- Email: [email protected]
Proton nuclear magnetic resonance study of the histidine residues of pituitary bovine growth hormone
Biochimica et BiophysicaActa, 994(198911166-171
166
Elsevier
BBA33290
Proton nuclear magnetic resonance study of the histidine residues of pituitary bovine growth hormone Nell E. M a c K e n z i e 1, S,M. Plaisted 2 a n d D . N . Brems 2,. ! Departmentof PharmaceuticalSciences, Universityof Arizona, Tucson,AZ and 2 The Upjohn Company, Kalamazoo, MI (U.S.A.)
(Received26 September1988) Keywords: Pituitarybovinegrowthhormone;NMR, t H-; Histidineresidues The pit'.value of hisfldlues 20, 22 and 170 of pituitary bovine growth hormone (pbGH) and their C2H deuterium exchange rates have been determined by high.resolution I H.NMR spectroscopy. Partial assignment of these pm'ameters imllcate that His-170 is a surface residue with a half-life for deuterium exchange at the C2H position of approx. 36 h and a ! ~ , value of 5.68 4. 0,11. A reveralble conformaflonal change of pbGH has also been characterized in terms of the invelvmtent of either hisfldiues 20 or 22. At or near physiological pH values, one of these residues is expelled from a buried, h y d e position to become fully solvent-exposed. The d|spori~ in pK. values between hisfi~nes 20 and 22 (4.67 4. 0.20 and 5.94 4. 0.10, in some combination) has been shown by computer modeling, to be compatible with these residues residing in an a-helical region of the protein. InUmlmtion Bovine growth hormone (bGH) is a pituitary protein consisting of 191 amino acid residues [1]. It is globular in conformation but a high-resolution X-ray crystallographic analysis of the three-dimensional structure has not been accomplished [2]. Recently however, the X-ray structure, at 1.9 A resolution, of the highly homologous protein [3], porcine growth hormone, has been reported [4] to consist of a four-helix bundle with the helices arranged in an antiparallel fashion. High-resolution nuclear magnetic spectroscopy is capable of providing detailed information on protein structure in solution, but a prerequisite for this is the assignment of resonances in the spectrum to individual amino acid residues in the sequence. For proteins the size of bGH (approx. 22 kDa) resolution of the NMR spectrum even at the highest available field strength is limited and hence the sophisticated two-dimensional methodologies developed for the structural elucidation of small proreins [5] cannot be readily applied. The most common exceptions to the lack of resolution in the IH-NMR spectrum of a protein are the * Present address: Lilly Research Laborat~ry, Hi Lilly and Company, Indianapolis,IN 46285,U.S.A. Abbreviations: pbGH, pituitary bovine growth hormone; NMR, nuclear magneticresonance; GdnHCI,guaninehydrochloride; SDS; sodiumdodecylsulfate,
downfield resonances attributable to the C2H imidazole ring protons of histidine residues. Few protein-bound protons have been more useful as 1H-NMR spectroscopic probes of intrinsic tertiary solution structure [6]. The associated resonances are sensitive to the microenvironments of the histidine residues, and any spectral changes can be attributed to discrete molecular perturbations of the folded protein. Generally, the microenvironment of individ,,~al histkiines within a protein can be assessed by comparison of their protonation parameters of pK. and deuterium exchange rates. This paper describes 1H-NMR studies of histidines 20, 22 and 170 of pituitary bovine growth hormone (pbOH) wherein their protonation parameters are characterized and partially assigned.
Materials and Methods Materials
Pituitary-derived b o l l (pbGH) was obtained from A.F. Parlow, lot 7899C, the Harbour Medical Center of the University of California at Los Angeles. Deuterium oxide (2H20, 99.9 atom% enrichment) and 2Hs-glycine (98 atoms enrichment) were obtained from Cambridge Isotope laboratories. Guanidine hydrochloride (GdnHCI) was ultra pure grade from Schwarz/Mann while all other reagents were of analytical grade. Sample preparation
Correspondence: N.E. MacKenzie, Department of Pharmaceutical Sciences,Universityof Arizona,Tucson,A285721,U.S.A.
Protein samples for 1H-NMR spectroscopy were prepared by dissolving pbGH (20 mg) in 10 mM 2H 5-
0167-4838/89/$03.50 O 1989 Elsevier Science Publishers B.V. (Biomedical Division)
167 glycine in 2H20 (1 ml), pH 9.5 at room temperature. The pH meter readings were not corrected for isotope effects on the glass electrode. Under these conditions the readily exchangeable protons (amino, amido, hydroxyl and carboxyl) were fully exchanged with deuterium within a few minutes *. The solution pH was adjusted with microliter quantities of dilute 2HCI or N a O ' H and pH readings were taken before and after data acquisition. The spectrum was accepted only if the deviation of the pH was less than 0.04 units. Protein samples were incubated for at least 30 rain at 25°C before NMR analysis.
NMR spectroscopy t H spectra were recorded at 25°C on a Varian XL400 spectrometer operating at 399.96 MHz for t H nuclei in the Fourier-transform mode. A total of 500 transients obtained by quandrature detection for a spectral width of 6 kHz by using 15 000 data points and an acquisition time of 1.334 s were averaged for each spectrum. Unless stated otherwise, the free induction decay was multiplied by an exponential factor equivalent to a line broadening of 2 Hz prior to Fourier transformation. Chemical shifts (8, ppm) are expressed relative to internal dioxane (~ = 3.740 ppm) downfield from 4,4-dimethyl-4-silapentane-l-sulfonate. Residual HO2H was suppressed by presaturation (1 s) prior to data acquisition.
Histidine zH exchange pbGH (80 mg) was dissolved in 10 mM 2Hs-glycine in 2H20 (4 ml) at pH 9.50. The solution was incubated at 37 °C in a sealed vial. Aliquots (1 ml) were taken at 0, 16, 42 and 67 h and immediately frozen in liquid nitrogen. The sample was stored at - 4 ° C until NMR analysis was performed.
Preparation and isolation of zH-fragments of pbGH A portion (1 ml) of deuterated pbGH (20 mg) was diluted by the addition of 96% formic acid (2 ml). Cyanogen bromdie (40 mg) was added to this solution, and the reaction mixture was left to stand at room temperature in a sealed vial for 24 h. The reaction mixture was then diluted with water (45 ml) and lyophilized. The lyophilate was dissolved in 40% formic acid (2 ml) and chromatographed on a Sephadex G-50 column (1 × 50 cm) using 20% formic acid as the mobile phase. Elution was monitored by uRra-violet detection at 280 nm. The fractions containing the AB fragment
* The average expected time constant for amide proton exchange of unfolded b o l l at 2 0 ° C and pH 9.5 is approx. 0.1 ms [7]. Since within 100 s all the amide protons had exchanged, the degree of protection from exchange by the folded protein must be less than 100000-fold.
B
H3,. H1
--Y
AA H2,H3 H1-A
8'.5
81o
!
|
T.o
Bio PPr
Fig. 1. Aromatic region of the 400 MHz I H-NMR spectrum of pituitary bovine growth hormone (A) pH 1.91, (B) pH 9.73. Insets: Line-narrowed spectra of the histidine C2H resonances, H1, H2 and H3.
(residues 6-124 linked to 150-179) were collected, diluted 4-fold and lyophilized. The peptides were reduced for 1 h with 8 M GdnHC1 (1.2 ml) containing 0.1 M ammonium bicarbonate and dithiothreitol (2 mg/ml) and rechromatographed on the same column as above after the addition of 0.8 ml 96% formic acid. Fraction A consisting of residues 6-124 and fraction B consisting of residues 150-179 were collected and lyophilized. Results
The ~H-NMR spectrum of pbGH is typical for proteins in the molecular weight range of 20 kDa, consisting of broad regions of ill-resolved resonances. The aromatic region of tile 1H-NMR spectrum of pbGH at pH 1.91 and 9.73 is shown in Fig. 1A and B, respectively. The resonances X and Y (Fig. 1A) are assignable to the 2',6' and the 3',5' aromatic protons, respectively, of the six tyrosine residues of the protein. The average chemical shifts of these resonances are 7.059 ppm and 6.764 ppm and are almost identical to those for tyrosine in small peptides [8]. Similarly, the resonances H1, H2 and H3 can be assigned to the C2H imidazole ring protons of the three histidine residues of the pbGH which occur at positions 20, 22 and 170 in the amino acid sequence. The chemical shifts of 8.659 ppm, 8.636 ppm and 8.620 ppm, are consistent with those of fully protonated histidine reisdues both in the isolated amino acid [8] and small peptides [6]. The small differences in chemical shifts are due to amino acid primary sequence variations in the region of the histine residues. Based on
168 TABLE I
8.61
Chemical shifts and pKa values
8.4E 8.3~
Chemical shifts and pK a values derived for the histidine C2H resonances of pituitary bovine growth hormone by non-finear least-squares analysis of the titration data in Fig. 2. Figures in parentheses are the calculated standard errors.
8.23 ~8.1C 7.84
O (acidic)
7.71 758 74e
HI
H2
H3
8.629 (0.015)
8.616 (0.013)
8.679 (0.027)
8 (basic) I
2
~
I
3
4
~
~
~
8
I
9
I
10
a chemical shift argument, pbGH assumes an essentially random coil configuration at this pH. Bovine ~ , : t l i hormone is known to undergo a reversible~pH-induced conformational change [9] and indeed the 1H-NMR spectrum of the protein at pH 9.73 (Fig. 1B) is that of a specifically folded globular structure. The perturbations of the aromatic resonances seen in Fig. 1B can be ascribed to a combination of factors including direct ionization and proximal electrostatic interactions together with the environmental changes of various residues within the folded protein. All these factors presumably contribute to the upfield shift of resonances H1, H2 and H3 with the most significant effect on chemical shift arising from deprotonation of the imidazole rings. The chemical shifts of the histidine C2H resonances H1, H2 and H3 of pbGH, plotted against pH at 25 °C are shown in Fig. 2. Titration data for the bJstidines represented by H1, H2 and H3 were analyzed by a three.parameter fit to the Henderson-Hasselbach equation (Eqn. 1) using a non-linear regression program for each ionization event. S H ÷ - SIlo =
7.670
7.561
(0.008)
(0.011)
OA
1.077
1.046
1.118
pK a
5.944 (0.101)
5.680 (0.109)
4.674 (0.125)
11
Fig 2, Variation in chemical shirt (O) with pH for the resonances HI (o), H2 03) and H3 (z) of pituitary bovine growth hormone shown in Fig. 1. The solid lines were calculated as described in the text and the fitted parameters are given in Table I.
81l ÷ - ~
7.552
(0.008)
I
couldonly be followed over a lirc,;te~i'pH range as at values between approx. 6 to 8, this resonance could not be resolved from the envelope of aromatic resonances. Above the latter pH value, resonance H3 reappears downfield of this envelope and co-titrates with resonance H2 (Fig. 2).
Histidine ZH.exchange kinetics The time-course of deuterium incorporation into the C2H position of histidines 20, 22 and 170 is shown in Fig. 3. Resonances a, b and c (Fig. 3) are, in some combination, the I H-NMR resonances of the C2H pro-
Hr)
go
g,
+[H ÷ ]
(1)
The observed chemical shift at a given pH is taken as 8ob~, &H+ and 8H0 are the" chemical shifts of the protonated and unprotonated forms of the histidine residue, respectively, while K a is the acidic dissociation constant. The pK a values derived for the histidine residues of pbGH are summarized in Table I. The best fit to the pH titration data is represented by the solid lines for H1, H2 and H3 (Fig. 2). In general, resonances H1 and H2 follow a simple signoidal curve indicative of a single ionization. The titration of resonance H3 occurs over a relatively small pH range (about 2 pH units) and suggests that this protonation - deprotonation event is a cooperative process. Resonance H3
8. o
!
8. 5
8. o
8. 5
' PPM 8.60
Fig. 3. Deuterium exchange of the imidazole C2 protons of pituitary bovine growth hormone. The histidine C2H region (resonances a, b and c) of a series of 400 MHz I H-NMR spectra of pituitary bovine growth hormone preparations containing 2~$ (w/w) SDS. The protein had been deuterated at 37°C (pH 9.50) in 10 mmol 2Hs-glycine (in 2H20) for the time (in hours) indicated prior to data accumulation.
169 0.05~__________________~ ,~
TABLE II First-order rate constants
First-order rate constants for deuterium exchange of the C2 protons of the histidine residues of pituitary bovine growth hormone at 37 o C (pH 9.5) derived by least-squares analysis of the data shown in Fig. 4.
-o.~.
-~
~ H1 H2 H3
-0.45 ,
-o~
,
2'o
il
|
4o Time
e'o
k (s- i ) X 10 6
(Correlation coefficient)
k/k
1.17 0.96 1.91
(0.98) (0.99) (0.99)
0.61 0.50 1
fastest
|
(h)
Fig. 4. Deuterium exchange of the imidazole C2 protons of pituitary bovine growth hormone. The NMR srectra shown in Fig. 3 were integrated and plotted, as the fraction of unexchanged resonance, on a logarithmic scale versus time (in hours l to enable the rate constant (Table If) to be determined by linear-regression analysis.
tons of the histidines and differentially diminish with time. The spectra were recorded at pH 6 after the addition of 2~ (w/w) sodium dodecyl sulfate (SDS) to the deuterated protein preparations (see Materials and Meihods). Under these defL~-d conditions the three resonances were baseline-resolved and had identical line wid,:hs. This facilitated the accurate integration of the
area of the resonances. A plot of the C2H resonance area as a function of time is shown in Fig. 4. A non-exchangeable aromatic resonance was used as an integration reference and the data were fit by linear-regression analysis to the expression: log (fraction unexchanged) = - kt. The derived values for k, the first-order rate constant, are given in Table II.
Partial assignment of rate constants The 1H-NMR spectrum of the histidine C2H resonances of pbGH which had been deuterated for 42 h is
1.5
b AB C
0.75
t-
.o
B
~x 1.5 0 0'1
<
0.75
--I
8.85
i
8.80
I
8.75
i
8.70
l
8.65
w
8.60
I
8.55 PPIVi
Fig. 5. (A) The histidine C2H region of the 400 MHz 1H-NMR spectrum of pituitary bovine growth hormone after the addition of 2% (w/w) SDS. The protein had been deuterated at 37 o C (pH 9.50) in 10 mmoi 2Hs-glycine (in 2H20) for 42 h prior to data acquisition. (B) The histidine C2H region of the 400 MHz 1H-NMR spectrum of deuterated peptide, fragment B, (residues 6-124) derived from the protein sample described in (A). (C) As for (B) but of the deuterated peptide, fragment A (residues 150-179).
el 0
I
I
l
9 18 27 Elution volume (ml)
36
Fig. 6. Gel filtration chromatography of deuterated pituitary bovine growth hormone "fter cleavage with cyanogen bromide. The top chromatogram represents the elution profile of the complete cyanogen bromide reaction. Fractions corresponding to the fragment AB (residues 6-124 linked to 150-179) were collected and the disulfide was reduced (see Materials and Methods). The bottom chromatogram represents the elution profile of the reduced fragmem~ A (residues 6-124) and B (residues 1b0-179).
170 shown in Fig. 5A. This protein was then cleaved by reaction with cyanogen bromide. The elution profile of the complete reaction mixture from gel filtration chromatography is shown in the upper trace of Fig. 6. The peak labeled AB corresponds to the peptide fragment composed of residues 6-124 and 150-179 finked by a disulfide bridge btween Cys-53 and Cys-164 [11]. The AB fraction was collected and reduced with dithiothreitol. The chromatogram of reduced peptide fragments A (residues 6-124) and B (residues 150-179) is shown in the lower trace of Fig. 6. The 1H-NMR spectrum of the histidine C2H resonances of fragment A and B are shown in Fig. 5B and C, resp~tive!y. The C 2 E resonances of histidines 20 and 22 occur (in some combination) at 8.783 and 8.701 ppm (Fig. 5B) and that of His-170 (Fig. 5C) at 8.616 ppm. These can be compared to the chemical shifts of the C2H resonances in the deuterated intact protein (Fig. 5A) at 8.774, 8.704 and 8.631 ppm under identical solution conditions. From comparison of the chemical shift data and the relative integrated areas of the resonances in Fig. 5, His-170 corresponds to resonance C, which has been shown to be associated with the fastest deuterium exchange rate (Table II).
Partial assignment of pKa values To assign a pK, value to His-170, the C2H resonance C (Figs. 3, 5A and 7) was compared to H1, H2 or H3 (Figs. 1A, B and 7A) under identical solution condib
i42
o'l
H3
alb alb
'
8.4
'
8:2
'
'
S,O
'
"
PPM
Fig. 7. Inset: As described for Fig. 5A after 42 h of deuteration (A) The histidine C2H region of the 400 MHz IH.NMR spectrum of the deuterated protein described for the insert above but accumulated in the absence of SDS. (B) The hisfidine C2H region of the 400 MHz SH-NMR spectrum of pituitary bovine growth hormone prior to C2H deuterb~un exchange.
tions. In the absence of SDS the histidine C2H resonances of pbGH, H1, H2 and H3 are only sufficiently resolved, to allow accurate integration, over the small pH range of approx. 4.80 to 5.20. Fig. 7B shows the ~H-NMR spectrum of the histidine C2H resonances of pbGH at pH 4.90. This can be compared with the resonances of deuterated pbGH recorded under the same conditions (Fig. 7A). Comparison of the relative integrated areas of the resonances in Fig. 7A shows that the central resonance, corresponding to H2, has lost the most area and can be assigned to resonance C, of Fig. 7 (inset). The peak assignments are therefore as follows: H1, H3 -- C2H of histidines 20 and 22 (in some combination), H2 - C2H of histidine 170. Diseussion The linewidth of resonances H1, H2 and H3 in the random coil configuration (Fig. 1A) at pH 1.91 is 7 Hz. This increases to only 9 Hz in the folded protein at pH 9.73. This indicates that pbGH has a substantially mobile or flexible structure in the latter pH range. This is corroborated by the fast deuterium exchange of all amide protons at this pH and the fast deuterium exchange rates of the histidine C2H protons (Table If). The half-life of the deuterium exchange of His-170 is approx. 36 h and can be compared with the rate of exchange of a surface residue [11] or those histidine residues known to reside in very mobile protein regions [12,131. The pKa values of histidines 20, 22 and 170 (Table I) are all lower than the pK a value (6.65, 21 ° C) of Nacetylhistidinemethylamide, the standard model for exposed histidine residues free from specific perturbation [14,15]. An environmental factor that can cause the lowering of a histidine pK a is the proximity of a positive charge to the imidazole ring [6,14]. Thus the ion-pair interaction of the histidine residues H1 and H2 of pbGH with a prctonated basic residue may account for their low pK, values. In addition, histidine residues which are buried in the protein in the unprotonated state have been assigned very low pK a values [13,16,1'7]. It is suggested here that the very iow pKa value of resonance H3 (histidine 20 or 22) is due to its burial within the protein. The resonance H3 is significantly shifted upfield in the unprotonated state with the lowest measurable chemical shift value of 7.533 ppm (Fig. 2). At higher pH values, this resonance moves further upfield and cannot be resolved from the envelope of aromatic resonances. This unusually low chemical shift value for this resonance presumably is a consequence, in part, of ring current effects and so in the buried (unprotonated) state this histidine residue is closely associated with an aromatic residue(s). Similarly, resonance H1 (Fig. 2, Table I) is shifted upfield in the deprotonated state at all alkaline pH values and again this arioes from
171
the proximity of the associated histidine residue to an aromatic residue(s). At pH values greater than approx. 8, resonance H3 reappears at a chemical shift coincident with that of resonance H2 (Fig. 2). It is suggested that a conformational change of pbGH occurs at or near physiological pH values causing expulsion of this histidine residue and exposure to solvent. This is supported by the now 'normal' chemical shift value for a deprotonated histidine C2H resonance. The large difference in the pK a values of histidines 20 and 22 (H1 and H3, Table II) indicates that their microenvironments are very different. Model studies show that for two histidine residues separated by a single amino acid residue in a B-pleated sheet, the separat,on between the two imidazole rings will be within the range 2 to 8 ,~. This is close enough to allow the two imidazole rings to interact and compete for the same proton. For similarly situated residues in an ~x-helix the imidazole rings lie on opposite sides of the helix, far enough apart ( > 12 A) that they may reside in sufficiently diverse environments to accommodate the pK a values determined in this study. Based on secondary structure predictions. [18] the segments 10-34, 66-87, 111-127 and 186-191 should be helical [19]. The X-ray diffraction studies of poricne growth hormone indicates four helical regions located within residues 7-34, 75-87, 106-127 and 152-183 [4]. The pK a values of histidines 20 and 22 presented here are compatible with the predicted secondary structure of bGH in the regions of the histidine residues ano, by analogy, with the X-ray detrmined structure of porcine growth hormone.
References 1 Wallis, M. (1973) FEBS Lett. 35, 11-14. 2 Bell, J.A., Moffat, K., Vanderhaar, Golde, D.W. (1985) J. Biol. Chem. 260, 8520-8525. 3 Seeburg, P.H., Sias, S., Adelman, J., De Boer, H.A., Hayflick, J., Thurani, P., Goeddel, D.V. and Heyneker, H.L. (1983) DNA 2(1), 37-45. 4 Abdel-Meguid, S.S., Shieh, H.S., Smith, W.W., Dayringer, H.E., Violand, B.N. and Bentle, L.A. (1987) Proc. Natl. Acad. Sci. USA 84, 6434-6437. 5 Wiithdch, K. (1986) in NMR of Proteins and Nucleic Acids, Wiley Interscience, New York. 6 Markley, J.L. (1975) Acc. Chem. Res. 8, 70-80. 7 Molday, R.S., Englander, S.W. and Kallen, R.G. (1971) Biochemistry 11, 150-158. 8 Bundi, A. and Wiithrich, K. (1979) Biopolymers 18, 285-297. 9 Burger, H.G., Edelhoch, H. and Condliffe, P.G. (1966) J. Biol. Chem. 241, 449-457. 10 Fellows, R.E., Jr. and Rogol, A.D. (1969) J. Biol. Chem. 244, 1567-1575. 11 Turner, C., Cary, P.D., Grego, B., Hem, M.T.W. and Chapman, G.E. (1983) Biochem. J. 213, 107-113. 12 Gronenborn, A.M. and CIore, G.M. (1982) Biochemistry 21, 4040-4048. 13 Clore, G.M. and Gronenborn, A.M. (1982) Biochemistry 21, 4048-4053. 14 Tanokura, M., Tasumi, M. and Miyazawa, T. (1987) Chem. Lett. 739-742. 15 Boscho, P., Seidel, W., Muradian, J., Tominaga, M., Paiva, A.C.M. and Juliano, L. (1983) Bioorg. Chem. 12, 34=44. 16 Williams, T.C., Corson, D.C., McCuggin, W.D., Oikawa, K., Kay, C.M. and Sykes, B.D. (1986) Biochemistry 25.. 1826-1834. 17 Maghuin-Rogister, G., Degelaen, J. and Roberts, G.C.K. (1979) Eur. J. Biochem. 96, 59-68. 18 Chou, P.Y. and Fasman, G.D. (1974) Biochemistry 13, 222-245. 19 Chen, C.-j.H. and Sonenberg, M. (1977) Biochemistry 16, 2110-2118.
166
Elsevier
BBA33290
Proton nuclear magnetic resonance study of the histidine residues of pituitary bovine growth hormone Nell E. M a c K e n z i e 1, S,M. Plaisted 2 a n d D . N . Brems 2,. ! Departmentof PharmaceuticalSciences, Universityof Arizona, Tucson,AZ and 2 The Upjohn Company, Kalamazoo, MI (U.S.A.)
(Received26 September1988) Keywords: Pituitarybovinegrowthhormone;NMR, t H-; Histidineresidues The pit'.value of hisfldlues 20, 22 and 170 of pituitary bovine growth hormone (pbGH) and their C2H deuterium exchange rates have been determined by high.resolution I H.NMR spectroscopy. Partial assignment of these pm'ameters imllcate that His-170 is a surface residue with a half-life for deuterium exchange at the C2H position of approx. 36 h and a ! ~ , value of 5.68 4. 0,11. A reveralble conformaflonal change of pbGH has also been characterized in terms of the invelvmtent of either hisfldiues 20 or 22. At or near physiological pH values, one of these residues is expelled from a buried, h y d e position to become fully solvent-exposed. The d|spori~ in pK. values between hisfi~nes 20 and 22 (4.67 4. 0.20 and 5.94 4. 0.10, in some combination) has been shown by computer modeling, to be compatible with these residues residing in an a-helical region of the protein. InUmlmtion Bovine growth hormone (bGH) is a pituitary protein consisting of 191 amino acid residues [1]. It is globular in conformation but a high-resolution X-ray crystallographic analysis of the three-dimensional structure has not been accomplished [2]. Recently however, the X-ray structure, at 1.9 A resolution, of the highly homologous protein [3], porcine growth hormone, has been reported [4] to consist of a four-helix bundle with the helices arranged in an antiparallel fashion. High-resolution nuclear magnetic spectroscopy is capable of providing detailed information on protein structure in solution, but a prerequisite for this is the assignment of resonances in the spectrum to individual amino acid residues in the sequence. For proteins the size of bGH (approx. 22 kDa) resolution of the NMR spectrum even at the highest available field strength is limited and hence the sophisticated two-dimensional methodologies developed for the structural elucidation of small proreins [5] cannot be readily applied. The most common exceptions to the lack of resolution in the IH-NMR spectrum of a protein are the * Present address: Lilly Research Laborat~ry, Hi Lilly and Company, Indianapolis,IN 46285,U.S.A. Abbreviations: pbGH, pituitary bovine growth hormone; NMR, nuclear magneticresonance; GdnHCI,guaninehydrochloride; SDS; sodiumdodecylsulfate,
downfield resonances attributable to the C2H imidazole ring protons of histidine residues. Few protein-bound protons have been more useful as 1H-NMR spectroscopic probes of intrinsic tertiary solution structure [6]. The associated resonances are sensitive to the microenvironments of the histidine residues, and any spectral changes can be attributed to discrete molecular perturbations of the folded protein. Generally, the microenvironment of individ,,~al histkiines within a protein can be assessed by comparison of their protonation parameters of pK. and deuterium exchange rates. This paper describes 1H-NMR studies of histidines 20, 22 and 170 of pituitary bovine growth hormone (pbOH) wherein their protonation parameters are characterized and partially assigned.
Materials and Methods Materials
Pituitary-derived b o l l (pbGH) was obtained from A.F. Parlow, lot 7899C, the Harbour Medical Center of the University of California at Los Angeles. Deuterium oxide (2H20, 99.9 atom% enrichment) and 2Hs-glycine (98 atoms enrichment) were obtained from Cambridge Isotope laboratories. Guanidine hydrochloride (GdnHCI) was ultra pure grade from Schwarz/Mann while all other reagents were of analytical grade. Sample preparation
Correspondence: N.E. MacKenzie, Department of Pharmaceutical Sciences,Universityof Arizona,Tucson,A285721,U.S.A.
Protein samples for 1H-NMR spectroscopy were prepared by dissolving pbGH (20 mg) in 10 mM 2H 5-
0167-4838/89/$03.50 O 1989 Elsevier Science Publishers B.V. (Biomedical Division)
167 glycine in 2H20 (1 ml), pH 9.5 at room temperature. The pH meter readings were not corrected for isotope effects on the glass electrode. Under these conditions the readily exchangeable protons (amino, amido, hydroxyl and carboxyl) were fully exchanged with deuterium within a few minutes *. The solution pH was adjusted with microliter quantities of dilute 2HCI or N a O ' H and pH readings were taken before and after data acquisition. The spectrum was accepted only if the deviation of the pH was less than 0.04 units. Protein samples were incubated for at least 30 rain at 25°C before NMR analysis.
NMR spectroscopy t H spectra were recorded at 25°C on a Varian XL400 spectrometer operating at 399.96 MHz for t H nuclei in the Fourier-transform mode. A total of 500 transients obtained by quandrature detection for a spectral width of 6 kHz by using 15 000 data points and an acquisition time of 1.334 s were averaged for each spectrum. Unless stated otherwise, the free induction decay was multiplied by an exponential factor equivalent to a line broadening of 2 Hz prior to Fourier transformation. Chemical shifts (8, ppm) are expressed relative to internal dioxane (~ = 3.740 ppm) downfield from 4,4-dimethyl-4-silapentane-l-sulfonate. Residual HO2H was suppressed by presaturation (1 s) prior to data acquisition.
Histidine zH exchange pbGH (80 mg) was dissolved in 10 mM 2Hs-glycine in 2H20 (4 ml) at pH 9.50. The solution was incubated at 37 °C in a sealed vial. Aliquots (1 ml) were taken at 0, 16, 42 and 67 h and immediately frozen in liquid nitrogen. The sample was stored at - 4 ° C until NMR analysis was performed.
Preparation and isolation of zH-fragments of pbGH A portion (1 ml) of deuterated pbGH (20 mg) was diluted by the addition of 96% formic acid (2 ml). Cyanogen bromdie (40 mg) was added to this solution, and the reaction mixture was left to stand at room temperature in a sealed vial for 24 h. The reaction mixture was then diluted with water (45 ml) and lyophilized. The lyophilate was dissolved in 40% formic acid (2 ml) and chromatographed on a Sephadex G-50 column (1 × 50 cm) using 20% formic acid as the mobile phase. Elution was monitored by uRra-violet detection at 280 nm. The fractions containing the AB fragment
* The average expected time constant for amide proton exchange of unfolded b o l l at 2 0 ° C and pH 9.5 is approx. 0.1 ms [7]. Since within 100 s all the amide protons had exchanged, the degree of protection from exchange by the folded protein must be less than 100000-fold.
B
H3,. H1
--Y
AA H2,H3 H1-A
8'.5
81o
!
|
T.o
Bio PPr
Fig. 1. Aromatic region of the 400 MHz I H-NMR spectrum of pituitary bovine growth hormone (A) pH 1.91, (B) pH 9.73. Insets: Line-narrowed spectra of the histidine C2H resonances, H1, H2 and H3.
(residues 6-124 linked to 150-179) were collected, diluted 4-fold and lyophilized. The peptides were reduced for 1 h with 8 M GdnHC1 (1.2 ml) containing 0.1 M ammonium bicarbonate and dithiothreitol (2 mg/ml) and rechromatographed on the same column as above after the addition of 0.8 ml 96% formic acid. Fraction A consisting of residues 6-124 and fraction B consisting of residues 150-179 were collected and lyophilized. Results
The ~H-NMR spectrum of pbGH is typical for proteins in the molecular weight range of 20 kDa, consisting of broad regions of ill-resolved resonances. The aromatic region of tile 1H-NMR spectrum of pbGH at pH 1.91 and 9.73 is shown in Fig. 1A and B, respectively. The resonances X and Y (Fig. 1A) are assignable to the 2',6' and the 3',5' aromatic protons, respectively, of the six tyrosine residues of the protein. The average chemical shifts of these resonances are 7.059 ppm and 6.764 ppm and are almost identical to those for tyrosine in small peptides [8]. Similarly, the resonances H1, H2 and H3 can be assigned to the C2H imidazole ring protons of the three histidine residues of the pbGH which occur at positions 20, 22 and 170 in the amino acid sequence. The chemical shifts of 8.659 ppm, 8.636 ppm and 8.620 ppm, are consistent with those of fully protonated histidine reisdues both in the isolated amino acid [8] and small peptides [6]. The small differences in chemical shifts are due to amino acid primary sequence variations in the region of the histine residues. Based on
168 TABLE I
8.61
Chemical shifts and pKa values
8.4E 8.3~
Chemical shifts and pK a values derived for the histidine C2H resonances of pituitary bovine growth hormone by non-finear least-squares analysis of the titration data in Fig. 2. Figures in parentheses are the calculated standard errors.
8.23 ~8.1C 7.84
O (acidic)
7.71 758 74e
HI
H2
H3
8.629 (0.015)
8.616 (0.013)
8.679 (0.027)
8 (basic) I
2
~
I
3
4
~
~
~
8
I
9
I
10
a chemical shift argument, pbGH assumes an essentially random coil configuration at this pH. Bovine ~ , : t l i hormone is known to undergo a reversible~pH-induced conformational change [9] and indeed the 1H-NMR spectrum of the protein at pH 9.73 (Fig. 1B) is that of a specifically folded globular structure. The perturbations of the aromatic resonances seen in Fig. 1B can be ascribed to a combination of factors including direct ionization and proximal electrostatic interactions together with the environmental changes of various residues within the folded protein. All these factors presumably contribute to the upfield shift of resonances H1, H2 and H3 with the most significant effect on chemical shift arising from deprotonation of the imidazole rings. The chemical shifts of the histidine C2H resonances H1, H2 and H3 of pbGH, plotted against pH at 25 °C are shown in Fig. 2. Titration data for the bJstidines represented by H1, H2 and H3 were analyzed by a three.parameter fit to the Henderson-Hasselbach equation (Eqn. 1) using a non-linear regression program for each ionization event. S H ÷ - SIlo =
7.670
7.561
(0.008)
(0.011)
OA
1.077
1.046
1.118
pK a
5.944 (0.101)
5.680 (0.109)
4.674 (0.125)
11
Fig 2, Variation in chemical shirt (O) with pH for the resonances HI (o), H2 03) and H3 (z) of pituitary bovine growth hormone shown in Fig. 1. The solid lines were calculated as described in the text and the fitted parameters are given in Table I.
81l ÷ - ~
7.552
(0.008)
I
couldonly be followed over a lirc,;te~i'pH range as at values between approx. 6 to 8, this resonance could not be resolved from the envelope of aromatic resonances. Above the latter pH value, resonance H3 reappears downfield of this envelope and co-titrates with resonance H2 (Fig. 2).
Histidine ZH.exchange kinetics The time-course of deuterium incorporation into the C2H position of histidines 20, 22 and 170 is shown in Fig. 3. Resonances a, b and c (Fig. 3) are, in some combination, the I H-NMR resonances of the C2H pro-
Hr)
go
g,
+[H ÷ ]
(1)
The observed chemical shift at a given pH is taken as 8ob~, &H+ and 8H0 are the" chemical shifts of the protonated and unprotonated forms of the histidine residue, respectively, while K a is the acidic dissociation constant. The pK a values derived for the histidine residues of pbGH are summarized in Table I. The best fit to the pH titration data is represented by the solid lines for H1, H2 and H3 (Fig. 2). In general, resonances H1 and H2 follow a simple signoidal curve indicative of a single ionization. The titration of resonance H3 occurs over a relatively small pH range (about 2 pH units) and suggests that this protonation - deprotonation event is a cooperative process. Resonance H3
8. o
!
8. 5
8. o
8. 5
' PPM 8.60
Fig. 3. Deuterium exchange of the imidazole C2 protons of pituitary bovine growth hormone. The histidine C2H region (resonances a, b and c) of a series of 400 MHz I H-NMR spectra of pituitary bovine growth hormone preparations containing 2~$ (w/w) SDS. The protein had been deuterated at 37°C (pH 9.50) in 10 mmol 2Hs-glycine (in 2H20) for the time (in hours) indicated prior to data accumulation.
169 0.05~__________________~ ,~
TABLE II First-order rate constants
First-order rate constants for deuterium exchange of the C2 protons of the histidine residues of pituitary bovine growth hormone at 37 o C (pH 9.5) derived by least-squares analysis of the data shown in Fig. 4.
-o.~.
-~
~ H1 H2 H3
-0.45 ,
-o~
,
2'o
il
|
4o Time
e'o
k (s- i ) X 10 6
(Correlation coefficient)
k/k
1.17 0.96 1.91
(0.98) (0.99) (0.99)
0.61 0.50 1
fastest
|
(h)
Fig. 4. Deuterium exchange of the imidazole C2 protons of pituitary bovine growth hormone. The NMR srectra shown in Fig. 3 were integrated and plotted, as the fraction of unexchanged resonance, on a logarithmic scale versus time (in hours l to enable the rate constant (Table If) to be determined by linear-regression analysis.
tons of the histidines and differentially diminish with time. The spectra were recorded at pH 6 after the addition of 2~ (w/w) sodium dodecyl sulfate (SDS) to the deuterated protein preparations (see Materials and Meihods). Under these defL~-d conditions the three resonances were baseline-resolved and had identical line wid,:hs. This facilitated the accurate integration of the
area of the resonances. A plot of the C2H resonance area as a function of time is shown in Fig. 4. A non-exchangeable aromatic resonance was used as an integration reference and the data were fit by linear-regression analysis to the expression: log (fraction unexchanged) = - kt. The derived values for k, the first-order rate constant, are given in Table II.
Partial assignment of rate constants The 1H-NMR spectrum of the histidine C2H resonances of pbGH which had been deuterated for 42 h is
1.5
b AB C
0.75
t-
.o
B
~x 1.5 0 0'1
<
0.75
--I
8.85
i
8.80
I
8.75
i
8.70
l
8.65
w
8.60
I
8.55 PPIVi
Fig. 5. (A) The histidine C2H region of the 400 MHz 1H-NMR spectrum of pituitary bovine growth hormone after the addition of 2% (w/w) SDS. The protein had been deuterated at 37 o C (pH 9.50) in 10 mmoi 2Hs-glycine (in 2H20) for 42 h prior to data acquisition. (B) The histidine C2H region of the 400 MHz 1H-NMR spectrum of deuterated peptide, fragment B, (residues 6-124) derived from the protein sample described in (A). (C) As for (B) but of the deuterated peptide, fragment A (residues 150-179).
el 0
I
I
l
9 18 27 Elution volume (ml)
36
Fig. 6. Gel filtration chromatography of deuterated pituitary bovine growth hormone "fter cleavage with cyanogen bromide. The top chromatogram represents the elution profile of the complete cyanogen bromide reaction. Fractions corresponding to the fragment AB (residues 6-124 linked to 150-179) were collected and the disulfide was reduced (see Materials and Methods). The bottom chromatogram represents the elution profile of the reduced fragmem~ A (residues 6-124) and B (residues 1b0-179).
170 shown in Fig. 5A. This protein was then cleaved by reaction with cyanogen bromide. The elution profile of the complete reaction mixture from gel filtration chromatography is shown in the upper trace of Fig. 6. The peak labeled AB corresponds to the peptide fragment composed of residues 6-124 and 150-179 finked by a disulfide bridge btween Cys-53 and Cys-164 [11]. The AB fraction was collected and reduced with dithiothreitol. The chromatogram of reduced peptide fragments A (residues 6-124) and B (residues 150-179) is shown in the lower trace of Fig. 6. The 1H-NMR spectrum of the histidine C2H resonances of fragment A and B are shown in Fig. 5B and C, resp~tive!y. The C 2 E resonances of histidines 20 and 22 occur (in some combination) at 8.783 and 8.701 ppm (Fig. 5B) and that of His-170 (Fig. 5C) at 8.616 ppm. These can be compared to the chemical shifts of the C2H resonances in the deuterated intact protein (Fig. 5A) at 8.774, 8.704 and 8.631 ppm under identical solution conditions. From comparison of the chemical shift data and the relative integrated areas of the resonances in Fig. 5, His-170 corresponds to resonance C, which has been shown to be associated with the fastest deuterium exchange rate (Table II).
Partial assignment of pKa values To assign a pK, value to His-170, the C2H resonance C (Figs. 3, 5A and 7) was compared to H1, H2 or H3 (Figs. 1A, B and 7A) under identical solution condib
i42
o'l
H3
alb alb
'
8.4
'
8:2
'
'
S,O
'
"
PPM
Fig. 7. Inset: As described for Fig. 5A after 42 h of deuteration (A) The histidine C2H region of the 400 MHz IH.NMR spectrum of the deuterated protein described for the insert above but accumulated in the absence of SDS. (B) The hisfidine C2H region of the 400 MHz SH-NMR spectrum of pituitary bovine growth hormone prior to C2H deuterb~un exchange.
tions. In the absence of SDS the histidine C2H resonances of pbGH, H1, H2 and H3 are only sufficiently resolved, to allow accurate integration, over the small pH range of approx. 4.80 to 5.20. Fig. 7B shows the ~H-NMR spectrum of the histidine C2H resonances of pbGH at pH 4.90. This can be compared with the resonances of deuterated pbGH recorded under the same conditions (Fig. 7A). Comparison of the relative integrated areas of the resonances in Fig. 7A shows that the central resonance, corresponding to H2, has lost the most area and can be assigned to resonance C, of Fig. 7 (inset). The peak assignments are therefore as follows: H1, H3 -- C2H of histidines 20 and 22 (in some combination), H2 - C2H of histidine 170. Diseussion The linewidth of resonances H1, H2 and H3 in the random coil configuration (Fig. 1A) at pH 1.91 is 7 Hz. This increases to only 9 Hz in the folded protein at pH 9.73. This indicates that pbGH has a substantially mobile or flexible structure in the latter pH range. This is corroborated by the fast deuterium exchange of all amide protons at this pH and the fast deuterium exchange rates of the histidine C2H protons (Table If). The half-life of the deuterium exchange of His-170 is approx. 36 h and can be compared with the rate of exchange of a surface residue [11] or those histidine residues known to reside in very mobile protein regions [12,131. The pKa values of histidines 20, 22 and 170 (Table I) are all lower than the pK a value (6.65, 21 ° C) of Nacetylhistidinemethylamide, the standard model for exposed histidine residues free from specific perturbation [14,15]. An environmental factor that can cause the lowering of a histidine pK a is the proximity of a positive charge to the imidazole ring [6,14]. Thus the ion-pair interaction of the histidine residues H1 and H2 of pbGH with a prctonated basic residue may account for their low pK, values. In addition, histidine residues which are buried in the protein in the unprotonated state have been assigned very low pK a values [13,16,1'7]. It is suggested here that the very iow pKa value of resonance H3 (histidine 20 or 22) is due to its burial within the protein. The resonance H3 is significantly shifted upfield in the unprotonated state with the lowest measurable chemical shift value of 7.533 ppm (Fig. 2). At higher pH values, this resonance moves further upfield and cannot be resolved from the envelope of aromatic resonances. This unusually low chemical shift value for this resonance presumably is a consequence, in part, of ring current effects and so in the buried (unprotonated) state this histidine residue is closely associated with an aromatic residue(s). Similarly, resonance H1 (Fig. 2, Table I) is shifted upfield in the deprotonated state at all alkaline pH values and again this arioes from
171
the proximity of the associated histidine residue to an aromatic residue(s). At pH values greater than approx. 8, resonance H3 reappears at a chemical shift coincident with that of resonance H2 (Fig. 2). It is suggested that a conformational change of pbGH occurs at or near physiological pH values causing expulsion of this histidine residue and exposure to solvent. This is supported by the now 'normal' chemical shift value for a deprotonated histidine C2H resonance. The large difference in the pK a values of histidines 20 and 22 (H1 and H3, Table II) indicates that their microenvironments are very different. Model studies show that for two histidine residues separated by a single amino acid residue in a B-pleated sheet, the separat,on between the two imidazole rings will be within the range 2 to 8 ,~. This is close enough to allow the two imidazole rings to interact and compete for the same proton. For similarly situated residues in an ~x-helix the imidazole rings lie on opposite sides of the helix, far enough apart ( > 12 A) that they may reside in sufficiently diverse environments to accommodate the pK a values determined in this study. Based on secondary structure predictions. [18] the segments 10-34, 66-87, 111-127 and 186-191 should be helical [19]. The X-ray diffraction studies of poricne growth hormone indicates four helical regions located within residues 7-34, 75-87, 106-127 and 152-183 [4]. The pK a values of histidines 20 and 22 presented here are compatible with the predicted secondary structure of bGH in the regions of the histidine residues ano, by analogy, with the X-ray detrmined structure of porcine growth hormone.
References 1 Wallis, M. (1973) FEBS Lett. 35, 11-14. 2 Bell, J.A., Moffat, K., Vanderhaar, Golde, D.W. (1985) J. Biol. Chem. 260, 8520-8525. 3 Seeburg, P.H., Sias, S., Adelman, J., De Boer, H.A., Hayflick, J., Thurani, P., Goeddel, D.V. and Heyneker, H.L. (1983) DNA 2(1), 37-45. 4 Abdel-Meguid, S.S., Shieh, H.S., Smith, W.W., Dayringer, H.E., Violand, B.N. and Bentle, L.A. (1987) Proc. Natl. Acad. Sci. USA 84, 6434-6437. 5 Wiithdch, K. (1986) in NMR of Proteins and Nucleic Acids, Wiley Interscience, New York. 6 Markley, J.L. (1975) Acc. Chem. Res. 8, 70-80. 7 Molday, R.S., Englander, S.W. and Kallen, R.G. (1971) Biochemistry 11, 150-158. 8 Bundi, A. and Wiithrich, K. (1979) Biopolymers 18, 285-297. 9 Burger, H.G., Edelhoch, H. and Condliffe, P.G. (1966) J. Biol. Chem. 241, 449-457. 10 Fellows, R.E., Jr. and Rogol, A.D. (1969) J. Biol. Chem. 244, 1567-1575. 11 Turner, C., Cary, P.D., Grego, B., Hem, M.T.W. and Chapman, G.E. (1983) Biochem. J. 213, 107-113. 12 Gronenborn, A.M. and CIore, G.M. (1982) Biochemistry 21, 4040-4048. 13 Clore, G.M. and Gronenborn, A.M. (1982) Biochemistry 21, 4048-4053. 14 Tanokura, M., Tasumi, M. and Miyazawa, T. (1987) Chem. Lett. 739-742. 15 Boscho, P., Seidel, W., Muradian, J., Tominaga, M., Paiva, A.C.M. and Juliano, L. (1983) Bioorg. Chem. 12, 34=44. 16 Williams, T.C., Corson, D.C., McCuggin, W.D., Oikawa, K., Kay, C.M. and Sykes, B.D. (1986) Biochemistry 25.. 1826-1834. 17 Maghuin-Rogister, G., Degelaen, J. and Roberts, G.C.K. (1979) Eur. J. Biochem. 96, 59-68. 18 Chou, P.Y. and Fasman, G.D. (1974) Biochemistry 13, 222-245. 19 Chen, C.-j.H. and Sonenberg, M. (1977) Biochemistry 16, 2110-2118.