Characterization of Phycocyanin from Chroomonas Species*

SUMMARY Phycocyanin, a chromoprotein from Chroomonas sp., is characterized in regard to its size, subunit structure, amino acid composition, and spectroscopic properties. It is a monodisperse protein of 50,000 daltons and is composed of two polypeptide chains of 10,000 and two chains of 16,000 daltons. The proposed structure of the “native” protein is CU$Z. The /I s?~,,, is 4.4 S, and the partial specific volume is 0.73. Unlike C-phycocyanin, a functionally related chromoprotein from blue-green and red algae, phycocyanin does not form a number of different aggregates. We suggest that the absence of these larger aggregates is related to its in vivo location. The amino acid composition of phycocyanin from Chroomonas sp. differs extensively from that of C-phy-cocyanin, with much greater amounts of serine, half-cystine, and lysine. Likewise, the circular dichroism and fluorescence spectra are very different, indicating major functional modifications.


SUMMARY
Phycocyanin, a chromoprotein from Chroomonas sp., is characterized in regard to its size, subunit structure, amino acid composition, and spectroscopic properties.
It is a monodisperse protein of 50,000 daltons and is composed of two polypeptide chains of 10,000 and two chains of 16,000 daltons. The proposed structure of the "native" protein is CU$Z. The /I s?~,,, is 4.4 S, and the partial specific volume is 0.73. Unlike C-phycocyanin, a functionally related chromoprotein from blue-green and red algae, phycocyanin does not form a number of different aggregates.
We suggest that the absence of these larger aggregates is related to its in vivo location.
The amino acid composition of phycocyanin from Chroomonas sp. differs extensively from that of C-phycocyanin, with much greater amounts of serine, half-cystine, and lysine.
Likewise, the circular dichroism and fluorescence spectra are very different, indicating major functional modifications.
C-Phycocyanin and phycocrythrin are located on the stromal side of the photoaynt,hetic lamellae in the form of large, discrete aggregates callrd l)hvcobili:ollles.
This paper shows that the physical prop&es and structure of phycocyanin from Chroomoms sp. are rclutcd to the int,riguing morphological differences (7-9) betwrcn it ant1 C-phycocyanin.
This correlation is particularly trident w-lien the differences in aggregation properties are considered.
In this pnpcr, the term phycocyanin refers specifically to Chr~~monas sp. and Cphycocyanin to the chromoprotein from blue-green algae. l'ur&xdm of I'/~ycoc,yanin-A culture of the eukaryotic alga, Chroov2onas sp., was obtained from Dr. L. I'rovasoli, IIaskins Laboratory, Se\r-Haven, Conn. The alga was grown using white fluorescent lamps with an illumination of 80 foot-candles in 500 ml of UV media (10) in 2-liter Erlenmeyer flasks without agitation at 18" and was harvested by centrifugation.
The cells wcrc then resuspended in sodium phosphate buffer, pH 6.0, I 0.1.' Following a second centrifugation, the pelleted cells were stored frozen at -26".
'i'o extract the phycocyanin, the cells were suspendcd in small amounts of pI1 6.0 buffer and subjected to three cycles of freezing and thawing.
The protein which appeared in the supcrnatant was then purified (lither by ammonium sulfate gradic>nt, chromatography (11) or by tliffcrent~ial ammonium sulfa& fractionation.
(Plbsorbancc: ratio > 5.0 was considcxred an index of purity.) 'I'hcsc solutions Iv(sre ctlntrifuged at 30,000 x g to rclnlove cell debris prior lo each xltlition of ammonium sulfate. 'l'hc purified protein was usually ,~;tored in 68 c' ,(, saturated ammonium sulfntc at 4". In purification b!-ammonium sulfate gradient chromatography, the procedures n-erc as i'ollo~s. 'I'wenty milliliters of crude Chroolnonas sp. oxtl,act (AF,ij = 480) x\cre adjusted lo 80% saturated ammonium sulfate in sodium l)hosphate buffer, pII 6.0, I 0.1. This solution (100 ml) was mixed (or 20 mill with 6 g of Cclitc 545 a11d then poured into R column with a final bed size of 1 x 12 cm. Following an initial wash with 80 % saturated ammonium sulfate, a linear ammonium sulfate gradient from 80 to Oyi saturation in pH 6.0 bufl'cr was passed through the column. The initial crude extract had an ~46~~:~280 ratio of about 0.9. ISy following this proc~edurc, about 59(/A of the phycocyanin was purified to a ratio of 25.0 (Fig. 1). 'I'hc material which was initially washed from the column was soluble in 80 y0 saturated ammonium sulfate and had an A?FO:&Q ratio of 1.93, indicative of a high proportion of nucleic acid. l'hc biliprotcin undoubtedly represents a large proport,ioll of the soluble protein of the organism, since the elution profile rcvcaled no other large protein peaks.
The freeze-thaxing procttlure is quite (,fficient in releasing phycocyanin from Clrroomonas sp. For csamplc, from a packed cell volume of 24.5 cm3, r ~0 ml of soluble phycocyanin solution with an Asi15 of 49.3 and an /i645:A28" ratio oT 0.92 was obtained. whcrc w is angular velocity; R, the gas constant; ?', the temperat,ure; v', the partial specific volume; and p, the solvent density. A least squares analysis of ln absorption oersus r2 was used to obtain d ln A/dr2. A linear In A against r2 plot is expected for a homogeneous and thermodynamically ideal solute. The photoelectric scanner, multiplexer, and monochromator attachment, set at eitller 580 or 620 nm, were used. All sediment'ation equilibrium experiments were carried out in sodium phosphate buffer, pH 6.0, either in 100% H20, or in a mixture of 9% 1120 and 91% DsO. The mixture was used to determine simultaneously the partial specific volume (v) and molecular weight of the protein as demonst,rated by Edelstein and Schachman (13). These experiments were performed at 20" and 29,500 rpm in an AN-F rotor with 12-mm double sector Kel-F centerpieces and sapphire windows.
Gel Electrophoresis in Sodiuna Dodecyl Suljate-Gel electrophoresis iu the presence of sodium dodecyl sulfate, introduced by Shapiro et nl. (14), \vns performed with the procedures of Weber and Osborn (Lj).
Gels with both the normal and double the normal amomits of cross-linking reagent were used. Acrylamide (clectrophorcsis grade) was purchased from Eastman; sodium dodecyl sulfate (99%) was obtained from Sigma. Awlno Acid Co?nposition-Phycocyanin solut,ions were dialyzed exhaustively into distilled water and lyophilized. The protein was treated by a(%1 hydrolysis in 6 N RCI in evacuated tubes at 110" for 24, 48, aud 72 hours. The HCl n-as removed by rotary evaporation under reduced pressure at 40". .\miuo acid analyses were pcrformcd by the method of Spackrnan el wl. (16). Separate analyses were done for cysteic acid (17).
The two subunits of phycocyanin were separated on the gels and eluted and prepared for amino acid analysis by the method of Webcr aud Osborn (15). The two bands were clearly visible on the gels; the faster migrating band was green and the slower, blue.
Phycocyanin (0.1 mg) was placed on each gel, and the two bands were cut out of the gels and eluted overnight into a 0.1 y0 sodium dodccyl sulfate solution at 37". The solutions were separately pooled and lyophilized. Distilled water R'as added to make a 1% sodium dodecyl sulfate solution, and 9 parts of acetone were added in the cold to dissolve the detergent. TWO additional washings with acetone were performed at 0". The separated chains were then dissolved in 6 s HCl, and the amino acid analysis was accomplished as indicated aborcl. X small amount of phenol (0.5 g per liter) was present in the 6 s HCI during acid hydrolysis of the cy-and P-bands to prevent destruction of tyrosine. Xpectroscopic Measurerrlenls-Fluorescence measurements were performed on a Isaird Atomic fluorescence spcct,rometer, model SF-l, as described previously (18). Fluorescence measurements were made in buffer, p1-I 6.0, I 0.1. Visible and ultraviolet absorption and C'D spectra were obtained on a ('ary 14 spcctrophotometer and a Gary 61 circular dichrometer, resprctively. The extinction coefficient of phycocyanin was determined at 645 nm.
After the absorption of a purified sample in pH 6.0 buffer was measured, the solution was exhaustively dialyzed into distilled water. Aliquots were pipetted into w-cighetl \-esscls, dried at llO", and weighed.
A 1 mg per ml solution was calcillated to possess an absorption of 11.4 for a l-cm path length.

Gel Filtration in Cuanidine
Hydrochloride-The subunit molecular weights were estimated by gel filtration in guanidinc hydrochloride (Hcico, Delaware Water Gap, Pa.) by the m&hod of Fish et al. (19). A column, 2.1 x 41 cm, of Sepharosc 613, equilibrated with 6 RI guanidine hydrochloride and 0.1 31 2-mercaptocthanol, was used in these experiments. The column ITas operat'cd in an upward flow configuration at room tcmprrature, and fractions of 60 drops (-1.6 ml) were collected at :I flow rate of 6 ml per hour.
Samples containing 5 to 10 rng of lyophilizcd protein were dennturcd with 1.0 to 1.5 ml of the elution solvent for a minimum of 4 hours prior to application on the column.

RESULTS
Size of Native I+otein-The molecular weight of l)ll?-cocyanin in the absence of denaturation agents was determined by sedimentation equilibrium and Sephadex G-200 gel filtratioll. The average molecular weight obtained from sedimentation data from different samples in buffer, pH 6.0, I 0.1, was 4.9 x 10J =t 950 (Fig. 2). From gel filtration in phosphate buffer, p11 7.0, a molecular weight, of 5.3 x lo4 \vas obtained (Fig. 3). I%>-cocyanin was elutcd from the Sephadcx G-200 column n a sill& symmetrical peak, and there was no indication of the &c hctcrogeneity commonly seen with C-phycocyanin (3). itlcl~tical molecular weights were obtained with sedimentation equilibrium in pH 6.0 buffer composed of either 100% Ii,0 or n nlisture of 9% I-Is0 and 91 o/, 1)20. The partial specific volume determined from these studies was 0.724.
In addition, values in the vicinity of 4.9 x lo4 tlalto~ wcrc obaervtd, whether tllcl rllolloc,llroruatol was set at 580 or 620 nm. Spectrophotometric measurements at 580 and 620 nm showed a linear relationship between absorption and concentration over the range of protein concentrations at which equilibrium experiments were performed. Sedimentation velocity experiments with both schlieren and absorption optics (Fig. 4) showed a single symmetrical boundary. Several experiments over a loo-fold concentration range (between 0.04 and 5 g per liter) yielded the same sedimentation coefficient.
The s&,+ is 4.4 S. This independence of the concentration of the sedimentation coefficient suggests that a single species is present.
The sedimentation equilibrium data produced a linear In A versus r2 plot (Fig. 2), which also indicates the presence of a single phycocyanin species. To test this conclusion further, sedimentation equilibrium data were analyzed to obtain number and weight point-average slopes at various distances from the center of rotation, as indicated in the following equations (12,20). Jeffrey and Pont (20) have demonstrated that this method provides a better criterion of homogeneity, since a linear 11~4 versus r2 plot may be misleading in certain situations. The results of these analyses showed no variation in either number-or weightaverage molecular weights at any distance from the center of rotation.
In chloride and sodium dodecyl sulfate gel electrophoresis experiments both showed that phycocyanin is composed of two polypeptide chains.
In the experiments with detergent (Fig. 5), two bands were visible, one blue and the other green.
No additional protein bands were detected upon staining with Coomassie blue. When the mobilities of the two phycocyanin bands were compared with those of several proteins with known molecular weights (Fig. 6), the green band ((Y chain) was calculated to have a molecular weight of 9.9 X lo3 f 510 and the blue band (0 chain) one of 1.59 x lo4 + 570. The standard deviations were calculated from the results of six separate experiments. Identical results were obtained with a ratio of 22.2 g of acrylamide to either 0.6 or 1.2 g of methylenebisacrylamide.
When 2-mercaptoethanol was omitted from the solutions, the two bands were still observed, indicating that they are not joined by disulfide bonds.
Gel filtration in 6 M guanidine hydrochloride resolved the phycocyanin clearly into two peaks (Fig. 7) corresponding to 19,900 and 9,900 daltons (Fig. 8). The larger of these appeared blue-green and, despite a high absorbance at 400 nm, possessed a clearly defined absorption maximum at 600 nm. tion maxima of 640 and 560 nm. These probably correspond to the blue and green bands seen by sodium dodecyl sulfate electrophoresis.

Amino
Acid Composition-The amino acid compositions of phycocyanin and its two component polypeptide chains are given in Table I. The partial specific volume of phycocyanin, calculated by the method of Cohn and Edsall (al), was 0.729; this compared favorably with the 0.724 determined from sedimentation equilibrium.
A value of 0.73 was subsequently used in all calculations.
The residues in the o( and /? chains were evaluated, based on 9,000 and 15,000, respectively.
These molecular weights were based! on the sodium dodecyl sulfate gel electrophoresis and gel filtration results. If equal numbers of each chain arc present in the native protein, the sum of the QI and p residues, or an integral multiple thereof, should equal that of the whole protein, provided the choice of molecular weights was judicious.
The amino acid composition of phycocyanin differs from the reported analyses of C-phycocyanin (3-6, 18, 22, 23). To com- on their amino acid compositions (Table II). Its implications are discussed below.
ties of this protein are therefore related to its functions. Fluores-Spectroscopic Characteri&-Functionally, phycocyanin is an cence spectra (excitation, emission, and polarization) were accessory pigment affiliated with Photosystem 2 in photosynthe-obtained for a sample at pH 6.0 and A645 = 0.20 (Fig. 9). Exsis. Ilani  were calculated as described previously (17). by the selection of an excitation wave length (520 to 630 nm). Both of these facts are consistent with the suggestion that the band at 585 nm transfers its energy to the 635-nm band prior to The visible ultraviolet and CD spectra are illustrated in Fig.  10. The CD spectrum correlates well with the visible absorption maxima.

Phycocyanin
from Chroomonas sp. is a single species with a particle weight of 5.0 x lo4 and is composed of two different polypeptide chains of about 1.0 X lo4 and 1.6 X lo4 daltons. These data have led us to conclude that the native protein has a quaternary structure of (Y&Q. The 1: 1 ratio of a! and /3 chains is an assumption supported by the amino acid data. From the Svedberg equation, using the particle weight (50,000), .s!&~ (4.4 x lo-l3 S), and partial specific volume (0.73), a D of 7.9 x 10W7 cm* s-1 was calculated.
Subsequent calculation of a frictional ratio (f:finin) yielded 1.11, a value typical for globular proteins (27). The physical properties of phycocyanin are summarized iu Table 111.
The contrast in aggregation behavior between this phycocyanin and C-phycocyanin is great. C-Phycocyanin at pH 6.0 exists as a complex mixture of aggregates (3 S, 6 S, 11 S, and 19 S) . The proportion of each aggregate depends on concentration, temperature, the history of the sample, solvent, and the algal origin of the protein (3, 23, 28). Phycocyanin from Chroomonas sp. has a single and generally smaller quaternary structure (4.4 S) over a large concentration range. The difference may be indicative of the variation in the in vivo states of these chromoproteins.
C we propose that self-assembly is necessary to produce phycobilisomes. The correlation would then follow between (a) the presence of phycobilisomes and the 11 S and 19 S aggregates of C-phycocyanin and (b) the absence of phycobilisomes and the lack of the ability of assembly beyond a 4.4 S subunit with phycocyanin from Chroomonus sp.
The loss of this aggregation trait is produced by a significantly distinct amino acid composition, as shown by the difference index. The lower the indices, the more closely related are the two proteins. Fondy and Holohan (29) have used this index on other functionally related families of proteins, the pyridine-nucleotide-linked dehvdrogena,ses.
For glgceraldchyde S-phosphate dchydrogcnase, "for esampkl, taking the enzyme from pig muscle as a base, they found the following values: rabbit muscle 2.7, chicken muscle 3.5, halibut muscle 4.6, yeast 5.8, and Bscherichia coli 8.3. 'I'he indices for cytochrome c were greater; with rabbit's heart as a base, they found pig and chicken 3.8, tuna 9.2, and bakers' yeast 10.2.
Comparcd with these selected values, phycocyanin is seen to be very different from the C-phycocyanins, since the average difference index is 12.0 (Table II).
This result also indicates some residual homology, ho\vever, since 1Ietzger et al. (24) found that 75y0 of all the randomly selcctecl proteins they tested had difference indices abovc 20. The principal contributors to this difference iudes are the much larger quantities of serine, half-cyst&, and lysine in phycocyanin. The a! and fl chains of phycocyanin also differ from each other. The cx chain is richer in glutamic acid and lysine; the /3 chain is richer in serine and leucine.
It was not,cd that the a chain possesses a higher proportion of polar amino acids (glutamic acid, aspart,ic acid, serine, thrconinc, lysine, histidine, and argininc). 'I'he larger aggrcgatcs of C-phycocyanin have an increased abilit,y for energy transfer, as indicated by their greater relative fluorescence efficiency and lower fluorescence polarization (3, 30). Since phycocyanin from Chroomonas sp. is smaller, it might be expected to be comparatively inefficient in energy transfer and this would be reflected in a high fluorcsccnce polarization with respect to C-phycocyanin of similar size. However, its Uuorescence polarization at 620 nm is $0.10, compared to a p value of +0.29 for the 6 S subunit of C-phycocyanin (30). Since lower polarization means more extensive internal energy transfer (31) and since this is combined with a smaller size, phycocyanin has apparently been equipped with a system \vhereby cncrgy transfer is more efficient.
It is significant that the 9 645:A586 ratio is constant across the entire phycocyanin peak obtained by ammonium sulfate chromatography.
This indicates either that the chromophores responsible for these maxima are present in a fixed ratio in the native protein or that they are inseparable by this technique. Other experiments provide support for the first hypothesis: no spectral changes were observed when Chroomonas extracts were exarnined under nondenaturing conditions by centrifugation, gel filtration, or isoelectric focusing. This may indicate a very high association constant for the subunits. 0' hEocha el al. (32) have reported purification of phycocyanin from the Cryptomonad Hemiselmis vireocens, which is spectrally similar to the biliprotein in Chrocmonas. Phycocyanin from E-l. vireocens had an A645:A585 ratio of about 1.1 (compared to 1.05 for ours) but an AGbs:A 280 ratio of only 3.5 (calculated from Fig.  1, ltcf. 32). The ammonium sulfate gradient purification procedure yielded peak fractions with A 6d5:A280 ratios greater than 5.5, while more than half of the phycocyanin applied was recovered with a ratio greater than 5.0.
C-Phycocyanin possesses two positive CD bands, at 630 and 595 nm, while phycoerythrin has a negative band at 567 nm and a positive band at 548 nm (33, 34). The CD spectrum of phycocyanin exhibits a strong negative band at 650 nm, a strong positive at 588 nm, and a positive shoulder at 613 nm.
Both the CD and the absorption spectra indicate at least two different types of chromophorcs, or distinctly different environments for similar tetrapyrrole chromophores, as was found for C-phycocyanin and allophycocyanin (35). The molar ellipticity [01Z09 nm was calculated to be -20,200 deg cm2 per dmole. This corresponds t)o about 55% o( helis using poly(L-lysine) as a standard (36) and assuming that the chromophorc makes no contribution to t,hc dichroism at 205 nm. The behavior of phycocyanin in denaturing solvents differs from that of C-phycocyanin.
C-Phycocganin in sodium dodecyl sulfate gel electrophoresis yields two bands (22,(37)(38)(39); but the a-band of phycocyanin seems from our esperinients to have a significantly smaller molecular weight than the corrcsponding band in C-phycocyanin, whereas the @-bands nlay be similar in molecular weight. The monodisperse nature of phycocyanin is not a result of the purification procedures, since crude solutions csamined by ultracentrifugation with schlieren optics also showd the 4.4 $ boundary :ls the principal component.
In contrast, fresh estracsts of C-phycocyanin possess an even greater proportion of larger aggrcgatcs than does purified C:-phycocyanin (40). Jsoelcctric focusing in sucrose densit'y gradients2 revealed a complex pattern of bands which are isoelectric in the range of PI-I 4.2 to 6.9. 'Illis complex pattern is apparently not artil'aci.ual, for several of the phycocyanin zones have been re-electrofocused and found to return to their original isoelectric points.
Tor is this complex pattern due to photodegradation, as has been reported for phycoerythrin from another Cryptophycean alga (41). Other studies in this laboratory have indicat'rd that the degree of charge heterogeneity seen with phycocyanin is not observed with the C-phycocyanins, even though the latter have multiple aggregate forms. Chroomonas sp. is a eukaryotic and Uagellated organism representative of the Cryptophyta. The cxperimcnt,s presented here show that its chromoprotein is quite different, in terms of its properties and amino acid composition from the chromoprotein from the prokaryotic, blue-green algae. This supports the earlier observation (42) that the Cryptophyta have evolved to a significantly different phylogenetic position than the blue-green algae, a position more dist,ant from the prokaryotic than another eukaryote, the red algae.