Fluorescence of Histones H1 A TYROSINATE-LIKE FLUORESCENCE EMISSION IN CERATITIS CAPITATA H1 AT NEUTRAL pH VALUES*

We have examined the fluorescent properties of his- tones H1, and of some peptides derived from them, from calf thymus and from the fruit fly Ceratitis capitata. The fluorescent emission spectrum of folded histone H1 from C. capitata at neutral pH is characterized by a maximum at 303 nm and a shoulder at 340 nm. The overall quantum yields of fluorescence do not increase upon folding, although the fluorescence of the single tyrosyl residue of calf H1 is enhanced when the protein folds. As expected, the excitation maximum of calf H1 is shifted to longer wavelengths when the protein folds and its position does not depend upon the wavelength at which the fluorescence is observed. However, Ceratitis H1 exhibits two excitation maxima. The first cor- responds to emission at 303 nm and it is slightly red-shifted upon protein folding, whereas the second, which corresponds to emission at 340 nm, is displaced from 280 nm in the denatured protein to -285 nm in the folded histone. This suggests that the two tyrosyl residues of the insect histone behave as independent fluo- rophores. The shoulder at 340 nm does not appear at pH 2, even when the protein is folded. Titration to neutral pH values results in the appearance of the shoulder, the process being characterized by a p c = 3.7. The fluo- rescence spectrum of insect histone has been resolved into the contributions of the individual tyrosyl residues and the results suggest that the emission at

We have examined the fluorescent properties of histones H1, and of some peptides derived from them, from calf thymus and from the fruit fly Ceratitis capitata. The fluorescent emission spectrum of folded histone H1 from C. capitata at neutral pH is characterized by a maximum at 303 nm and a shoulder at 340 nm. The overall quantum yields of fluorescence do not increase upon folding, although the fluorescence of the single tyrosyl residue of calf H1 is enhanced when the protein folds. As expected, the excitation maximum of calf H1 is shifted to longer wavelengths when the protein folds and its position does not depend upon the wavelength at which the fluorescence is observed. However, Ceratitis H1 exhibits two excitation maxima. The first corresponds to emission at 303 nm and it is slightly redshifted upon protein folding, whereas the second, which corresponds to emission at 340 nm, is displaced from 280 nm in the denatured protein to -285 nm in the folded histone. This suggests that the two tyrosyl residues of the insect histone behave as independent fluorophores.
The shoulder at 340 nm does not appear at pH 2, even when the protein is folded. Titration to neutral pH values results in the appearance of the shoulder, the process being characterized by a p c = 3.7. The fluorescence spectrum of insect histone has been resolved into the contributions of the individual tyrosyl residues and the results suggest that the emission at 340 nm originates in a tyrosinate that may be formed in the excited state by proton transfer to the carboxylate anion of a glutamyl residue.
The results obtained from these experiments have also aided in resolving the n e a r -W circular dichroism spectrum of insect histone (Barbero, J. L., Franco Fluorescence spectroscopy is often a method of choice to investigate conformational transitions, especially in tryptophan-free proteins, which are usually referred to as class A proteins (1,2). Their fluorescence is entirely due to tyrosine, because the phenylalanine contribution is considered negligible (3). Early research (4) showed that in most class A proteins tyrosine emits with a higher quantum yield when the protein is denatured, but recent data suggest that in some proteins, including calf histone H1, the quantum yield of tyrosine fluorescence may be enhanced when the tyrosyl residue is transferred to a hydrophobic environment (5)(6)(7)(8). Apart from this anomaly, some examples have been described in which tyrosine emits at an unusually high wavelength, as opposed to the normal value of about 304 nm. They include cattle adrenodoxin (9-ll), parsley plastocyanin ( E ) , human serum albumin (13), and two cytotoxins isolated from Indian cobra (Naja nuja) venom (14). There is no agreement as to the cause of this abnormal emission, and Szabo et al. (14) have claimed that it may originate in the tyrosinate form of tyrosine, which would be produced in the excited state by a proton transfer to some adjacent acceptor group. This mechanism had been substantiated by Rayner et al. (15) in the case of free tyrosine in the presence of acetate anions.
In the present paper we describe some fluorescent properties of histone H1 from the fruit fly Ceratitis capitata. We have previously reported that this histone has two tyrosyl residues in the region containing the globular head (16). CD and difference spectroscopy were used to study the environments of these residues and it was concluded that one of them (tyrosine "1") is buried in the hydrophobic core in an environment that resembles that of the single tyrosyl residue of calf H1. The additional tyrosyl residue (tyrosine "2") seems to be more exposed to the solvent than tyrosine 1. The CD spectrum in the near-UV region of folded calf H1 is characterized by a minimum at 276 nm (16). However, it was not possible to decide whether the spectrum of Ceratitis H1, which exhibits a small maximum at 287 nm and a minimum at 266 nm (16), results from the summation of independent contributions from the tyrosyl residues or whether it is due to an interaction between both these residues. The results given in the present paper have aided in solution of this question and, on the other hand, they strongly suggest that a mechanism similar to that proposed by Rayner et al. (15)

MATERIALS AND METHODS
Isolation of Histones HI and Their Peptides-Calf thymus H1 was isolated by the method of Johns and Butler (17) and it was purified by CM-cellulose chromatography as described by Johns (18).
Pure Ceratitis H1 was isolated as described elsewhere (19) and it was purified by a modification (19) of the chromatographic method of Johns (18). The trypsin-resistant cores from both histones were prepared and purified by gel permeation chromatography as previously described (16). Ceratitis HI was cleaved at its single methionyl 315 residue by cyanogen bromide and the larger peptide, containing the COOH-terminal three-quarters of the molecule, was isolated by gel permeation chromatography in a column (2 X 100 cm) of Sephadex G-100 equilibrated and eluted with 0.01 M HCI. The purity of all the preparations was checked by polyacrylamide gel electrophoresis in the presence of 2.5 M urea (20). The presence of tryptophan-containing impurities was ruled out by examining the first derivative absorption spectrum (16): no peak nor shoulder at 290 nm was observed in any sample. Fig. 1 shows the electrophoretic pattern of histones HI and their TRCs.' Spectroscopic Methods-Absorption and CD spectra were recorded as previously described (16). Fluorescence measurements were carried out in a spectrofluorometer RRS-1000 (Schoeffel Instruments). Fluorescence was measured at right angles. The entrance slit of the excitation monochromator was 2.0 nm and the exit slit was adjusted to 1.0 nm. The analog electrical outputs from both the wavelength drive and the photomultiplier were digitized with an HP 43710A A/D converter and stored in a HP 9815A computer interfaced with a HP-IB 98135 unit. Spectra were recorded on a plotter (HP 7225A). To obtain emission spectra, the voltage of the emission photomultiplier (Q 4283B 520) was adjusted to 1060 V. The photometer sensitivity was adjusted to 30 or 100 nA, depending on the fluorescence of the sample. The scan speed was 100 nm/min.
To obtain corrected excitation spectra, the fluorescence was measured with a Q 42834 S-20 photomultiplier operating at 980 V. The fluorescence of a rhodamine B quantum counter was used as reference signal. Rhodamine fluorescence was measured with an IP-28 photomultiplier operating at 500 V. The signals from both photomultipliers were processed in the spectrofluorometer ratio computing circuitry, which gives an analog signal that equals 10A/B (A = output of Q 42834 S-20; B = output of IP-28). This signal was digitized and the excitation spectrum was plotted as described above, except that the scan speed was 20 nm/min. The photometer sensitivities were adjusted to 30 nA (Q 42834 S-20) and 100 d (IP-28).
Quantum yields of fluorescence were determined using as standard L-tyrosine. In order to avoid artifacts due to quenching by external ions, the histones were dissolved in water; if required, NaCl was added to obtain the desired concentration and finally the calculated amount of HCI or NaOH to bring the pH to the desired value was added. The actual pH was then determined and a solution of L-tyrosine in the same solvent was carefully titrated to the same pH. To determine the quantum yields, the following relation was applied:

h(S)/h(T) = Ah(T).Fh(S)/Ah(S).Fh(T)
where and @ A ( T ) are the quantum yields of the sample and of L-tyrosine, respectively, for excitation at a wavelength A. Fh(S) and R ( 7 ' ) stand for the total fluorescence when exciting at X; they were determined from the area under the emission spectra and were expressed in arbitrary units. Finally, AA(S) and AA(T) are the absorbances at a wavelength h of the sample and L-tyrosine, respectively.

RESULTS
The fluorescence emission spectra of calf thymus and C. capitata H1 in the folded state are shown in Fig. 2. I t is apparent that the quantum yield of Ceratitis H1 emission is about 30% of the corresponding value for calf H1 (see also Table I). Both spectra show a maximum at 303 nm, characteristic of tyrosine emission (4), but the spectrum of Ceratitis H1 shows a shoulder at 340 nm (see the inset of Fig. 2), the significance of which will be discussed later.
When the emission spectra were recorded under denaturing conditions (H20, pH < 7), the curves of both histones were very similar and the shoulder a t 340 nm was no longer noticeable in the spectrum of Ceratitis H1 (see also Fig. 5). Table I shows that the quantum yield of the emission of both histones in the denatured state is also very similar.
The quantum yield of calf H1 emission rises in going from the denatured state to a partially folded one (0.3 M NaCl, pH 6.3), in accordance with the results of other authors (7,8,21). Ceratitis H1 behaves in a different way in two respects.   the overall quantum yield does not vary in the transition from a disordered conformation (H20, pH 5.8) to a partially folded state (0.3 M NaCl) and it even diminishes in a late stage of folding (1.0 M NaCl, pH 6.4), as Table I shows. Second, the shape of the spectrum changes as the protein folds, and the more the folding, the more noticeable the shoulder at 340 nm.
As pointed out before, the overall quantum yield of fluorescence does not increase, so that the appearance of the shoulder is obviously accompanied by a fall in the fluorescence intensity a t 303 nm.
Conformational changes of a class A protein do not affect the position of the tyrosine fluorescence spectrum (4), SO that the appearance of the shoulder is not likely to result from a red shift in tyrosine emission induced by folding. In order to determine whether the emission at 340 nm was associated with a change in the absorption of Ceratitis H1, we have recorded the excitation spectra of calf thymus and insect histones, observing the fluorescence at two wavelengths, 300 and 350 nm. The position of the excitation maxima is given in Table 11. It is clear from the values given that the excitation spectrum of calf H1 is red-shifted as the protein folds. This

Ceratitis HI
6.0 Folded 0.04 a The degree of folding was estimated by CD and expressed as the percentage of cy-helix relative to the folded state (see Barber0 et al. (16)). Note that NaCl is less efficient than KF in folding histones HI.

TABLE I1
Excitation maxima of histones H1 Under the experimental conditions used, the position of the excitation maxima is affected by an error of f0.5 nm when the observation is carried out at 300 nm and by an error of *I nm when the emission is observed at 350 nm. finding is in accordance with the previously reported red shift in the absorption spectrum (16,21). The position of the excitation maximum did not vary with the wavelength at which the fluorescence emission was observed, although, as a consequence of the lower fluorescence intensity at 350 nm, the excitation spectrum was very flat when the emission was observed at this wavelength. The behavior of Ceratitis H1 is different. As shown in Fig. 3 when the emission was observed a t 300 nm, a slight red shift was detected in the excitation spectrum as the protein folds, but if the emission was observed at 350 nm, folding resulted in a large displacement of the excitation spectrum. This finding supports the idea that the presence of a maximum and a shoulder in the fluorescence of Ceratitis H1 is due to the presence of two fluorophores that behave very distinctly in the folded conformation of the histone.
We have also studied the influence of pH on the fluorescence of Ceratitis HI. It has been mentioned that the peculiarities of the fluorescence of insect H1 are only apparent in the folded state so that in order to avoid artifacts due to a change of conformation, the influence of pH must be studied without affecting the degree of folding. This could be accomplished in the presence of 1.2 M NaC1. Actually, CD measurements revealed that no change in the secondary structure of the histone was noticeable between pH 1.2 and 10 in the presence of 1.2 M NaCl. However, the fluorescence of the histone was markedly affected by the variation in pH. Fig. 4 shows that the fluorescence intensity a t 303 nm shows an abrupt fall as the pH rises, albeit the fluorescence intensity at 340 nm is scarcely influenced by a change of pH between 2 and 6. It is also interesting to note that the fluorescence emission spectrum of folded Ceratitis H1 at pH 2.0 shows no shoulder at 340 nm, suggesting that, although the protein is folded under these conditions (see above), both tyrosyl residues fluoresce at the same wavelength. The overall quantum yield is still lower than that of folded calf H1 (Table I), but this may be easily explained taking into account that tyrosine 2 is more exposed to the solvent than tyrosine 1 (16) and, therefore, its emission would be more quenched.
Assuming that the tertiary structure of Ceratitis H1 remains unaltered during titration between pH 2 and 6 in the presence of 1.2 M NaC1, as the secondary structure does, it is possible to conclude that the abnormal emission of Ceratitis H1 at 340 nm appears only when a functional group of the protein with pKh z 3.7 (Fig. 4) has been titrated.
The spectra depicted in Fig. 5 show the effects of folding on the fluorescence of Ceratitis H1 at pH 6.4. It is clear from this figure that folding results in a decrease of the overall quantum yield of fluorescence as previously mentioned, but the quantitative comparison of both spectra, as shown in the inset, conveys the most prominent result of this experiment. The curve in the inset was obtained by subtracting one-half of the fluorescence of the denatured histone from the fluorescence of the folded molecule. The reason for this subtraction relates to the fact that if both tyrosines were equally exposed to the solvent in the denatured histone, each one would contribute to 50% of the total fluorescence of the protein. Thus, assuming that the fluorescence of tyrosine 1 undergoes little or no change during folding (an assumption that will be explained under "Discussion"), the curve shown in the inset of Fig. 5 would roughly represent the fluorescence spectrum of tyrosine 2 in the folded histone. This curve practically superimposes with the fluorescence spectrum of tyrosinate.
It seems clear from the above results that for the abnormal emission of Ceratitis H1 to appear, two requirements have to be fulfilled, i.e. the histone has to be folded, and the pH must be above a critical value. That folding is essential for the appearance of the fluorescence shoulder at 340 nm was also verified by dissolving Ceratitis H1 in 0.5% (w/v) sodium dodecyl sulfate at neutral pH. Under such conditions, the fluorescence spectrum of Ceratitis H1 exhibited a single, symmetrical peak, centered at 303 nm.
The near-UV CD spectrum of Ceratitis H1 in 0.05% sodium dodecyl sulfate was also recorded. The positive peak that characterized the CD spectrum of folded H1 at neutral pH values (16) did not appear in the presence of the detergent, and the shape of the spectrum was like that of calf thymus H1, although the ellipticity was doubled. This result posed the question as to whether the appearance of the maximum at 287 nm and the distorted minimum at 266 nm characteristic of the CD spectrum of Ceratitis H1 are due to the same causes as the abnormal fluorescence of the fly histone. For practical reasons, this question could not be answered using intact H1, but it was solved by studying the fluorescence properties of histone fragments.
Fluorescence and CD of HI Peptides-Experiments similar to those described in the preceding section were also carried out with the TRCs of calf and fly histones and with peptide PLCNB~ from Ceratitis H1. In general, TRCs behave like their parent molecules and, to avoid reiteration, in the present section we mainly refer to the peculiarities of the TRCs.
It can be seen in both Fig. 2 and Table I that the quantum yield of fluorescence of TRC from calf thymus does not increase upon folding in so noticeable a manner as that of intact H1. The reasons for this behavior will be given under "Discussion." The behavior of the TRC of Ceratitis H1 is similar to that of the parent molecule and the shoulder at 340 nm appears in the fluorescence spectrum of the folded TRC at neutral pH values.
:50k W , I  The CD spectra in Fig. 6 proved that the shape of the CD spectrum of folded TRC is actually pH-dependent. Both spectra in Fig. 6 were recorded in the presence of 1.2 M NaC1, which, as pointed out before for intact H1, prevents changes in the secondary structure of TRC in going from pH 1.7 to 10.
Finally, the fluorescence and CD of peptide PLCNB~ from Ceratitis H1 were studied. Methionine is the 13th residue in TRC, so that the NH2-terminal segment of the H1 molecule and the 13 fust residues of TRC are not present in PLCNB~ (22). The fluorescence spectra of PLCNB~ at neutral pH did not show the tyrosinate-like emission at 340 nm, irrespective of the degree of folding. The CD spectrum of folded PLCNB~ in the near-UV region is similar to that of calf H1, i.e. there is a single negative band with a minimum at 276 nm. Thus, the presence in Ceratitis H1 of the CD maximum at 287 nm seems to be associated with the tyrosinate-like emission and any treatment of the histone which causes the former to vanish also leads to the disappearance of the latter.

DISCUSSION
The fluorescence emission spectrum of folded Ceratitis H1 is characterized by the presence of a maximum at 303 nm and a shoulder at 340 nm. It is obvious that the two tyrosyl residues of the histone molecule behave as different fluorophores. Fig. 3 and Table I1 clearly show that the fluorescence at 340 nm arises from a tyrosyl residue whose absorption maximum, which has to coincide with the excitation maximum, is placed at -284 nm, whereas the emission at 303 nm is due to the tyrosyl residue that absorbs at 279 nm. The possibility that a tryptophan-containing impurity would be responsible for the emission at 340 nm can be ruled out not only in view of the purity of both H1 and its TRC, but also by the excitation spectra themselves.
The excitation spectra were recorded by observing the fluorescence emission at 300 and 350 nm rather than 303 and 340 nm in order to minimize the cross-influence of both emitting species. The single tyrosyl residue of calf H1 emits at 303 nm and, due to the similarities between this residue and tyrosine 1 of Ceratitis H1 ( X ) , we assume that tyrosine 1 is responsible for the emission of Ceratitis H1 at 303 nm. According to the results of Table 11, it is also possible to conclude that tyrosine 1 moves to a hydrophobic environment as the protein folds, but the resulting red shift is not so large as that displayed by the tyrosine of calf H1. In a previous paper (16) we showed that the difference spectra between folded and unfolded histones H1 of both calf and Ceratitis are very similar and we suggested that the tyrosine 1 of Ceratitis histone is entirely responsible for the observed red shift. Nevertheless, in the present paper, taking advantage of the fluorescence properties of Ceratitis H1 that allow us to distinguish between the two tyrosines, we have shown that, upon folding, the absorption of both tyrosyl residues is red-shifted. However, the red shift of tyrosine 2 cannot be attributable to a change to a hydrophobic environment because: (i) a red shift of -6 nm seems to be too large to be ascribed to the transfer of a tyrosyl residue to a nonpolar environment; (ii) the fluorescence emission of tyrosine is not red-shifted when the amino acid moves to a hydrophobic environment (4) and the emission of tyrosine 2 goes from 303 nm in the denatured protein to 340 nm in the folded state. Thus, the possibility that the red shift of tyrosine 2 is due to the transfer of the aromatic ring to a nonpolar environment can be ruled out and we have to conclude that the red shift is due to an interaction of tyrosine 2 in its ground state with some group of the protein that is moved to the vicinity of the tyrosyl residue when the protein folds. The interaction of tyrosine with phosphate ion results in both hyperchromicity and a displacement of the spectrum about 1 nm to longer wavelength (23). In the present instance, the observed red shift is larger, but the tightness imposed by a rigid tertiary structure may account for this difference.
We can assume that the curve in the inset of Fig. 5 represents the emission spectrum of tyrosine 2 in folded Ceratitis H1. To construct this curve, we have supossed that the fluorescence of tyrosine 1 undergoes little or no change during folding. Although this assumption cannot be actually proved, it should be remembered that the increase in the quantum yield of fluorescence of a tyrosyl residue upon protein folding seems to be associated with its transfer to a highly hydrophobic, compact environment. This has been shown in Calf H1 ( 7 , 8 ) where the folding is accompanied by a red shift of 3 nm in the absorption maximum of tyrosine (Table 11). The excitation spectrum of tyrosine 1 of Ceratitis H1 is displaced only 1 nm to the longer wavelength when the protein folds and, therefore, a large increase in its quantum yield of fluorescence does not seem probable.
If the fluorescence at 340 nm may be attributable to tyrosinate, apart from the proposed interaction of tyrosine 2 in the ground state with a functional group of the protein, there would also be an interaction capable of removing the phenolic proton from the excited state of tyrosine 2. A carboxylate anion would play both roles, as predicted in the following straightforward scheme:

Ground state c " -----
The excitation step (I) would be red-shifted due to the interaction with the carboxylate anion and the fluorescent emission (11) would be mainly from the tyrosinate form. It is interesting to note that a glutamic acid is the 12th residue of the TRC of Ceratitis H1 (22) and, therefore, this glutamyl residue is not present in peptides PLcNB~, which do not show the tyrosinatelike emission. The carboxyl group of this particular glutamyl residue may, therefore, be the group responsible for the abnormal behavior of tyrosine 2. The pK: value of tyrosine 2 in the ground state is 9.8 (16). If the pKb value estimated from the inset on Fig. 4 corresponds to the acid-base equilibrium of excited tyrosine 2, ApK (i.e. PKexcited -pKground) would be -6.1. It is interesting to note that Rayner et al. (15) have calculated that, in the presence of acetate buffer, ApK for free tyrosine is -6.1. In spite of this remarkable coincidence, the pKh value of 3.7 (Fig. 4) may also correspond to the carboxyl group of glutamic acid. In this instance, for tyrosine 2 ApK S -6.1, but a pK as large as -6.7 has been calculated for 2naphthol (24).
The single tyrosyl residue of bovine adrenodoxin, which also emits at an abnormally high wavelength (9), seems to be strongly hydrogen-bonded to a glutamyl residue (25). The interaction with carboxylate anions seems to be a neccesary condition for the appearance of the abnormal emission of tyrosine (Refs. 11, 14, 15, and present results). However, it is not a sufficient one, for the single tyrosyl residue of vitamin D-dependent calcium binding protein from bovine intestine is hydrogen-bonded to glutamic acid (26) and yet the fluorescence of this protein does not show any abnormality (27). Szabo et al. (14) have claimed that this abnormal emission of tyrosine is due to tyrosinate and Graziani et al. (12) have also suggested this possibility. Our present results may also be interpreted in these terms. However, some other observations (11, 25) are less compatible with the idea that tyrosinate is the species that gives rise to the abnormal tyrosine fluorescence, although they do not rule out this possibility.
On the other hand, the results presented in this paper strongly suggest that both the abnormal fluorescent emission and the positive CD peak of Ceratitis H1 have a common cause. Both the tyrosines of Ceratitis H1 are independent fluorophores and it may be supposed that the CD spectrum of the histone also results from the summation of their individual contributions. Taking into account that the positive peak does not appear with folded TRC at pH 2.2 (Fig. 6), the CD spectrum of tyrosine 2 can be obtained by subtracting the curve at pH 2.2 from that at pH 6.5. The resulting curve (not shown) exhibits a positive peak at 284 nm, i.e. the wavelength at which tyrosine 2 absorbs when the protein is folded.
Finally, we wish to comment on the differences found between the quantum yields of fluorescence of calf H1 and its TRC. It is obvious from Table I that the emission of the tyrosyl residue of calf H l TRC is scarcely enhanced when the peptide folds, while the quantum yield of fluorescence of H1 increase -2.7 times in going from the denatured state to a folded conformation. We have recently found that the NH2terminal segment and the COOH-terminal region of histone H1 have an influence on the tertiary structure of the globular head of the molecule, so that removal of both the COOH-and NH2-terminal regions by limited tryptic digestion results in a relaxation of the globular head (28). In view of the above arguments, the transfer of the tyrosyl residue to an environment less compact than in the intact molecule results in only a slight increase of the quantum yield of fluorescence.