The haemoglobins of healthy and anaemic Xenopus laevis

The African Clawed Toad, Xenopus laevis (Daudin), like most other tetrapods, makes different haemoglobins during larval (tadpole) and adult life respectively. Moreover, both in larval and adult life, more than one haemoglobin is manufactured at a time. It is now well authenticated that, under conditions of anaemia, some adult animals alter the pattern of haemoglobins produced. In man enhanced synthesis of the foetal haemoglobin, HbF, occurs in various anaemias (Beaven, Ellis & White, 1960; Shahadi, Gerald & Diamond, 1962). Some breeds of sheep and goats, when rendered anaemic by injection of phenylhydrazine, synthesize abnormally large amounts of a ‘Haemoglobin C’ which, in healthy adults, is present in only small amounts (Boyer, Hathaway, Pascasio, Orton & Bordley, 1966; Beale, Lehmann, Drury & Tucker, 1966; Huisman, Adams, Dimmock, Edwards & Wilson, 1967).

We have artificially induced anaemia in adult Xenopus laevis and find that many such anaemic individuals possess substantial amounts of one of their tadpole haemoglobins which, in healthy adults, is present only in trace amounts. Moreover, during recovery from anaemia these experimental animals yield adult haemoglobins in which the normal ratio between the different types is altered.

We have already demonstrated, by immunofluorescent techniques, that during the metamorphic changeover from tadpole to adult haemoglobins in Xenopus, some red blood cells contain haemoglobin of both the adult and the tadpole types (Jurd & Maclean, 1970). By applying the same immunofluorescent techniques to red blood cells from anaemic adult Xenopus, we conclude that whilst every red cell contains adult haemoglobins, the tadpole haemoglobin produced during anaemia is distributed throughout a large proportion of the cell population.

Mature adult male and female Xenopus laevis were obtained from Harris’ Biological Supplies, Weston-super-Mare, England. Animals from selected adult pairs were each injected, over a 2-day period, with 350 i.u. of Chorionic Gonadotrophin, B.P. (Organon Labs., Ltd., Morden, England) to induce amplexus. Tadpoles reared from the resulting eggs were used in our experiments. Tadpoles were kept in tap water which had been ‘aged’ by standing overnight to eliminate chlorine and to equilibrate it with the room temperature of 20 °C. Adult toads were fed twice weekly on chopped ox heart and monthly on chopped ox liver; tadpoles were fed daily with a thin suspension of Complan (Glaxo, England) in water.

Except where otherwise specified, all reagents were supplied by British Drug Houses Ltd., Poole, England.

Two methods of induction were used. (1) Mature adult Xenopus, between 60 and 100 mm long from mouth to anus, were anaesthetized by immersion in 0·2 % methane tricaine sulphonate (MS 222, Sandoz Products Ltd., London). The body cavity was opened by making a small ventral incision in the skin and body wall just posterior to the xiphisternal cartilage, and between 2 and 11 ml of blood, depending upon the size of the toad, were extracted from the heart by ventricular puncture. The ventral incision was then sutured and the toad was allowed to recover from anaesthesia. A volume of Rugh’s Amphibian Ringer Solution (Rugh, 1962) equal to the volume of blood extracted was injected into the dorsal lymph sac to restore the salt balance of the animals. A count of red blood cells per ml of blood before the induction of anaemia was made. (2) Other adult Xenopus, of similar size, were made anaemic by 2 subcutaneous injections, each of 0·05 ml of a 1 % solution of phenylhydrazine in 0·01 N HC1, on day 1 and on day 5. A red blood cell count of peripheral blood extracted from a severed vein in the interdigital foot web was made before the first injection.

Fourteen days after the toads were initially bled, or 14 days after the first injection of phenylhydrazine hydrochloride, the animals were anaesthetized in MS 222, and 0·5 ml of blood was collected by ventricular puncture. A red blood cell count was again made. In some animals further samples of blood were taken 28 and 42 days after the induction of anaemia.

The haemoglobin from 0’5 ml of blood extracted from adult toads before or after the induction of anaemia was labelled by incubating the blood cells in vitro with [³H]leucine, as previously described (Maclean, Brooks & Jurd, 1969). After incubation the cells were lysed in 5·0 ml of distilled water to which 0·3 ml of toluene had been added. The mixture was shaken thoroughly and stood for 10 min to ensure complete cell lysis; after centrifuging for 60 min at 8000g, the clear red supernatant haemolysate was collected.

The haemolysate to be used for carboxymethyl-cellulose cation exchange chromatography was dialysed overnight at 2 °C against 200 vol. of stirred 0·01 M sodium phosphate at pH 6·10, containing 0·01 % dithiothreitol (DTT - Calbiochem Ltd., Los Angeles, U.S.A.) to reduce disulphide bridges between adjacent globin chains, and so prevent polymerization of the haemoglobin (Cleland, 1964; Sullivan & Riggs, 1967). Haemolysates to be used for polyacrylamide gel disk electrophoresis were similarly dialysed against a tris-(hydroxymethyl)-methylamine (Tris)/diamino-ethane-tetra-acetic acid (EDTA)/boric acid buffer at pH 8·6, also containing 0·01 % DTT. Haemolysates were oxygenated by shaking in air or by bubbling oxygen through them for 10 min at 20 °C prior to electrophoresis and chromatography.

Control haemolysates were similarly prepared from tadpole red blood cells (extracted by incising the hearts of anaesthetized tadpoles whilst they were immersed in Ringer solution), and from artificially mixed tadpole and adult cells. All cells were incubated with tritiated leucine prior to lysis.

Haemolysates from the incubated red blood cells were subjected to cation-exchange column chromatography on Whatman Chromedia grade CM52 preswollen carboxymethyl cellulose (sold by H. Reeve Angel & Co. Ltd., London) in a sodium phosphate pH gradient, as previously described (Maclean et al. 1969). The pH values of fractions eluting from the chromatography column were measured; the fractions were then assayed for optical density at 410 nm, and, by liquid scintillation counting, for tritium activity.

Polyacrylamide-gel disk electrophoresis was carried out in a Shandon SAE 2731 disk electrophoresis apparatus using a continuous TEB buffer system. The buffer used was an aqueous solution containing 10·9 g/1. Tris, 0·6 g/1. EDTA, and 3·1 g/1. boric acid. The pH was corrected to 8·60 with 0·1 N NaOH; 0·01 % DTT was added as a thiol agent. Gels containing 10% acrylamide were used: it was found that the addition of 10% glycerol to the gels greatly reduced diffusion of the protein bands.

The system was pre-run for 2 h at 2·5 mA per gel tube to effect equilibration and to remove excess persulphate ions (added to facilitate the polymerization of the acrylamide) which could cause an artificial heterogeneity of the proteins (Mitchell, 1967).

Oxygenated haemolysates from red cells from anaemic and healthy control Xenopus adults, dialysed against the TEB electrophoresis buffer as described above, were diluted with the buffer to bring their haemoglobin concentration to 15 mg/ml; 100 mg of sucrose was added to each ml of haemolysate to increase density, then 0·02 ml of the haemolysate was layered on to the top of the polyacrylamide gel and electrophoresis of the proteins was effected at 2·5 mA per gel tube for between 2 and 4 h. Haemoglobin migrated towards the anode. The apparatus was maintained at 2 °C in a refrigerator throughout electrophoresis.

On completion of electrophoresis the gels were photographed against transmitted white light, and were then fixed and stained for 24 h in a solution consisting of 50 ml methanol, 50 ml glycerol, 10 ml glacial acetic acid and 0·8 g naphthalene black 12B. Stain not bound to protein was removed electrophoretically by placing the gel at 90° to an applied electric field in a Petri dish containining 7% acetic acid, as described by Shaw (1969); unbound dye migrated towards the cathode. Gels were subsequently stored in 5 % acetic acid. Some gels were stained specifically for haemoglobin with benzidine immediately after electrophoresis, as described by Moss & Ingram (1968a).

Antisera specific to the pooled adult haemoglobins (‘Xenopus-HbA’) and to the pooled tadpole haemoglobins (‘Xenopus-HbF’) of Xenopus were raised in rabbits and guinea-pigs respectively as previously described (Jurd & Maclean, 1969, 1970). The γ-globulin fraction was separated on diethylaminoethyl cellulose at pH 7·5 according to the method of Stanworth (1960). The anti-Xenopizj-HbA immunoglobulin was treated with acetone-dried pig-liver powder (Burroughs Wellcome & Co. Ltd., London) and with Xenopus-HbF to remove nonspecific activity; similarly the anti-Xenopus-HbF was treated with pig-liver powder and Xenopus-HbA.

Ouchterlony (1949) agar-gel immunodiffusion plates were prepared using a 1% Bacto-Agar (Difco Labs., Detroit, U.S.A.) gel containing 0 ·15% sodium azide as a bacteriostatic agent.

Haemoglobin from anaemic animals was tested on the plates against anti-Xenopus-HbA and anti-Xenopus-HbF antisera, and the appearance of any precipitin lines between the immunoglobulin and the haemoglobin was noted. Control wells on the same immunodiffusion plates were set up using haemoglobin taken from the same adult toad before induction of anaemia, and Xenopus-HbF from young tadpoles. All haemolysates contained 0·01 % DTT.

Permanent preparations of the plates, stained with 0·1% azocarmine, were subsequently made following the method of Uriel (1964).

Anti-Xenopus-HbA and anti-Xenopus-HbF immunoglobulins, prepared and purified as described above, were conjugated with fluorescein isothiocyanate (FITC - Calbiochem Ltd., Los Angeles, U.S.A.) after the method of Naim (1969). The anti-adult haemoglobin and antitadpole haemoglobin conjugates, known as anti-Xenopus-HbA-FITC and anti-Xewopiw-HbF-FITC respectively, were further purified by a second absorption with pig-liver powder to remove non-specific fluorescence (Curtain, 1958).

Red blood cells from anaemic Xenopus were washed in Ringer solution and smeared on to slides which were then air-dried and fixed for 2 min in 95% ethanol - a method modified from Sainte-Marie (1962) and Wild (1970). The smears were treated with the fluorescent immunoglobulin as previously described (Jurd & Maclean, 1970). The proportion of cells fluorescing when the smear was viewed by ultraviolet illumination at 360 nm was noted.

Red blood cell smears from anaemic animals were treated with anti-Xenopus-HbA-FITC and anti-Xenopus-HbF-FITC. Both fluorescent antisera were also used to test the blood cells from the same animals prior to the induction of anaemia, and cells from other control animals, from tadpoles, and from artificial mixtures of adult and tadpole cells.

The elution pattern obtained when adult Xenopus red blood cells were incubated with tritiated leucine and the haemoglobin chromatographed on carboxymethyl cellulose in a sodium phosphate pH gradient has already been described (Maclean et al. 1969). We now feel that this pattern must be slightly modified in the light of further experimentation. Fig. 1 illustrates an elution pattern typical of the haemoglobins of healthy adult toads. Optical absorption at 410 nm indicates 2 significant haemoglobin peaks. The first is eluted from the column at pH 7·38 and contains 92% of the total haemoglobin eluted (as measured by optical absorption), and 92 % of the tritium activity; the second peak is much smaller, eluting at pH 7·85 and containing 5 % of the haemoglobin as measured by optical absorption, and 5 % of the radioactivity. Similar elution patterns were obtained from the haemoglobins of all other healthy adult toads tested: chromatography data from 8 such toads are summarized in the top part of Table 1. It will be noted that most of the toads show very small optical absorption and radioactivity peaks in the fraction eluting from the column at pH 6·5 (+ 0 ·1) and pH 7 · 15 (±0 ·1): the amount of haemoglobin in these peaks is always less than 2% of the total.

The main adult haemoglobin peak, which elutes at pH 7·35 (± 0·07), we designate ‘Xenopus-HbA, ’: this haemoglobin corresponds to the combined ‘H₁’ and ‘H₂’ haemoglobins described by Maclean et al. (1969): our recent studies show that the ‘H₂’ peak closely associated with the ‘H/peak is an artifact caused by varying degrees of oxygenation of the haemoglobin. The minor adult haemoglobin eluting at pH 7·85 (± 0 ·05) is designated ‘Xenopus -HbA₂’, and corresponds to the previously described ‘H₃’ haemoglobin (Maclean et al. 1969).

Fig. 2 shows the elution pattern obtained when the pooled red blood cells from 10 stage-50 (Nieuwkoop & Faber, 1967) tadpoles were incubated with tritiated leucine and the haemoglobin was subsequently chromatographed on carboxymethyl cellulose. Three elution peaks are seen. The first elutes at pH 6·53 and contains 56% of the haemoglobin as measured by optical absorption at 410 nm, and 57% of the tritium activity; the second elutes at pH 7·13, contains 20% of the haemoglobin and 17% of the radioactivity; the final peak is eluted from the column at pH 7·36 and comprises 22 % of the haemoglobin and 23 % of the radioactivity. That this pattern is repeatable is shown from the data in the lower part of Table 1. It will be noticed that, as the tadpoles get older the proportion of haemoglobin eluting at pH 7·10 (± 0·05) decreases slightly and the proportion eluting at pH 6·50 (± 0·08) increases.

We find that the haemoglobin from tadpoles eluting from the column at pH 7·35 (± 0·05) gives a positive precipitin reaction with specific anti-Xenopui-HbA antiserum: the haemoglobins eluting at pH 6·5 (±0·08) and pH 7·10 (±0·05) give no such reaction. We therefore conclude that the pH 7·35 haemoglobin is identical to Xenopus-HbA₁ This conclusion is further justified by the fact that as tadpoles get older the relative amount of this haemoglobin progressively increases until, about 10 weeks after metamorphosis, it comprises over 80% of the haemoglobin present and the overall elution pattern resembles that of a mature adult toad.

The pH 6·50 and pH 7·10 eluate fractions are designated ‘Xenopus^,-HbF₁’ and ‘Xenopus-HbF₂’ respectively. After metamorphosis these haemoglobins disappear almost completely: the existence of very small absorption peaks at pH 6·50 and pH 7·15 in haemolysates from mature adult toads may be explained by the persistence of minute amounts of the foetal haemoglobins into post-metamorphic life.

When haemoglobins from artificially mixed tadpole and adult red blood cells are chromatographed, 4 optical absorption peaks are observed which elute at pH values corresponding to the 2 foetal and the 2 adult haemoglobin peaks respectively.

Haemoglobin from the red blood cells of healthy adult Xenopus, when electro-phoresed on polyacrylamide gel in a continuous TEB buffer system at pH 8·60, migrates towards the anode in 2 visible bands. Most of the haemoglobin migrates in one fast band (a₁ in Fig. 5), but a faint slower band (a₂) is also detectable. Benzidine staining shows that both bands are haemoglobins. Naphthalene black staining also reveals these 2 bands, together with 2 fainter, faster bands which are not detectable by benzidine staining. These faster bands probably represent non-haem proteins such as carbonic anhydrase present in the cell lysate.

When haemolysates from Xenopus tadpole red blood cells are similarly electro-phoresed, 4 visible bands, which also stain with benzidine, are present in the gel. A strong band (t₃ in Fig. 5) migrates towards the anode somewhat faster than the main adult band. Another band, migrates more slowly than the main adult, but faster than the slow adult band. A third band (t₂) migrates at a speed intermediate to and t₃, and at the same speed as the main adult band. A faint, very fast fourth band (t₄) is also detectable. If the bands are excised from the gel and the haemoglobins are eluted into 0·3 ml of 0·85 % saline, eluates from the fast (t₃), very fast (t₄), and slow (t₄) tadpole bands give a positive precipitin ring reaction with specific anti-Xenopus- HbF antiserum, but not with anti-Xenopus-HbA antiserum; the t₂ eluate, and the eluates from both the adult bands give a positive precipitin ring reaction with anti-Xenopus-HbA, but not with anti-Xenopus-HbF antiserum.

We therefore conclude that the tadpole haemoglobin comprising electrophoresis band t₂ is identical to that which comprises the main adult haemoglobin band. This view is reinforced by the observation that in haemolysates from young, pre-stage 50, tadpoles the t₂ band is almost or completely absent, but as the tadpoles get older the proportion of this band increases. The remaining 3 tadpole bands, t₁t₃, t₄, are genuine tadpole haemoglobins. Artificial mixtures of adult and tadpole haemoglobins show the presence of both adult bands and all 3 true tadpole bands.

When the Xenopus-HbA₁ (pH 7·35) eluate fractions from carboxymethyl-cellulose chromatography were dialysed and concentrated by ultrafiltration against TEB buffer at pH 8-6o and subsequently electrophoresed on polyacrylamide it was found that the haemoglobin migrated at the same speed as the main adult (a₁) and the t₂ tadpole band: it is thus concluded that the main adult band (which is also identical to the t₂ tadpole band) represents Xenopus-Hb A₁ haemoglobin. Using similar techniques it can be shown that the pH 7·85 Xenopus-HbA₂ chromatography eluate corresponds to the slow adult electrophoresis band a₂, that the pH 6·50 Xenopus-HbF₁ eluate corresponds to the fast tadpole band t₃, and that the pH 7·15 Xenopus-HbF₂ eluate corresponds with the slow tadpole band It has not proved possible to correlate the very fast tadpole band t₄ with any of the carboxymethyl cellulose fractions.

Following the induction of anaemia in adult Xenopus by bleeding, the blood cell count 14 days later was reduced by between 28% and 63%. It was noted that the peripheral blood cells contained abnormally large numbers of early and mid-polychromatic erythrocytes in addition to the usual mature erythrocytes - the nomenclature used is taken from Lucas & Jamroz (1961). Fig. 6 shows a smear made from blood cells from an adult toad before it was made anaemic and its red blood cell count was 731 × 10⁶ red cells/ml; Fig. 7 shows blood cells from the same animal 14 days after 6·3 ml of blood had been extracted from the ventricle and the cell count had been reduced to 281 × 10⁸ cells/ml.

Injection of phenylhydrazine hydrochloride reduced the blood cell count of adult Xenopus by between 54% and 98% after 14 days. In extreme anaemia the majority of circulating cells are seen to be blast cells and erythroblasts (Fig. 8). Four out of 10 animals made anaemic in this way died between 14 and 20 days after the initial phenylhydrazine injection; the others recovered. The count of circulating cells started to increase again after about 21 days, but it was between 6 and 12 weeks before the normal population of mature erythrocytes was restored.

Table 2 shows the results of chromatography of haemolysates from red blood cells of 18 adult Xenopus before and 14 days after they were made anaemic. Prior to chromatography the cells were incubated with tritiated leucine: as might be expected, in anaemia there is an overall increase in the synthetic activity of the red blood cells measured in terms of tritiated leucine incorporation. It will be noted that in 15 out of the 18 animals (the exceptions are animals 3, 4 and 7) there is a significant increase in the proportion of tritium activity in the pH 6·5 Xenopus-HbF₁ elution peak. Before anaemia there was never more than 1 % of the total radioactivity in the fractions eluted from the column in the pH 6·5 peak: in 14 anaemic animals the proportion of the total radioactivity in the pH 6·5 peak varies between 12% and 48%. The proportion of radioactivity in the pH 7·15 Xenopus-HbF₂ peak increases by up to 3%, but the percentage of radioactivity in the pH 7·85 Xenopus-HbA₂ peak does not increase. In three of the animals made anaemic (Nos. 9, 11 and 14) there is a markedly enhanced optical absorption peak, measured at 410 nm, in the pH 6·5 fraction. It is noticeable that the proportion of tritium activity in the pH 6·5 fractions of these 3 animals is not as high as in some toads which do not show increased optical absorption, demonstrating that after an appreciable amount of the new haemoglobin has been made, its synthesis tends to slow down. The optical absorption data from animal 9 are represented graphically in Fig. 3.

When the haemolysate from anaemic toad 9 was submitted to polyacrylamide-gel disk electrophoresis a faint haemoglobin band was seen migrating at the same speed as Xenopus-HbF₁: the usual Xenopus-HbF₂ and Xenopus-AFbX, bands were also present (Fig. 9).

Six toads which had been made anaemic with phenylhydrazine recovered. On the 28th and 42nd days after the initial phenylhydrazine injection blood cells extracted from the animals were chromatographed in the usual way. The results are shown in Table 3. In 5 of the toads, after both 28 and 42 days, an abnormally high proportion of the haemoglobin (between 21 % and 58%), measured in terms both of radioactivity and optical absorption, eluted from the chromatography column in the pH 7 ·85 Xenopus-HbA₂elution peak. The amount of haemoglobin eluting from the column in the pH 6 ·5 Xenopus-HbF₁ and pH 7 ·15 Xenopus-HbF₂ fractions was not significantly greater than in healthy adult toads. The optical absorption data for animal 11, bled 42 days after the induction of anaemia, are represented in Fig. 4.

One toad, No. 15, was bled and its haemoglobin chromatographed 120 days after the induction of anaemia. It was found that the haemoglobin profile had regained normality. Optical absorption measurements showed the haemoglobin to be present in the following proportions: Xenopus-HbF₁ <0 ·5%; Xenopus-HbF₂ <0·5%; Xenopus-HbA₁, 93%; Xenopus-HbF₂, 4%.

Polyacrylamide-gel disk electrophoresis of the 28-day haemolysate from toad No. 14 is shown in Fig. 10. Two bands can be seen, both of approximately equal strength; one migrates at the same speed as Xenopus-HbA₁ and the other at the same speed as Xenopus-HbA₂.

Fig. 11 shows an Ouchterlony agar-gel immunodiffusion plate on which haemoglobin taken from toad 9 (Table 2) 14 days after the induction of anaemia was tested against specific anti-Xenopus-HbF antiserum. Precipitin lines are visible between the antiserum well and the wells containing the haemoglobin from the anaemic toad. No line is detectable between the antiserum and a control well containing haemoglobin removed from the animal before the induction of anaemia.

Fig. 12 shows a plate on which haemoglobin from toad 11 (Table 2) 42 days after the induction of anaemia was tested against specific anti-Xenopus-HbA antiserum. Two precipitin lines between the antiserum and the haemoglobin from the recovering toad are visible; only one line is detectable between the antiserum and the control well containing haemoglobin removed from the toad before the induction of anaemia.

From these results it is concluded that a greatly enhanced amount of Xenopus-HbF₁ was present in the blood of toad 9, 14 days after the induction of anaemia: similar positive results for the presence of tadpole haemoglobin were obtained from toads 11 and 14 (Table 2).

The 2 precipitin lines between the anti-Xenopus-HbA immunoglobulin and the haemolysate taken from toad 11, 42 days after the induction of anaemia, suggest the presence of a haemoglobin which normally occurs in adult toads in very small amounts. This haemoglobin cannot normally be detected on immunodiffusion plates but is detectable by the host animal when it is injected so that antibody is raised against it. On recovery from anaemia adult Xenopus have sufficient of this haemoglobin for it to be detectable by immunodiffusion. The haemoglobin is not a polymer because it is unaffected by the presence or absence of DTT. The chromatographic and electrophoretic evidence strongly implies that the enhanced amount of haemoglobin produced during recovery from anaemia is Xenopus-HbA₂.

When blood smears from healthy Xenopus adults are treated with anb-Xenopus- HbA-FITC the percentage of cells fluorescing is 100%. Treatment of tadpole cells with the same conjugate results in the number of cells fluorescing never exceeding 2% of the total. If tadpole cells are treated with anti-Xenopus-HbF-FITC the proportion of fluorescing cells always exceeds 98%; no more than 1·5% of the cells fluoresce when adult blood cells are treated with the same conjugate.

When artificial mixtures containing known proportions of adult and tadpole red blood cells were treated with anti-Xenopus-HbA-FITC or anti-Xenopus-HbF-FITC, the proportion of cells fluorescing corresponded to the known proportions of adult or tadpole cells present (Fig. 13).

Blood smears from anaemic Xenopus 9 (Table 2) taken 14 days after the induction of anaemia, and treated with anti-Xenopus-HbA-FITC showed that 100% of the cells fluoresce, and thus 100% of the cells contain adult haemoglobins. Treatment with anti-Xenopus-HbF-FITC (Fig. 14) shows that only 1-5% of the cells fluoresce, our normal background error for the technique, despite the fact that the amount of Xenopus-HbF₁ eluting from the chromatography column comprises 11 % of the total. It is therefore concluded that the tadpole haemoglobin is spread throughout a large proportion of the blood cells. Similar results were obtained from other anaemic toads.

An altered pattern of haemoglobin synthesis during anaemia is already known to occur in man (Beaven et al. 1960; Shahadi et al. 1962), sheep (Beale et al. 1966; Gabuzda, Schuman, Silver & Lewis, 1968), and goats (Huisman et al. 1967). We have reported 2 distinct changes in the haemoglobin pattern of anaemic Xenopus. One resembles the phenomenon in man in which a foetal haemoglobin, normally present in the adult as less than 1 % of the total haemoglobin, appears in increased amounts following blood loss. The other change observed in the blood picture of anaemic Xenopus is the synthesis of increased amounts of Xenopus-HbA₂ a haemoglobin normally present only in the adult, but in very small amounts. This parallels the phenomenon observed in some breeds of sheep and goats, in which the ratios of the adult haemoglobins are changed during anaemia (Beale et al. 1966; Huisman et al. 1967). In Xenopus, therefore, the evidence is that 2 haemoglobins normally present only in very small quantities in the adult appear in increased amounts after blood loss; one of these haemoglobins is a major haemoglobin in the tadpole, the other is exclusively adult. Moreover, there is complete asynchrony between the 2 changes, one appearing early and the other late after the induction of anaemia.

We have discussed elsewhere (Jurd & Maclean, 1970) the evidence for more than one haemoglobin co-existing in the same blood cell. It seems clear that in man, both adult and foetal haemoglobins occur together in some cells of the new-born (Hosoi, 1965; Dan & Hagiwara, 1967; Schneider & Haggard, 1955; Rosenburg, 1970). We predict that a similar situation occurs in induced types of anaemia in man, that the foetal haemoglobin will be found widely distributed in many blood cells together with adult haemoglobin. Our evidence from the use of fluorescent antisera with anaemic Xenopus strongly suggests that the tadpole haemoglobin is present, in low concentration, in many red blood cells.

Two conclusions emerge from this work. The first is that there is considerable flexibility inherent in the machinery responsible for haemoglobin synthesis. We believe that this should not necessarily be regarded as flexibility operated at the genetic level. On the one hand, the altered pattern of synthesis is distributed throughout many cells, and is not achieved by a total * switch ‘in a few cells. On the other hand, in no case is it apparent that the synthesis of a previously absent protein is involved. The production of haemoglobin C in some breeds of sheep (Beale et al. 1966) and in goats (Huisman et al. 1967) during anaemia appears to involve a change in proportions of the different haemoglobins, but not the appearance of a totally novel protein. So, in Xenopus, we find greatly increased production after blood loss of 2 haemoglobins previously present only in very small amounts. It follows that the phenomenon of altered haemoglobin synthesis may involve only a rate control mechanism; rate control of the synthetic product at a translational level is already established as a mechanism in the synthesis of haemoglobin (Baglioni & Colombo, 1964).

The second conclusion is that the experience of a cell during erythropoiesis probably determines its eventual synthetic products. Since the tadpole haemoglobin seems to be widely distributed amongst many cells during anaemia, its reappearance cannot be attributed to the entry into circulation of the descendants of stem cells committed solely to its production. Instead it appears that the protein-synthetic pattern of many cells is altered by their exposure during erythropoiesis to plasma factors released during anaemia. Evidence for this is particularly clear from the work of Gabuzda on anaemic sheep (Gabuzda et al. 1968) and the work of Thurmon et al. (1970) provides strong evidence for identifying one plasma factor as an erythropoietin. What should not be overlooked is that the process of increased rate of cell division in the erythropoietic tissue induced by anaemia may itself lead to the altered pattern of globin synthesis, an idea originally proposed by Baglioni (1963).

The changing pattern of haemoglobin synthesized during the metamorphosis of amphibians (Moss & Ingram, 1968a, b; Jurd & Maclean, 1970) also provides information on the regulation of protein synthesis. It is noteworthy that the adult type haemoglobin which appears in the blood of thyroxin-treated bullfrog tadpoles (Moss & Ingram, 1968b) is located predominantly within apparently immature erythrocytes.

We are currently investigating these problems further by assaying the haemoglobin synthesized by erythropoietic tissues cultured in vitro in the presence and absence of various plasma factors.

This work was supported by grants from the Wellcome Trust and the Medical Research Council. The technical assistance of Mrs B. Streets and Mrs Y. Baynes is gratefully acknowledged.

Figs. 5-8. For legend see next page.

Figs. 6—8. Blood smears stained with Wright’s Stain, × 1250 approx.

Figs. 9, 10. Samples electrophoresed at 2·5 mA per gel in Tris/EDTA/borate buffer at pH 8·60 for 3 h at 2 °C; arrows indicate origin. Photographed unstained.

Figs. 11, 12. Ouchterlony agar-gel diffusion plates, stained with azo-carmine by the method of Uriel (1964) before photography.