NMR characterization of self-association of a helical peptide using deuterium exchange experiments

COLLOIDS

;

AND Colloids and Surfaces A: Physicochemical and Engineering Aspects 11511996) 39 45

ELSEVIER

A

SURFACES

NMR characterization of self-association of a helical peptide using deuterium exchange experiments Imma Fernfindez, Josep Ubach, David Andreu, Miquel Pons * Departament de Quimica Or gtinica, Unit~ersit~tt de Barceloncl, Marti i l:'r~mqu~s 1-11, E-08028 Barc~,lo~7~l, Sp~li~l

Received 11 September 1995: accepted 14 January 1996

Abstract

The 15-residue hybrid peptide containing residues 1-7 from cecropin A and residues 2 9 from melittin, CA(I 7)M(2 9), is a potent antibiotic with broader activity than cecropin A but without the cytotoxic character of melittin. In the presence of the helix inducer hexafluoroisopropanol the peptide forms aggregates of amphipathic :~helices. Aggregation causes very slow proton-deuterium exchange in some amide protons in the C-terminal region. This provides a method for estimating the association constant (~10 ~' M -~) as well as the stoichiometry of the aggregates. Exchange could be mediated by helix breathing or could inxolve complete disruption of the helix. These two mechanisms can be differentiated by comparing the decay of nuclear Overhauser effect cross-peaks involving two anaide protons with the decay of each individual proton.

K~,ywords. Antibiotic peptides; Deuterium exchange: NMR; Peptide aggregation 1. Introduction Peptide-peptide interactions play a key role in many biologically relevant processes including channel formation, signal transduction, and regulation of binding. Aggregation phenomena involving peptides are also relevant in drug delivery processes and can therefore strongly influence the bioavailability of peptide drugs. Vancomycin antibiotics have been shown to dimerize strongly and this has been suggested to be relevant to their biological activity [ 1]. Deuterium exchange experiments [2] are widely used to study protein folding, the rationale being that amide protons of residues buried in the interior of a folded protein cannot exchange efficiently with the solvent while exposed residues exchange * Corresponding author. 0927-7757'96,'$15.00 c~'>1996 Elsevier Science B.V. All rights reserved PII S0~727-7757(96)03617-5

with rates of the order of 103S ! [3]. Small peptides are completely exposed to the solvent and therefore complete exchange of all the amide protons usually prevents their observation in D20. Amphipathic helices are a common structural motif found in many membrane-active peptides. These peptides are often not well structured in water but they adopt a helical conformation in liposomes or in the presence of helix-inducing solvents such as trifluoroethanol (TFE) or hexafluoroisopropanol (HFIP). P e p t i d e peptide interactions can also contribute to the stabilizalion of the helical structure. In the course of studies aimed at finding peptides with improved antibacterial activity', a 15-amino acid hybrid peptide consisting of residues 1 7 of cecropin A followed by residues 2 9 of melittin (CA( 1 7)M(2-9); Table 1 ) was found to retain the ability to adopt an amphipathic :~-helical confor-

40

I. Ferngmde: et al./Colloids Surfiwes A: Physicochem. Eng. Aspects 115 (1996) 39 45

Table 1 Sequences of the peptides cecropin A, melittin and CA(I 7}M(2 9t Peptide

Sequence

Cecropina A Melittin CA(1 7)M(2 9)

H-KWKLFKKIEKVGQNIRDGIIKAGPAVAVVGQATQIAKK NH 2 H GIGAVLKVLTTGLPALISSWIKRKRQQ NH 2 H KWKLFKKIGAVLKVL-NH2

mation and to display a broad and potent antibacterial spectrum and no cytotoxic character [4]. In spite of the small size of C A ( 1 - 7 ) M ( 2 - 9 ) 15 residues , N M R spectra recorded in 88% D 2 0 : 12% HFIP-d2 showed some of the amide protons in the C-terminal part of the peptide unexchanged even after several hours while the remaining amide protons had disappeared after a few minutes. In this paper we report our studies of deuterium exchange of C A ( 1 - 7 ) M ( 2 9) and show that it is the result of aggregation and can provide useful information on the structure of the aggregate.

2. Experimental The peptide CA(1 7)M(2-9), with the sequence shown in Table 1, was synthesized by solid-phase techniques and characterized as previously described [-5]. All spectra were recorded at 5 0 0 M H z in a VXR-500 spectrometer. Assignment of the spectra was done by combining the information from double-quantum filtered correlation spectroscpy (DQF-COSY), total correlation spectroscopy (TOCSY) (mixing time 80min) and nuclear Overhauser effect spectroscopy (NOESY) (mixing times from 100 400min) experiments. Presaturation was used to suppress the water signal in all the experiments. Proton-deuterium exchange of amide groups was studied in samples of CA(1-7)M(2 9) lyophilized from H 2 0 and dissolved in 12% HFIP-d6/88% D20. Onedimensional (1D) spectra were acquired every 2 min over the first 4 h. Fast NOESY experiments to study the decay of N H - N H cross-peaks were carried out using 100 increments (32 scans) and extending the first

dimension using linear prediction. Each experiment lasted 10 min. This time is short enough, compared to the rate of exchange, that broadening of the signals due to apparent fast relaxation during the evolution period is minimal.

3. Results and discussion Standard N M R methods [6] were used to carry out a complete sequential assignment of CA(1 7) M ( 2 - 9 ) in 15%DzO/85%H20 or in 15%D20/ 12%HFIP/73%H20. No significant changes in chemical shift were observed when the peptide concentration was varied from 0.5 to 6 mM in 12% H F I P or when the H F I P percentage was raised to 50% (v/v). Deviations from aqueous random coil values for the He and N H chemical shifts of CA(1 7)M(2-9) in H 2 0 and H F I P are consistent with the adoption of an :~-helical conformation [-7]. The helical conformation was supported by the observation of strong NH/ NHi+ 1 NOEs together with weaker Hi NHi+a connectivities and suggested the presence of a helical structure throughout most of the sequence. The lower intensity of N H - N H crosspeaks involving protons near the N-terminal region (e.g. Trp2 Lys3, or Lys3-Leu4) as well as the smaller upfield shifts in this region suggested some helical fraying in this terminal region. In contrast, the helical structure appeared well formed up to the last residue in the C-terminus. Fig. 1 shows the decay in the intensity of the signal from one of the slowly exchanging amide protons after having dissolved samples of two different concentrations in 88%DzO:12%HFIP (pH3.4). The observed decay is clearly nonexponential, especially at the higher concentration, but it could be fitted to a double exponential. The

L Fernandez et al./ColloMs Su(laces A: Physicochem. Eng. Aspects 115 (1996,) 39 45

41

1,0-

0.8

~

mM

a

........... Z

0.4

02 1 F ...... 0

I

I

I

50

1 O0

150

------

I

--

200

I

F

250

300

time (rain)

1.0

0.8

b

..~06

6 mM 04

0.2

I

I

50

100

I

150 time (rain)

I

200

- -

I

250

--

1

300

Fig. 1. Amide proton exchange curve of residues Leul2 la) and Va114 (b) at 1 and 6 mM concentrations respectively in 129~ HFIPd 2 : 88% D20 (pH 3.4; 25 'C). Experimental points were obtained from 1D speclra recorded at 2 rain intervals. Solid lines arc the best fits to a biexponential decay.

I. Ferndndez et al./Colloi&" Surfaces A: Physieochem. Eng. Aspects 115 ( 19963 39 45

42

relative amplitudes and rates of the two components are presented in Table 2. One important observation is that the slower rate is strongly concentration-dependent. This indicates that protection against deuterium exchange is the result of aggregation. This result contrasts with the lack of any observable changes in chemical shifts or linewidths when the concentration was varied in the range 0.5 6 mM, but the presence of aggregation was confirmed by ultracentrifugation, electrospray mass spectrometry (ES-MS) and concentrationdependent CD spectra, all of them indicating substantial aggregation in the presence of HFIP [8]. The observed protection could have two possible origins: a decrease in the breathing of the helix associated with the enhanced stability of the helical structure in the aggregate and/or a protection of some NH residues by intermolecular contacts. In an attempt to clarify this point we have studied the rates of exchange of pairs of adjacent NH residues involved in the helical structure and, therefore, displaying strong sequential NOE cross-peaks in NOESY spectra. When the rate of decay of the cross-peak between two NH protons is compared with the rates of decay of each individual proton two different results are possible, as illustrated schematically in Fig. 2a. If each of the amide protons exchanges independently of its neighbour, i.e. the event leading to the exchange of one of the protons does not necessarily cause the other proton to become exposed, the decay in the intensity of

the cross-peak will be faster than that of the individual components. This is due to the fact that molecules in which only one of the protons is exchanged cannot give an NH-NH cross-peak but still contribute to the intensity of the unexchanged proton. However, if a single event causes the simultaneous exchange of both protons, the rates of exchange of the cross-peaks and the individual protons will be the same. Fig, 3 shows a comparison of the decay in the intensities of cross-peaks between the amide proton and the non-exchangeable alpha proton in the same residue and, therefore, involving only one amide proton, with the decay of the cross-peak between two amide protons. In Fig. 3a, corresponding to residues 10 and 11, the three curves overlap, indicating that the exchange of the two protons takes place simultaneously. A less obvious case, involving protons in residues 11 and 12, is shown in Fig. 3b. In this case the rates of exchange of the individual amide protons are different and this may suggest that in this case the two processes are independent. A comparison with the decay curve of the NH NH cross-peak shows, however, that it does not decay faster than the faster of the two components, indicating that both protons are simultaneously exposed to solvent although their intrinsic exchange rates may be different. Quantitative interpretation of exchange rates is usually done in terms of protection factors that relate the observed rate to that of fully exposed

Table 2 H D exchange rates of CA(1-7)M(2 9) amide protons a Residue

Phe 5 Ile 8 Gly 9 Val 10c Val 11 Leu 12 Lys 13 Val 14

1 mM

6mM

fl b

k 1 × 104 (S 1)

.[~b

k2 × 104 (S 1)

0.53

2

1.00 0.47

12.4 50

0.61 0.85

2 0.8

0.39 0.15

60 50

0.66

4

0.34

70

fib

kl × 104 (s l)

f2 b

k2 × 104 (s i)

0.85 0.70

0.7 0.8

1.00 0.15 0.30

7 5 60

0.67 0.69 0.74 0.58

0.5 0.3 1 2

0.33 0.31 0.26 0.42

1.3 50 70 14

a Measured in 12%HFIP-d 6 at pH 3.4 (electrode reading) and 25°C. b fl and f2 refer to the relative contributions of the slow (rate kl) and fast (rate k2) components to the observed decay. c Overlaps with aromatic.

L Ferndndez et al./Colloid~ Surlizces A." Phrsicochem. Eng. Aspects 115 ( 1996 ) 39 45

-CO-NH-CHR-CO-ND-

-COINH-CHRICO-NH-

43

a

-CO-ND-CHR-CO-ND-

-CO-ND-CHR-CO-NH-

1

~ ,

0.8 0.6 0.4

b

. NH-CHcross-peak "" NH-NHcross-peak ,,"" Independentexch. k~v~-~,Q NH-NHcross-peak

0.2 0

0.5

1

1.5

2

time

Fig. 2. (a) Stepwise (thin arrows) versus simultaneous (bold arrow) exchange of neighbouring amide protons, lb) Simultaneous exchange causes the NH N H cross-peak intensity to decay at the same rate as lhe individual protons. Stepwise exchange would result in a faster decay of the N H - N H cross-peak.

NH protons. Exchange rates are dependent on experimental conditions (pH, temperature, ionic strength) [9,10] and the peptide sequence [11]. However, the presence of two components exchanging at different rates provides an internal control by which it may be possible to derive a relative protection factor that would specifically reflect the effect of aggregation on deuterium exchange. Assuming a specific model for the aggregate, two types of information can, in principle, be derived from the exchange data: the association constant and the size of the aggregate. The case of a linear aggregate is illustrated in Fig. 4. In this model fast exchange components would correspond to molecules located in the extremes of the chain while molecules with slowly exchanging protons would be located in the interior. If a chain breaks, two interior molecules, with slowly exchanging protons, will become terminal and their amide protons will exchange faster. Conversely, the interacting ends of two chains that are joined

together become protected. One can thus define a pseudodimerization constant (K,,) for the process of joining two chains of n/2 molecules to give an aggregate of n monomers: 2P,,2 ~ P

[P,,] K, -- [ p , 212

This constant can be determined from the ratio R between the slow and fast rate constants of exchange and the total peptide concentration C,~ using a straightforward modification of the method described by Mackay et al. [ 12]: K,-

n (R--l) - 8('T R 2

kslow with

R

--

](fast

In the same model, the relative amplitudes of the slow and last exchanging components provide an estimate of the length of the chain. The slow exchanging component represents ~ 60 70% of the total population. If the distribution of sizes was

44

L Fernhndez et al./Colloids Surfaces A." Physicochem. Eng. Aspects 115 (1996) 39-45

---:r,°j 11 lO-ll

0.8-

c

E "~ 0.6 N E

0.4

0

I

]

I

I

I

[

I

10

20

30

40

50

60

70

time (min)

b 12-.-1112 0.8

:>,

0.6 E o

0.4

0

I

I

I

I

l

I

F

10

20

30

40

50

60

70

time (min)

Fig. 3. Normalized intensity of N H - N H cross-peak intensity (D) and NH CHe cross-peaks (A, II) corresponding to (a) residues 10 and 11 and (b) residues 1l and 12 of CA(1 7)M(2 9) measured using fast NOESY spectra at 10min intervals. Peptide concentration was 6 mM (pH 3.0; 25 'C).

L Ferndndez et aL/Colloids Surflu'es A." Physicochem. Eng. A.~pects 115 1996i 39 45

45

Kpseudodim

f slow f fast

=

f .nner

=

f end

2 n=6 4

Fig. 4. Schematic representation of a linear aggregate of CA( 1 7 )M(2 9 t. Each circle represents a molecule in a helical conformation. Filled circles correspond to molecules showing slowly-exchanging protons while open circles represent lhe fastcr component.

monodisperse, this would indicate a chain of six molecules ( 4 / 6 = 0 . 6 7 ) in agreement with the ES-MS and ultracentrifugation results. The Leul2 amide proton of C A ( 1 - 7 ) M ( 2 9) has a protection factor (R -~) of around 60 at 1 mM and ~ 170 at 6 raM. Amide protons located either side in the sequence have lower values. The pseudodimerization constant derived from these protection factors has a value of .~3 x 10 ~ M

in= 6).

4. Conclusions Our results show that deuterium exchange experiments can provide a diagnostic test for peptide aggregation under N M R conditions and, in favourable cases, can also provide an estimate of the association constant and the size of the aggregate. Moreover, measurement of the rates of decay of N O E cross-peaks have provided evidence for the sinmltaneous exchange of pairs of amide protons.

Acknowledgments We acknowledge the use of the N M R facilities of the "Serveis Cientifico-T~cnics" of the University

of Barcelona. This work was supported in part by funds from the CICYT {project PB94-924).

References [ I ] J.P. Mackay, U. Gerhard, D.A. Beauregard, M.S. Westwell, M.S. Searle and D . H Williams, J. Am. Chem. Soc., 11~3(1994) 4581. [2] S.W. Englander and L. Mayne, Almu. Rev. Biophys. Biomolec. Struct., 21 (1992) 243. [3] R.S. Molday, S.W. Englander and R G . Kallcn, Biochendstry, 11 (1972) 150. [4] D. Wade, D. Andreu, S.A. Mitchell, A.V. Silvcria, l.A. Boman. H.G. Boman and R.B. Merrifield, Int. J. Pcpl. Protein Res., 40 (1992) 429. [5] D. Andreu, J. Ubach, 1.A. Boman. B. W~ihlin, D. Wade and R.B. Merrifield, FEBS Lett., 296 t 19921 190. [6] K. Wt).thrich, N M R of Proteins and Nucleic Acids, John Wiley, New York, 1986. [7] M. Bruix, M. Perello, J. Herranz, M. Rico and J.1.. Nieto, Biochem. Biophys. Res. Commun.. 167 11990i 1009. [8] 1. Fernandez, J. Ubach, M. Euxreiter. J.M. Andrcu. D. Andreu and M. Pons. Chcm. Eur. J., in press. [9] J.J. Englander, J.R. Rogero and W Enghmder, Anal. Biochem., 147 (1985) 234. [10] P.S. K i m a n d R.L. Baldwin, Biochemistry, 21 119S2) 1. [11] A.D. Roberston and R L . Baldain. Biochemistry, 30 11991 ) 9907. [ 12] J.P. Mackay, U. Gerhard, DA. Beauregard, R.A. Maplestone and D.H. Williams, J. Am. ('hem. Snc., 116 ( 19941 4573.