules on the exterior. Most probably, injection of the Hb–rFSA3 clusters into a cat would not generate Hb antibodies in the blood. The O2 affinity of this cluster became higher (P50 = 9 Torr) than that of the naked Hb, mainly because of the interference of the quaternary structure plasticity by the modification of surface lysine residues. Overall, the Hb–rFSA3 clusters possess three important beneficial features for use as an RBC substitute: (1) a negatively charged molecular surface, (2) equivalent immunogenicity with FSA, and (3) moderately high O2 affinity. These characteristics are favourable for long-term circulation in the bloodstream as an artificial O2 carrier, for maintaining immunological safety in vivo, and for avoiding a vasopressor response. Both the rFSA and the Hb–rFSA3 clusters will be of great medical importance as a potential plasma expander and O2-delivering fluid for veterinary blood transfusion in many clinical situations.


Materials and apparatus

Native FSA was purified from feline plasma (ESI). Bovine Hb was obtained from bovine blood purchased from Tokyo Shibaura Zouki Co. Ltd. All other chemical reagents were purchased from commercial sources as special grades and were used without additional purification. Deionized water (18.2 MΩ cm) was prepared using water purification systems (Elix UV and Milli Q Reference; Millipore Corp.). SDS-PAGE and Native-PAGE were conducted using a 5–20% polyacrylamide precast gradient gel (SuperSep; Wako Pure Chemical Industries Ltd). Isoelectric focusing (IEF) was performed using IEF protein gels (Novex pH 3–10; Thermo Fisher Scientific K.K.). The UV-visible absorption spectra were obtained using a UV-visible spectrophotometer (8543; Agilent Technologies Inc.) equipped with a Peltier temperature controller (89090A; Agilent Technologies Inc.). Circular dichroism (CD) spectra were recorded using a circular dichroism spectrometer (J-820; Jasco Corp.). Mass spectra were measured using a MALDI-TOF mass spectrometer (Autoflex equipped with a pulsed N2 laser (337 nm); Bruker Daltonics K.K.). As a matrix, 50% acetonitrile solution including 0.1% sinapic acid and 0.05% trifluoroacetic acid was used. The molecular size of the clusters was determined by dynamic light scattering (DLS) measurement using a zeta-potential and particle size analyser (ELSZ-2000; Otsuka Electronics Co., Ltd).

Expression and purification of rFSA

Complementary (c)DNA was prepared via reverse transcription using feline liver RNA. The full-length cDNA of FSA was amplified by PCR using the following oligonucleotide primer set: forward [5′-TCGAAACGAGGAATTCGGCACAATGAAGTGGGTAACC-3′] and reverse [5′-TGTCTAAGGCGAATTCTTAGGCTAAGGCAGCTTGAGC-3′]. The amplified fragment was cloned into the EcoRI site of the pHIL-D2 plasmid. The obtained pHIL-D2-FSA vector was linearized by SalI and was used for transformation of the GS115 strain of Pichia pastoris (Thermo Fisher Scientific K.K.) using electroporation.

Expression and purification of rFSA were conducted by our protocols as reported previously.33 Transformed clone cells were grown in a buffered mineral glycerol-complex (BMGY) medium (total 4 L), and subsequently in a buffered mineral methanol complex (BMMY) medium (total 0.8 L) in a shaking incubator (Bio-Shaker G·BR-200; Taitec Corp.) (200 rpm, 30 °C). During cultivation, methanol (1% volume of medium) was added every 24 h. The amount of rFSA reached 724 mg/0.8 L media after 10 days.

The growth medium was centrifuged, and the supernatant was brought to 50% saturation with ammonium sulphate [(NH4)2SO4]. After the turbid solution was centrifuged, the supernatant was brought to 95% saturation with (NH4)2SO4. The precipitated protein was collected by centrifugation and was dialyzed against deionized water at 4 °C with subsequent addition of 11% volume of 500 mM sodium phosphate (pH 5.8). The resultant protein solution in 50 mM sodium phosphate (pH 6.0) was subjected to affinity chromatography (Toyopearl AF-Blue F3G-A; Tosoh Corp.). After washing with 50 mM sodium phosphate (pH 6.0), rFSA was eluted with 50 mM sodium phosphate (pH 7.4) including 3 M NaCl. Then the eluent was dialyzed against deionized water at 4 °C, and 25% volume of 100 mM Tris–HCl (pH 8.0) was added. Next, the resulting 20 mM Tris–HCl solution (pH 8.0) of the protein was subjected to anion exchange chromatography (Hitrap Q FF; GE Healthcare UK Ltd) using 20 mM Tris–HCl (pH 8.0) as the running buffer. After washing adequately with 20 mM Tris–HCl (pH 8.0) containing 100 mM NaCl, the rFSA was eluted with 20 mM Tris–HCl (pH 8.0) containing 200 mM NaCl. The eluent was dialyzed against deionized water at 4 °C with subsequent addition of 11% volume of 10× phosphate-buffered saline (PBS, pH 7.4). Finally, the rFSA solution was concentrated to 30 mL and was sterilized with a sterile membrane filter (DISMIC-25CS, 0.2 μm pore; Toyo Roshi Kaisha Ltd). All purification processes were checked by SDS-PAGE or Native-PAGE analysis. The concentration of rFSA was measured using a protein assay kit (Pierce 660 nm; Thermo Fisher Scientific K.K.). The sulfhydryl group assay of rFSA was conducted by reaction with 4,4′-dithiopyridine (4,4′-DTP).48

Crystallization, data collection, and structure determination of rFSA

The crystallization of rFSA was achieved in the Japanese Experiment Module “Kibo” of the ISS. The solution samples were launched aboard the Soyuz MS-02 spacecraft (48S) on 19 October 2016 from the Baikonur Cosmodrome (Kazakhstan) for the ISS. Crystal growth in space was started in the Protein Crystallization Research Facility (PCRF) in Kibo from 21 October 2016. Colourless crystals (0.15 × 0.05 × 0.05 mm) grew from a 47 mg mL−1 protein solution in 10 mM sodium phosphate (pH 7.4), 35% PEG3350, and 0.2 M sodium malonate (pH 4.8) at 293 K within 10 days. After the grown crystals were returned to Earth aboard the Soyuz MS-01 spacecraft (47S) on 30 October 2016, they were flash-frozen in liquid nitrogen. X-ray diffraction data were collected using a Dectris Pilatus 6M detector with a synchrotron radiation source at the SPring-8 beamline BL41XU (Hyogo, Japan). All data were collected at 100 K and were processed with XDS (Table S2, ESI).49 The rFSA structure was ascertained using molecular replacement with MOLREP in PHENIX.50 The rHSA structure (PDB ID: 3LU8) was used as the search model.51 Sequence assignment of the atomic model was conducted using the amino acid sequence of FSA.34 Further iterative model building and refinement were performed using PHENIX and COOT.52 Electron densities corresponding to residues 1–4, 88–91, and 101–102 (all in subdomain IA) were disordered in the model because of conformational flexibility at the N-terminal and loops. The atomic co-ordinates of the rFSA structure reported in this study were deposited in the Protein Data Bank (PDB) under accession code 5YXE. All molecular images were produced using PyMOL (Schrödinger LLC).

Molecular dynamics (MD) simulations of rFSA

MD simulations were performed using the Desmond application as implemented in the Maestro graphical interface.53,54 We used the optimized potentials for liquid simulations (OPLS_2005) force field implemented in the Desmond software for all molecules of the system.55 The crystal structures of rFSA obtained in this work (PDB ID: 5YXE) and rHSA (PDB ID: 1AO6)37 were used for the simulations. To neutralize the charges of the rFSA and rHSA models, seventeen and twelve sodium counter ions were added, respectively. The proteins were solvated with explicit TIP3P water56 with a buffering distance of 20 Å in an octahedral box. The systems were relaxed using a default relaxation protocol as implemented in the Desmond software. The NPT MD simulations were performed for 10 ns to maintain the 300 K temperature and the 1.01325 bar pressure. The time step was 2 fs, and information for analysis was printed every 10 ps. A 10 Å cut off was used for non-bonded interactions.

Preparation of Hb–rFSA3 clusters

A DMSO solution of succinimidyl-4-(N-maleimidomethyl)-cyclohexane-1-carboxylate (SMCC; Wako Pure Chemical Industries Ltd) (20 mM, 0.2 mL) was added dropwise into PBS solution (pH 7.4) of carbonyl Hb (0.1 mM, 2 mL). After stirring the mixture for 30 min in a CO atmosphere at 4 °C, excess SMCC was removed using gel filtration chromatography (GFC) (Sephadex G25 superfine; GE Healthcare UK Ltd). The obtained SMCC-bound Hb (maleimide-activated Hb, MA-Hb) was concentrated to 2 mL (0.1 mM) using a centrifugal concentrator (Vivaspin 20 ultrafilter, 10 kDa MWCO; GE Healthcare UK Ltd).

Aqueous dithiothreitol (DTT) solution (20 mM, 25 μL) was added to the PBS solution (pH 7.4) of rFSA (1 mM, 1 mL) ([DTT]/[rFSA] = 0.5 mol/mol) to reduce the partially oxidized Cys-34 of rFSA. The mixture was incubated for 1 h at 25 °C. The unreacted DTT was removed using GFC (Sephadex G25 superfine), and the eluent was condensed to 1 mL ([rFSA]= 1 mM) using a Vivaspin 20 ultrafilter (10 kDa MWCO). The mercapto-ratio of the Cys-34 was confirmed as 80–90% using the 4,4′-DTP procedure.48

Next the MA-Hb solution (0.1 mM, 1 mL) was added dropwise into the Cys-34 reduced rFSA solution (1 mM, 1 mL) in PBS. The mixture was stirred continuously in a CO atmosphere under dark conditions for 16 h at 4 °C. In all of the synthesis processes, the Hb was kept as a carbonyl complex under CO atmosphere to avoid unfavourable autoxidation of the prosthetic haem group. The preparation of the clusters under N2 atmosphere caused a reduction of the O2 affinity.26 An aliquot of the reactant was subjected to size-exclusion chromatography (SEC) on an HPLC system (Prominence LC-20AD/CTO-20A/SPD-20A; Shimadzu Corp.) with an SEC column (YMC-Pack Diol-300 S-5; YMC Co. Ltd) with 50 mM sodium phosphate (pH 7.4) as the mobile phase. The chromatogram showed new peaks at the high molecular weight region. The resultant mixture was subjected to GFC (Superdex 200 pg; GE Healthcare UK Ltd) using PBS as the running buffer, and all the fractions including clusters were collected. The concentrations of total protein and Hb were measured, respectively, using a protein assay (Pierce 660 nm Protein Assay; Thermo Fisher Scientific K.K.) and a general cyano-metHb assay. The average rFSA/Hb ratio of the product was determined as 3.0 ± 0.2. We designated these clusters as Hb–rFSA3, with an italicized subscript 3. The Hb–FSA3 clusters were prepared using the same procedure with native FSA.


To assess the reaction between the HbA–rFSA3 clusters and anti-HbA antibodies, a Quick Chaser Occult Blood Test kit (Mizuho Medy Co., Ltd), which detects human Hb (HbA) using an immuno-chromatographic method, was applied. The HbA–rFSA3 clusters were prepared using the same protocol as Hb–rFSA3 using human HbA.26 After the HbA–rFSA3 solution (0.1–1 μM) in PBS (pH 7.4) was dropped onto the assay plate, the appearance of the red test-line signal was evaluated.

The immunological response of the HbA–rFSA3 clusters against anti-FSA antibodies was evaluated using latex immuno-agglutination reagents (Feline albumin in urea; Shima Laboratories Co., Ltd). The Hb–rFSA3 cluster solution ([rFSA unit] = 27–300 μg mL−1, 30 μL) in PBS (pH 7.4) was diluted with saline (120 μL). An aliquot (2.25 μL) of the solution was diluted further with special buffer (60 μL). The obtained sample solution was mixed with an anti-FSA antibody-sensitized latex bead solution (13 μL) at 37 °C. The absorbance of the reactant was recorded at 694 nm using a clinical chemistry analyser (BioMajesty JCA-BM2250; JEOL Ltd).

O2 binding property and oxy form stability

The oxy form (O2 complex) of the Hb–rFSA3 cluster solution (PBS, pH 7.4, approximately 10 μM, 3 mL) was prepared using our previously reported technique33 and was transferred to an optical quartz cuvette (10 mm path length) with a rubber septum cap. The N2 gas was blown into the oxy form solution, yielding deoxy Hb–rFSA3 clusters. The UV-visible absorption spectra of these species were recorded at 25 °C. The O2 affinity (P50: O2 partial pressure where Hb is half-saturated with O2) and Hill coefficient (n) were determined using an automatic recording system for the O2 equilibrium curve (Hemox Analyzer; TCS Scientific Corp.) at 37 °C. The clusters in PBS solution (pH 7.4) were deoxygenated by flushing with N2 and were oxygenated by increasing the O2 partial pressure. The oxy form stability of the Hb–rFSA3 clusters was assessed using the first-order autoxidation rate constant (kox) of the core Hb using our earlier described procedures.25,33

Conflicts of interest

There are no conflicts to declare.


The authors acknowledge Dr Natsuhiko Sugimura and Dr Toshimichi Shibue (Materials Characterization Central Laboratory, Waseda University) for mass spectroscopy experimentation, and Ms Hiromi Asakawa and Dr Masakatsu Hashimoto (Shima Laboratories Co. Ltd) for immunogenicity measurements. The crystal growth in the Japanese Experiment Module “Kibo” of the ISS was conducted by JAXA High-quality Protein Crystal Growth (PCG) Experiment (2016B). The authors thank astronaut Mr Takuya Ohnishi (JAXA) for conducting protein crystal growth experiments aboard the ISS. This work was supported by a Joint Research Grant from the Institute of Science and Engineering, Chuo University, and a Research Grant from Naito Foundation.

Notes and references

  1. The Washington Post, 28 July 2014,, date of access: 15/01/2018.
  2. Worldatlas, 25 April 2017,, date of access: 15/01/2018.
  3. B. Gandolfi, R. A. Grahn, N. A. Gustafson, D. Proverbio, E. Spada, B. Adhikari, J. Cheng, G. Andrews, L. A. Lyons and C. R. Helps, PLoS One, 2016, 11, e0154973 Search PubMed.
  4. U. Giger and J. Bücheler, J. Am. Vet. Med. Assoc., 1991, 198, 411–418 CAS.
  5. J. E. Squires, Science, 2002, 295, 1002–1005 CrossRef CAS PubMed.
  6. H. W. Kim and A. G. Greenburg, Artif. Organs, 2004, 28, 813–827 CrossRef CAS PubMed.
  7. J. H. Jahr, A. Sadighi, L. Doherty, A. Li and H. W. Kim, in Chemistry and Biochemistry of Oxygen Therapeutics: From Transfusion to Artificial Blood, ed. A. Mozzarelli and S. Bettati, John Wiley & Sons, West Sussex, 2011, pp. 301–316 Search PubMed.
  8. R. Kluger and F. E. Lui, in Hemoglobin-Based Oxygen Carriers as Red Cell Substitutes and Oxygen Therapeutics, ed. H. W. Kim and A. G. Greenburg, Springer-Verlag, Berlin, 2013, pp. 159–183 Search PubMed.
  9. K. D. Vandegriff, M. Malavalli, J. Wooldridge, W. Lohman and R. M. Winslow, Transfusion, 2003, 43, 509–516 CrossRef CAS PubMed.
  10. B. M. Manjula, A. Tsai, R. Upadhya, K. Perumalsamy, P. K. Smith, A. Malavalli, K. Vandegriff, R. M. Winslow, M. Intaglietta, M. Prabhakaran, J. M. Friedman and A. S. Acharya, Bioconjugate Chem., 2003, 14, 464–472 CrossRef CAS PubMed.
  11. D. Li, T. Hu, B. N. Manjula and S. A. Acharya, Bioconjugate Chem., 2009, 20, 2062–2070 CrossRef CAS PubMed.
  12. R. A. Boykins, P. W. Buehler, Y. Jia, R. Venable and A. I. Alayash, Proteins, 2005, 59, 840–855 CrossRef CAS PubMed.
  13. J. H. Levy, Expert Opin. Biol. Ther., 2003, 33, 509–517 CrossRef PubMed.
  14. R. M. Winslow, Biochim. Biophys. Acta, 2008, 1784, 1382–1386 CrossRef CAS PubMed.
  15. E. E. Moore, F. A. Moore, T. C. Fabian, A. C. Bernard, G. F. Fulda, D. B. Hoyt, T. M. Duane, L. J. Weireter Jr., G. A. Gomez, M. D. Cipolle, G. H. Rodman Jr., M. A. Malangoni, G. A. Hides, L. A. Omert and S. A. Gould, J. Am. Coll. Surg., 2009, 208, 1–13 CrossRef PubMed.
  16. D. H. Doherty, M. P. Doyle, S. R. Curry, R. J. Vali, T. J. Fattor, J. S. Olson and D. D. Lemon, Nat. Biotechnol., 1998, 16, 672–676 CrossRef CAS PubMed.
  17. B. Yu, K. D. Bloch and W. Zapol, Trends Cardiovasc. Med., 2009, 19, 103–107 CrossRef CAS PubMed.
  18. C. Natanson, S. J. Kern, P. Lurie, S. M. Banks and S. M. Wolfe, J. Am. Med. Assoc., 2008, 299, 2304–2312 CrossRef CAS PubMed.
  19. R. Kluger, Curr. Opin. Chem. Biol., 2010, 14, 538–543 CrossRef CAS PubMed.
  20. A. Alayash, Trends Biotechnol., 2014, 32, 177–185 CrossRef CAS PubMed.
  21. B. Driessen, J. H. Jahr, F. Lurie, S. M. Griffery and R. A. Gunther, Br. J. Anaesth., 2001, 86, 683–692 CrossRef CAS PubMed.
  22. P. W. Buehler, R. A. Boykins, Y. Jia, S. Norris, D. I. Freedberg and A. I. Alayash, Anal. Chem., 2005, 77, 3466–3478 CrossRef CAS PubMed.
  23. European Medicines Agency, Oxyglobin, Haemoglobin glutamer 200, March 2012,'oxyglobin+side+effects, date of access: 15/01/2018.
  24. D. Tomita, T. Kimura, H. Hosaka, Y. Daijima, R. Haruki, K. Ludwig, C. Böttcher and T. Komatsu, Biomacromolecules, 2013, 14, 1816–1825 CrossRef CAS PubMed.
  25. H. Hosaka, R. Haruki, K. Yamada, C. Böttcher and T. Komatsu, PLoS One, 2014, 9, e110541 Search PubMed.
  26. T. Kimura, R. Shinohara, C. Böttcher and T. Komatsu, J. Mater. Chem. B, 2015, 3, 6157–6164 RSC.
  27. R. Haruki, T. Kimura, H. Iwasaki, K. Yamada, I. Kamiyama, M. Kohno, K. Taguchi, S. Nagao, T. Maruyama, M. Otagiri and T. Komatsu, Sci. Rep., 2015, 9, 12778 CrossRef PubMed.
  28. R. Shinohara, T. Yamada, B. Schade, C. Böttcher and T. Komatsu, J. Phys. Chem. Lett., 2017, 8, 819–824 CrossRef CAS PubMed.
  29. T. Peters Jr., All about Albumin, Biochemistry, Genetics and Medical Applications, Academic Press, San Diego, 1996 Search PubMed.
  30. G. Fanali, A. di Masi, A. V. Trezza, M. Marino, M. Fasano and P. Ascenzi, Mol. Aspects Med., 2012, 33, 209–290 CrossRef CAS PubMed.
  31. T. Kosa, T. Maruyama and M. Otagiri, Pharm. Res., 1997, 14, 1607–1612 CrossRef CAS.
  32. B. Pandjaitan, I. Swoboda, F. Brandejsky-Pichler, H. Rumpold, R. Valenta and S. Spitzauer, J. Allergy Clin. Immunol., 2000, 105, 279–285 CrossRef CAS PubMed.
  33. K. Yamada, K. Yokomaku, M. Kureishi, M. Akiyama, K. Kihira and T. Komatsu, Sci. Rep., 2016, 6, 36782 CrossRef CAS PubMed.
  34. UniProt KB-P49064 (ALBU_FELCA) for FSA, and UniProt KB-P02768 (ALBU_HUMAN) for HSA.
  35. S. Curry, H. Mandelkow, P. Brick and N. Franks, Nat. Struct. Biol., 1997, 5, 827–835 CrossRef PubMed.
  36. A. A. Bhattacharya, S. Curry and N. P. Franks, J. Biol. Chem., 2000, 275, 38731–38738 CrossRef CAS PubMed.
  37. K. Sugio, A. Kashima, S. Mochizuki, M. Noda and K. Kobayashi, Protein Eng., 1999, 12, 439–446 CrossRef PubMed.
  38. P. A. Zunszain, J. Ghuman, T. Komatsu, E. Tsuchida and S. Curry, BMC Struct. Biol., 2003, 3, 6 CrossRef PubMed.
  39. J. Ghuman, P. A. Zunszain, I. Petitpas, A. A. Bhattacharya, M. Otagiri and S. Curry, J. Mol. Biol., 2005, 353, 38–52 CrossRef CAS PubMed.
  40. K. A. Majorek, P. J. Porebski, A. Dayal, M. D. Zimmerman, K. Jablonska, A. J. Stewart, M. Chruszcz and W. Minor, Mol. Immunol., 2012, 52, 174–182 CrossRef CAS PubMed.
  41. T. C. Mueser, P. H. Rogers and S. Arnone, Biochemistry, 2000, 39, 15353–15364 CrossRef CAS PubMed.
  42. E. Antonini and M. Brunori, in Hemoglobin and Myoglobin in Their Reactions with Ligands, ed. A. Neuberger and E. L. Tatum, North-Holland Pub., Amsterdam, 1971, pp. 13–39 Search PubMed.
  43. Y. Zhang, V. S. Bhatt, G. Sun, P. G. Wang and A. F. Palmer, Bioconjugate Chem., 2008, 19, 2221–2230 CrossRef CAS PubMed.
  44. R. Kluger and J. Zhang, J. Am. Chem. Soc., 2003, 125, 6070–6071 CrossRef CAS PubMed.
  45. D. Hu and R. Kluger, Biochemistry, 2008, 47, 12551–12561 CrossRef CAS PubMed.
  46. R. J. Rohlfs, E. Bruner, A. Chiu, A. Gonzales, M. L. Gonzales, D. Magde, M. D. Magde Jr., K. D. Vandegriff and R. M. Winslow, J. Biol. Chem., 1998, 273, 12128–12134 CrossRef CAS PubMed.
  47. M. Intaglietta, in Microcirculatory Effects of Hemoglobin Solutions, ed. K. Messmer, K. E. Burhop and J. Hutter, Karger AG, Basel, 2004, pp. 8–15 Search PubMed.
  48. D. R. Grassetti and J. F. Murry Jr., Arch. Biochem. Biophys., 1967, 119, 41–49 CrossRef CAS PubMed.
  49. W. Kabsch, Acta Crystallogr., Sect. D: Biol. Crystallogr., 2010, 66, 125–132 CrossRef CAS PubMed.
  50. A. Vagin and A. Teplyakov, J. Appl. Crystallogr., 1997, 30, 1022–1025 CrossRef CAS.
  51. D. Buttar, N. Colclough, S. Gerhardt, P. A. MacFaul, S. D. Phillips, A. Plowright, P. Whittamore, K. Tam, K. Maskos, S. Steinbacher and H. Steuber, Bioorg. Med. Chem., 2010, 18, 7486–7496 CrossRef CAS PubMed.
  52. P. Emsley, B. Lohkamp, W. G. Scott and K. Cowtan, Acta Crystallogr., Sect. D: Biol. Crystallogr., 2010, 66, 486–501 CrossRef CAS PubMed.
  53. Desmond Molecular Dynamics System, D. E. Shaw Research, Maestro-Desmond Interoperability Tools, Schrödinger LLC, New York, NY, 2016.
  54. Maestro. Schrödinger Release 2017-3, Schrödinger LLC, New York, NY, 2017.
  55. J. L. Banks, H. S. Beard, Y. Cao, A. E. Cho, W. Damm, R. Farid, A. K. Felts, T. A. Halgren, D. T. Mainz, J. R. Maple, R. Murphy, D. M. Philipp, M. P. Repasky, L. Y. Zhang, B. J. Berne, R. A. Friesner, E. Gallicchio and R. M. Levy, J. Comput. Chem., 2005, 26, 1752–1780 CrossRef CAS PubMed.
  56. W. L. Jorgensen, J. Chandrasekhar, J. D. Madura, R. W. Impey and M. L. Klein, J. Chem. Phys., 1983, 79, 926–935 CrossRef CAS.


Electronic supplementary information (ESI) available. See DOI: 10.1039/c8tb00211h

This journal is © The Royal Society of Chemistry 2018