Regular paper on-line at: www.actabp.pl Isolation and characterisation of crocodile and python ovotransferrins

Transferrins play a major role in iron homeostasis and metabolism. In vertebrates, these proteins are synthesised in the liver and dispersed within the organism by the bloodstream. In oviparous vertebrates additional expression is observed in the oviduct and the synthesised protein is deposited in egg white as ovotransferrin. Most research on ovotransferrin has been performed on the chicken protein. There is a limited amount of information on other bird transferrins, and until our previous paper on red-eared turtle protein there was no data on the isolation, sequencing and biochemical properties of reptilian ovotransferrins. Recently our laboratory deposited ten new sequences of reptilian transferrins in the EMBL database. A comparative analysis of these sequences indicates a possibility of different mechanisms of iron release among crocodile and snake transferrin. In the present paper we follow with the purification and analysis of the basic biochemical properties of two crocodile (Crocodilus niloticus, C. rhombifer) and one snake (Python molurus bivittatus) ovotransferrins. The proteins were purified by anion exchange and hydrophobic chromatography, and their N-terminal amino-acid sequences, molecular mass and isoelectric points were determined. All three proteins are glycosylated and their N-glycan chromatographic profiles show the largest contribution of neutral oligosaccharides in crocodile and disialylated glycans in python ovotransferrin. The absorption spectra of iron-saturated transferrins were analysed. Iron release from these proteins is pH-dependent, showing a biphasic character in crocodile ovotransferrins and a monophasic type in the python protein. The reason for the different types of iron release is discussed.


INTRODUCTION
Iron plays a crucial role in many cellular functions, such as oxygen transport and mitochondrial energy metabolism.For this reason iron deficiency is hazardous to organisms.Its overload is also undesirable on account of its favouring toxic free radical formation (Hentze & Kühn, 1996;Aisen et al., 2001).One of the most important agents in iron homeostasis and metabolism is the iron-binding protein transferrin.Transferrins are widely distributed among vertebrates as well as invertebrates (Lambert et al., 2005b).They are involved mainly in ferric ion acquisition, transport, and level control, but also in antimicrobial defence and in cell growth and differentiation (for a review, see Gomme & Mc Cann, 2005).
Vertebrate transferrins include serotransferrins, lactoferrins, and ovotransferrins.In mammals, the dual role of transferrin (iron transport and scavenging) is divided between serotransferrin and lactoferrin.The first transports iron through the bloodstream and delivers it from places of assimilation (e.g., intestine) to places of utilization (e.g., bone marrow) (Bailey et al., 1988).The second captures free ferric ions in mammalian body secretions, thereby sequestering the metal from invading microorganisms (Valenti & Antonini, 2005).
In birds there is only one transferrin gene, expressed in the liver and in the oviduct.The hepatocyte-derived protein is delivered to the bloodstream similarly as mammalian serotransferrin, while the protein synthesised in tubular gland cells is deposited in egg white as ovotransferrin (Lee et al., 1980).A similar situation probably also exists in reptiles (Ciuraszkiewicz et al., 2006).
Transferrins are monomeric glycoproteins with molecular mass of about 80 kDa.Contemporary transferrins originated by gene duplication and fusion of a single-lobe ancestor protein, resulting in a bilobal structure (Park et al., 1985).The N-and Clobes are about 40% homologous and each contains two domains (NI and NII, and CI and CII) separated by a hydrophilic cleft where the iron-binding site is located (Baker et al., 2003).The ferric ions in both lobes are coordinated by a carboxylate oxygen of an Asp residue, two phenolate oxygens of two Tyr residues, and the imidazole nitrogen of one His residue.The two remaining coordination sites are occupied by two oxygen atoms from a synergistically bound carbonate anion, anchored in the protein by Arg and Thr residues (Baker et al., 2001).
Transferrins bind and release ferric ion in a pH-dependent manner.They have the capacity to uptake Fe 3+ with a very high affinity at near-neutral pH and to release it at acidic conditions.In vivo, iron-loaded transferrin binds at a slightly basic pH (pH 7.4) to cell membrane transferrin receptor.The transferrin-receptor complex is subsequently transferred to acidic endosomes (pH 5.5) where ferric ions are released from the protein (Gianetti et al., 2005).
Despite identical ligands in both lobes there are differences in the rate of iron release from the N-and C-lobes within a specific type of transferrin (Mason et al., 2005) and between different transferrins (sero-, lacto-, and ovotransferrins) (Baker et al., 2002).Our knowledge of the mechanism of iron binding and release by transferrin is based on seqences and structural data obtained only for a few mammalian sero-and lactoferrins and hen and duck ovotransferrins (Baker et al., 2002).Many details of this mechanism are still unclear.To respond to some unresolved questions, it is necessary to perform comparative functional and structural analyses of transferrins from different animal species.More than 70 sequences are now available for alignment (Lambert et al., 2005a), but only one of reptilian origin (Ciuraszkiewicz et al., 2006).Recently we have deposited ten reptilian transferrin cDNA sequences in the EMBL database, among them the Nile crocodile (AM262323) and Burmese python (AM262317) sequences.A comparative analysis of these sequences indicates a possibility of different mechanisms of iron release from the N-lobes among crocodile and snake transferrin.To prove it experimentally we follow in this paper with a description of the isolation and properties of ovotransferrins from two crocodile and one snake species.

MATERIALS AND METHODS
Animal material.Eggs of the Nile crocodile (Crocodylus niloticus), Cuban crocodile (C.rhombifer) and Burmese python (Python molurus bivittatus) were obtained from the zoological gardens in Wroclaw and Warsaw (Poland).
Purification of ovotransferrins from crocodile eggs.The crocodile egg whites separated from the yolk and shell were lyophilised, suspended in 0.05 M sodium phosphate buffer, pH 7.0, containing 0.15 M NaCl, and then centrifuged (20 000 × g, 10 min, 4°C).The supernatants were fractionated using 30%, 50%, and finally 80% ammonium sulphate saturation.Following dialysis against 0.02 M Tris/HCl buffer, pH 8.5, the solutions of proteins obtained in the last step of precipitation were applied on a chromatographic column.The purification procedure included anion exchange chromatography and hydrophobic chromatography.Anion exchange chromatography was performed using a MonoQ HR 5/5 column and a fast protein liquid chromatography (FPLC) system (Pharmacia).The column was equilibrated with 0.02 M Tris/HCl buffer, pH 8.5.Fractions of 1 ml were collected at a flow rate of 1 ml/min.Proteins were eluted with NaCl gradient (0-1.0M) with a total gradient volume of 30 ml.Hydrophobic chromatography was carried out on a PhenylSuperose HR 5/5 column (Pharmacia) using the FPLC system.The column was equilibrated with 0.05 M sodium phosphate buffer, pH 7.0, containing 1.7 M (NH 4 ) 2 SO 4 .Fractions of 1 ml were collected at a flow rate of 0.5 ml/min.Proteins were eluted with (NH 4 ) 2 SO 4 gradient (1.7-0 M) with a total gradient volume of 21 ml.
Purification of ovotransferrins from python eggs.Python ovotransferrin was isolated from egg yolk and white using the Kim and Nakai modified method (1996) as a first step.Briefly, egg yolk and white were diluted 10 times with distilled water, adjusted to pH 5.0 with 0.1 M HCl, kept for 2 h at 4°C, frozen at -20°C for at least 24 h, then slowly thawed at 4°C overnight, and finally filtered and centrifuged (20 000 × g, 20 min, 4°C).The supernatant was fractionated using 30%, 50%, and finally 80% ammonium sulphate saturation.The final pellet was dissolved and dialysed against 0.02 M Tris/HCl buffer, pH 8.5, and applied on a MonoQ HR 5/5 column equilibrated with the same buffer.Fractions of 1 ml were collected at a flow rate of 1 ml/min.Proteins were eluted with NaCl gradient (0-1.0M) with a total gradient volume of 26 ml.
Determination of protein concentration.Protein concentration was determined by the modified Bradford method (Zor & Selinger, 1996) using Roti reagent (Roth) and bovine serum albumin as a standard.
Polyacrylamide gel electrophoresis.SDS/ PAGE was carried out in 10% (w/v) or in gradient 4-16% (w/v) gel by the method of Schägger and von Jagow (1987).Proteins were stained with CBB R-250.
Amino-acid sequence analysis.After SDS/ PAGE, Nile crocodile and python proteins were transferred onto polyvinylidene fluoride membrane (Bio-Rad) in 10 mM Caps buffer, pH 11.0, with 10% (v/v) methanol for 1.5 h at a constant current of 0.15 A and subjected to amino-acid sequence analysis on a gas-phase sequencer (Model 491, Perkin Elmer-Applied Biosystems) (BioCenter of Jagiellonian University, Kraków, Poland).The phenylthiohydantoin derivatives were analysed by on-line gradient high-performance liquid chromatography on a Microgradient Delivery System Model 140C equipped with a Programmable Absorbance Detector Model 785A (both from Perkin Elmer-Applied Biosystems).
Determination of molecular mass.Molecular mass was determined using SDS/PAGE and mass spectrometric analyses.SDS/PAGE was carried out in gradient 4-16% (w/v) gel by the method of Schägger and von Jagow (1987).Proteins were stained with CBB R-250.Mass spectrometric analysis was performed with a MALDI-TOF Reflex IV mass spectrometer (Bruker-Daltonics) (Neurobiochemistry Unit, Faculty of Chemistry, Jagiellonian University, Kraków, Poland).Samples were dissolved in a methanol/water mixture (3 : 7, v/v) and mixed with an equal volume of saturated matrix solution (sinapinic acid dissolved in a water/acetonitrile mixture, 1 : 2, v/v).After mixing, 1 μl of the prepared sample was applied on a steel plate and left to dry.The plate with the dried sample was inserted into the spectrometer.The scan range was set to 10-200 kDa.The mass spectrometer was calibrated according to the standard procedure using Bruker's calibration mixture.
Determination of isoelectric points of transferrins.To determine the isoelectric points of the investigated proteins, chromatofocusing separations on a Mono P HR 5/20 column (Pharmacia) were performed.Lyophilised, purified and desalted protein samples (about 100 µg) were resolved in the respective starting buffer as stated below.The chromatography in pH gradient was performed at a flow rate of 1 ml/min.The pH interval for crocodile transferrins was from 7 to 5 with 0.025 M bis-Tris/HCl, pH 7.1, as a starting buffer and Polybuffer 74/HCl, pH 5.0, as eluent.For python transferrin pI determination, the pH interval was from 6 to 4 with 0.025 M bis-Tris/HCl, pH 6.3, as a starting buffer and Polybuffer 74/HCl, pH 4.0, as an eluent.Fractions (1.0 ml) were monitored spectrophotometrically at 280 nm.The pH of each collected fraction was measured using a Metrohm 691 pH Meter.The purity of samples containing eluted proteins was additionally checked by SDS/PAGE.
Absorption spectra of iron-saturated transferrins.Purified transferrins were dialysed against 0.1 M Tris/HCl buffer, pH 8.5, containing 5 mM Na 2 CO 3 and saturated with iron by incubating overnight at 10°C with FeCl 3 at a molar ratio of 1 : 4. Excess iron was removed by dialysing the iron-saturated proteins against 0.1 M Tris/HCl buffer, pH 8.5, containing 0.15 M NaCl.Absorption spectra were recorded for 1.9 mg/ml protein solutions over the range of 400-560 nm using a Beckman DU 640 Spectrophotometer.
The pH-dependent iron release.The pH-dependent iron release was monitored following dialysis of the protein solution against a series of buffers over the pH range 3.5-8.5, each containing 0.15 M NaCl.The buffers used for these experiments were: 0.05 M Tris/HCl in the pH range 7.0-8.5,0.05 M Mes/NaOH in the pH range 6.0-6.5, 0.05 M sodium acetate buffer in the pH range 4.0-5.5, and 0.05 M glycine/HCl buffer at pH 3.5.At each pH value the percentage of iron saturation was estimated assuming A 465 /A 280 at pH 8.5 as 100% and at pH 3.5 as 0% iron saturation.

Purification of transferrins from eggs
The applied procedure allowed for simple and efficient purification of crocodile and python ovotransferrins.As a result, electrophoretically pure proteins were obtained.The presence of individual proteins during preparation was monitored by SDS/ PAGE.Initially, transferrins were identified on the basis of molecular mass.
Crocodile transferrins were purified from egg white in a four-step procedure including egg white lyophilisation, ammonium sulphate fractionation, anion exchange and hydrophobic chromatography.Lyophilisation of egg white followed by its suspension in 0.05 M sodium phosphate buffer, pH 7.0, containing 0.15 M NaCl enabled the elimination of structural components, such as ovomucins.The relatively large protein loss at this stage indicates that a significant portion of soluble proteins was also entrapped within the insoluble pellet.Recovery attempts by pellet resuspension in a fresh buffer did not give satisfactory results.Ammonium sulphate fractionation allowed discarding the majority of the soluble egg white proteins in the first and second precipitation steps (30% and 50% salt saturation).Protein precipitated with 80% ammonium sulphate saturation was dissolved and loaded several times onto a MonoQ anion exchange column (Fig. 1A).Fractions (8-11) containing an 80-kDa protein, as determined by SDS/ PAGE (Fig. 2), were pooled, dialysed against 0.05 M sodium phosphate buffer, pH 7.0, and (NH 4 ) 2 SO 4 was added to a final concentration of 1.7 M. The protein solution was repeatedly applied onto a PhenylSuperose column (Fig. 1B).Starting from 6 g of lyophilised egg white, 7.5 mg and 4.1 mg of electrophoretically pure Cuban and Nile crocodile transferrins (fractions no.9), respectively, were obtained (Fig. 2).

Figure 1. Chromatography of crocodile egg white proteins.
Cuban (□) and Nile (Δ) crocodile egg white proteins were purified by ion-exchange chromatography on a MonoQ HR 5/5 column (A) and hydrophobic chromatography on a PhenylSuperose HR 5/5 column (B).About 10 mg of protein in 0.02 M Tris/HCl buffer, pH 8.5, was applied on a MonoQ column equilibrated with the same buffer.Proteins were eluted with a nonlinear gradient from 0 to 1 M NaCl (30 ml) at a flow rate of 1 ml/min (fraction volume: 1 ml).Pooled fractions (marked) were applied on a PhenylSuperose HR 5/5 column equilibrated with 0.05 M sodium phosphate buffer, pH 7.0, containing 1.7 M (NH 4 ) 2 SO 4 .A solution containing 5.2 mg of protein in starting buffer was applied on the column.Proteins were eluted with (NH 4 ) 2 SO 4 gradient (1.7-0 M) with a total gradient volume of 21 ml.Fractions of 1 ml were collected at a flow rate of 0.5 ml/min.Fractions no. 9 contained electrophoretically pure crocodile transferrins.

Crocodile and python ovotransferrins
Python ovotransferrin was purified in a threestep procedure including acidic extraction from whole egg content, ammonium sulphate fractionation, and anion exchange chromatography.Isolation was carried out from a mixture of yolk and white since the egg white content in Squamata eggs is very low and it is very difficult to separate it from the yolk.Mild acidic extraction followed by freezing, slow thawing, and filtration allowed discarding lipid egg content.Ammonium sulphate fractionation eliminated the great majority of egg proteins from the water soluble fraction.The protein precipitated with 80% salt saturation was dissolved and applied on a MonoQ anion exchange column (Fig. 3).As a result of multiple separations 4.6 mg of electrophoretically pure snake transferrin (fraction No. 17) was obtained (Fig. 4).
The purity of the isolated proteins was confirmed by overloaded (10 μg per well) SDS/PAGE (Fig. 5).

Amino-acid sequence analysis of purified transferrins
The purified proteins were identified as transferrins by the N-terminal amino-acid sequencing.Sequences were determined for the Nile crocodile and python transferrins only (Fig. 6), since both crocodile proteins were isolated by exactly the same procedure.The obtained results confirmed that the purified proteins belong to the transferrin family.

Molecular mass, isoelectric points, and glycosylation
The transferrins' molecular mass values determined by SDS/PAGE and MALDI-MS isoelectric points determined by chromatofocusing on a Mono P HR 5/20 column are shown in Table 1.Both crocodile transferrins exhibit very similar pI values (about 5.7), whereas the snake protein is more acidic (pI = 4.55).These values are significantly different from the values calculated for the amino-acid composition deduced from the available mRNA sequences (Nile crocodile transferrin EMBL ID: AM262323, Mm = 74262 Da, pI = 6.29;Burmese python transferrin EMBL ID: AM262320, Mm = 75514 Da, pI = 5.59).These discrepancies can be attributed to posttranslational modifications such as N-glycosylation.N-glycan profiling on a GlycoSep C column (Fig. 7) shows the largest contribution of neutral oligosaccharides in Nile crocodile ovotransferrin and acidic disialylated glycans in the python protein.In crocodile transferrin the predominance of neutral chains (with Mm about 2 kDa each) in the 6-kDa glycan moiety allows the assumption that all three potential N-glycosylation sites are occupied.The python protein glycan moiety has a lower molecular mass (4.6 kDa) and is composed mostly of larger sialylated chains, indicating that probably only two of the three potential N-glycosylation sites are actually occupied.Each time, 5.1 mg of protein in 0.02 M Tris/HCl buffer, pH 8.5, was applied on a MonoQ column equilibrated with the same buffer.Proteins were eluted with a nonlinear gradient from 0 to 1 M NaCl (26 ml) at a flow rate of 1 ml/min (fraction volume: 1 ml).Fraction no. 17 contained electrophoretically pure python transferrin.

Absorption spectra of iron-saturated transferrins and pH-dependent iron release
Transferrins saturated with ferric ions display a characteristic salmon-pink colour which is attributed to a charge transfer from tyrosine phenolate to the metal (Zak et al., 2002).Solutions of Fe-transferrins have an absorption minimum and maximum at about 410 nm and 465 nm, respectively.An A 280 / A max ratio near 22 indicates complete iron saturation and protein homogeneity (van Renswoude et al., 1982).
Absorption spectra over the range of 400-560 nm of the purified iron-saturated transferrins were recorded (Fig. 8).All three protein solutions display characteristic salmon-pink colour and have an absorption minimum at 410 nm and a maximum in the range of 460-480 nm.The calculated A 280 /A 465 ratios are 22.9, 23.0, and 24.9 for the Cuban crocodile, Nile crocodile, and Burmese python transferrins, respectively.These values confirm the saturation of iron binding to the examined proteins, as well as their homogeneity.A decrease in the A 465 /A 280 ratio at lower pH values confirms the pH-dependence of the iron-release process (Fig. 9).Acidification causes protonation of the amino-acid ligands, resulting in a conformational change in the protein structure which leads to interdomain cleft broadening and ferric ion release (Bali et al., 2001;Zak & Aisen, 2003).Sero-and ovotransferrin release iron in the pH range of 6-4, and the lactoferrins in the pH range of 4-2 (Ward et al., 1996;Baker et al., 2002).Despite identical iron ligands in both lobes, the rate of its release from the sero-and ovotransferrin N-lobe is faster than from the C-lobe    Percentage of iron saturation was estimated assuming the A 465 /A 280 ratio at pH 8.5 as 100% and at pH 3.5 as 0% iron saturation.

Figure 3 .
Figure 3. Ion-exchange chromatography on a MonoQ HR 5/5 column of Burmese python egg proteins.Each time, 5.1 mg of protein in 0.02 M Tris/HCl buffer, pH 8.5, was applied on a MonoQ column equilibrated with the same buffer.Proteins were eluted with a nonlinear gradient from 0 to 1 M NaCl (26 ml) at a flow rate of 1 ml/min (fraction volume: 1 ml).Fraction no. 17 contained electrophoretically pure python transferrin.

Figure 6 .
Figure 6.Comparison of the N-terminal amino-acid sequences of crocodile and python transferrins with corresponding sequences of avian and other reptilian ovotransferrins.Amino-acid residues identical in at least four sequences are shown in black boxes.Partial homology is indicated in grey boxes.Multiple sequence alignment was performed using the ClustalW program (http://www.ebi.ac.uk/clustalw).*For the duck transferrin, only the amino-acid sequence is available (SWISS-PROT P56410).

Figure 7 .
Figure 7. GlycoSep C separations of total N-glycans released from Nile crocodile and Burmese python transferrins.Glycans were fluorescently labelled with 2-AB, which is eluted in the second marked peak.The separation on the ion exchange column was performed at gradient conditions described in Materials and Methods.Elution positions of neutral sugars (N) and glycans with 1 to 3 charged groups are indicated by arrows.

Figure 8 .
Figure8.Absorption spectra of iron-saturated crocodile and python transferrins.Spectra were measured over the range of 400-560 nm for solutions with a protein concentration of 1.9 mg/ml.() Python transferrin; Cuban () and Nile () cocodile transferrin.

Figure 9 .
Figure 9. pH-dependent iron release from crocodile and python transferrins.Percentage of iron saturation was estimated assuming the A 465 /A 280 ratio at pH 8.5 as 100% and at pH 3.5 as 0% iron saturation.