- Email: [email protected]
Hydrogen exchange in hydrated films of proteins
Biophysical Chemistry 12 (1980) 279-284 Q North-Holland Publishinf Company
HYDROGEN APPLICATION
EXCHANGE
IN HYDRATED FILMS OF PROTEINS. LAC REPRESSOR CORE
TO THE E. COLI
Received 18 March 1980
An o@inal easy method of hydrogen to deuterium cschange in hydrated films of proteins, followed by infrared absorption measurements, is described and appiicd to films of the E. c&i lac repressor core, in order to csnmine the cffcct of isopropyl-p-D-thiogalnctoside (IPTG) binding. An estimation of about 255 (1 helical structure in this protein fraemcnt is deduced from the exchange curve. The bindinp of IPTG to the core does not affect the exchange curve within the cspcrimental error limits.
I _ Introduction
The hydrogen exchange rates of peptide NH groups are known to be very sensitive to protein conformations_ They depend upon hydrogen bonding in Q helix or p sheets, and upon hydrophobic and steric environment of the !abile protons in the molecule [l-3] _ Some difficulties in hydrogen exchange techniques, when measured by infrared absorption spectroscopy on solutions arise from the necessity to dissolve quickly (which is often no possible) the lyophilized protein in D,O, and to introduce this solution into a cell which is only a few microns thick in a very short period of time. The very fast exchange of some protons at neutral pH, with half-life of the order of one second [4,5] does not allow extrapolation of the exchange curve to zero time_ Thus the protein absorbance in the initial undeuterated state is not known without indirect estimation. A similar problem exists in determining the total number of labelled protons at zero time in tritium-hydrogen exchange experiments which are measured by gel filtration and radioactivity counting [2,6-S]. Furthermore, in this latter exchange-out method, very slow exchanging hydrogens are very difficult to label except * Present address: Division of Biological Sciences, N.R.C.C.. Ottawa, Canda KlA 0R6.
if the incubation time is increased or if the conditions of incubation are more drastic (for exatnple heating the solution). We describe here a simple method to follow the hydrogen-deuterium exchange of peptide N-H groups by infrared absorption measurements of the amide 11 absorbance at 1550 cm -I. in hydrated films of proteins. when exposed to heavy water DzO vapour. This method allows one to determine the initial point of the exchange curve, and does not require any lyophilization of the protein, as well it avoids the difficulties of infrared spectroscopy on dilute solutions. In this work we describe how we have used these techniques to study the Zac repressor core and its interaction with an inducer_ Under certain conditions proreolytic cleavage of the lac repressor occurs between residues 50 and 60 yielding two fragments: the N-terminal headpiece. which contains the major portion (if not all) of the DNA binding site, and a tetrameric core with full inducer binding activity [9,10]_ We were interested in studying the effect of an inducer, in this work isopropyl-P-D-thiogalactoside (IPTG), on this core. in order to obtain further information on its mechanical action. This core has the advantage of a greater solubility than the intact protein, allowing the formation of concentrated homogeneous films of the proiein and their study by infrared absorption spectroscbpy.
2. Materials and methods E c&i lac repressor from strain E&aH 493 (a generous gift of Dr. Beyreuthcr and Dr. ~~~IIer-Hill) was purified by ammonium sulfate fractionation and column ~llro~nato~aplly on phosphocelfulose as described by Miiller-Hill et al. [ 1 I 1. Proteolysis was performed as described by Geisler and \Veber [ 121 using clostripain QVorthfngton). The core preparation was analyzed on SDS polyacrylamide gel and run as a sin& band corresponding to molecular weight about 32 000 accounting for residues 60 to 360. Films of repressor core were deposited on calcium fluoride (CaF2) IR windows by gentle drying at ambient temperature and humidity. 1 ml of a solution 4.7 X 10e5 M core (i.e. 1.5 mg/ml) as measured by W spectroscopy (using ~28~ = 16 600 per protomer). in 10T3 M phosphate buffer, pH 7.3, iOw3 dithioerytbritol was used. In _wme experiments the sample was made IO-’ h? IPTG. to insure saturation of the inducer binding site in the core. The film on the window was then put into a ceil, closed with a second window, and a l~uillidity of 98% R.H. was achieved by pIacing I ml of a saturated salt solution of lead nitrate Pb(h’0~)~ in a teflon cylinder in the bottom of the cell. This teflon cylinder could be easily and quickly exchanged without opening the cell by having a new cylinder displacing the first one. The cell was placed in an IR 12 infrared spectrophotometer_ A similar celf, without any fIItn was placed in the reference beam. The cells were thermostated by water circulation at 29 * 2OC as measured with a thermistor in close contact with the window supporting the sample. The first spectrum from 1000 to 4000 cm -I _in the initial hydrogenated state. was run when ~quilibriunl of temperature and water content (as measured using the 3400 cm-* band of water) had been obtained (usually after 1 or 2 hours). At zero time. the teflon cylinder containing the salt saturated water solution was exchanged for saturated heavy water solution previously equilibrated to the same temperature. The absorbance of the film was recorded at fled frequency of 1550 cm-‘” (Amide II band) as a function of time. From time to time spectra were run from 1500 to 1660 cm-’ during the Mchange experiment_ A complete spectrum from 1000 to 4000 cm-l was run in the final deuterated state. The absorption measurements have been corrected
for non linearity of the refernce comb and the absorbance uncertainty was only 0.003 absorbance units. The core fifms were controfled to present rather good constant thickness, without any holes in the beam. The spectral slit width used was 5 cm-’ aliowing ne&$ble noise_ The film, containing about 1.5 mg of protein, contains about an equivalent quantity of water. as shown by the water absorbance at 3400 cm-’ _ On the other hand, the internal volume of the cell is about 10 ml which contains at 79*C about 0.3 mg water vapour at 98% R-H_ Thus the total initial amount of light water is negligible in terms of the quantity of heavy water (- 1 g) which is placed in the cell with the teflon cylinder. The exchange from light to heavy water in the vapour phase is very fast (halftime about 20 s) as measured by the disappearance of the 1560 cm-l water vapour absorption in control experiments_ The half time of the eschange between the absorbed water molecule in the thin protein film and the heavy water from the vapour phase, is between 1 and 10 minutes as seen by disappearance of the 3400 cm-l water absorption band in control experiments. This is also the order of mapitude of the initially observed peptide hydrogen exchange rate at the beginning of the experiment, as calculated from the exchange curves (see results). The rate constant for the exchange of the fastest hydrogens, in the case of random polypeptides is theoretically given by [3,5,13J: kn =: 50(10-pH
+ lo~H-6.0)X
lOd.O5(e--Q
min- t _
(1) This gives with 8 = 29°C and pH 7.3 a values of k. = 2.9 X IO3 min-1 , i.e. a halftime life of 2.5 X loo4 min. Thus the exchange of this class of proton cannot be observed by this method. since it is faster than the exchange of the water molecules between the film and the vapour phase.
3. Results and discussion Fig. I shows the exchange curves for the repressor core, in the presence and in the absence of IPTG. These curves represent the absarbance ratio A(t)/d(O) at 15.50 cm-’ as a function of time where A(r) is the absorbance and A(O) the initial absorbance. This ratio is a
c
b
T (mid
Fig. 1. aj Hydrogen C~CIIJII~~ data for frc~ (s) and for IPTG saturated (-) E. coli iac rcprcssor cox Kclntivc nbsorbnncc d(f)/A(O). which is the ratio of the absorbance at 1550 cm-* at time r. ACT), to the initial absorbance, A(O), versus time T: For figure clarity we show only a few of the csperimental points we used for calculations. Full lines rcprcscnt the curves calculated from parameters in table 1 (four-class analysis). b) and c). X 5 magnified deviations of cxpcrimcntal points from cakulated curves, b) for free and c) for IPTG saturated
rcprcssor
core.
measurement of the (tro - H~)/~z~ fraction (2), 12~ being the total number of peptide NH hydrogens in
the initial state, and jr1 the number of these hydrogen atoms exchanged for deuterium at time t. We have analyzed the exchange curves as a sum of first order terms (2): “0
-
___
?It =
“0
5 Lfi j=]
exp(-kjt),
“0
(2)
where ki is the exchange rate of a class containing )ti hydorgen atoms, and N the number of such classes. We used a computer successive substraction method for
this calculation: the weigJtted least square best fit of the last experimental points. on a semi logarithmic scale, gives the size fzJ /rro and the exchange rate k 1 for the slowest hydrogen class. Then after substraction of this class from tire experimental data. we obtained sizes and rates for the N classes by iterative techniques of tlte same calculation. Since Laikcr and Prinz [14] have sltown that a same exchange curve could be fitted by several computed curves, wit11 different numbers IV of classes, we also used the more soplristicated Provenchcr’s eigenfunctions expansion method [ 151 for the same calculations. This method permits objective determination of the number JV of classes. The results given by Provencher’s “discrete“ calculation program agree very well, within the errors limits. with the four-class analysis of the successive substractions method. which is valid here because the well separated eschange rate values of the different hydrogen classes. Exchange curves are shown on fig. 1 and results of the calculation are shown in table 1_ The calculated values for the half-times of tire two fastest classes in the four-class analysis are lower than JO minutss. i.e. in the order of the D,O penetration half-time in the film. For this reason we think that the separation of these two frrstest classes is an artifact. Control esperimerits shows that the sizes of these two classes aie not reproducible. whereas the sum of their sizes is reproducible. So, we will consider them thereafter to form only one class with an unknown half-time shorter than 10 minutes. St is clear from table I that the experimental data for free and IPTG saturated repressor core are best fir by very similar parameters within the experimental error. when only three reproducible classes are considered, i.e. that binding of IPT’G to the repressor core does not affect the exchange law. As often made by several authors [7,16] _ we can try to relate the size of the various classes to the a helical, p sheet and random part fractions of the protein. However, it is necessary to make clear that this relation is only tentative since it is well known that the secondary structure is only one among the elements which determine the esclzange rate of a residue [3,17,18]. Analysis of the circular dichroism spectrum of the Iac repressor core by Chou et al. [19] gave a content of
Siac
Protoll~
7% (NW
165--
210
(7.0) (‘) (1.1)“)
(73)
7.5 f 15 --
(940)
1~~~~1~~11~~ rate! (mine’)
(0.14)
>7 x 10-2
(O&94x
_
“r2)
0.8 x 10-2 - 1.2 x 10-2 __I
(0.74 x ‘CPf
0.5 x 10-3 - 1.1 x m-3 --
. ..__ _.._..___ _
6x’ - ‘400 Cl
-. . . ...
Il~lS.~iln~ b, (ah)
-_. _ _. . . __... - ..-.- _^.___....^_.. . _. _ - . .
(0.61)
t
(491
(0.16) GiX
37 - 62 --
(71)
(0.23)
16% t 4% -
12 - 80
.__.-.-- .^
nernbcr II/
23% L 3%’ --
____.-__-“.__..
llj/llo
-
Siac
_.
-
.__..
-
._ .
-
(0. I 8) 4 (0.36) (‘1
(OS4)
54’/& ?: 5g -_-
(0.22)
22% t 3% ..__-
(024)
24% t 2% --
_.
“j//l 0
_._._“.
-.__-
“66)
‘SOp
(68) ‘80
5x _ 17 P
(74)
68 - 80 -
liXCllillt~l! mte
ftnin-‘1
(0,201
<7 x to-2 -_
(I,‘2 x ‘o-2)
I,0 x lo-” - 1.4 x 10-2 --.
_ - ..__._... .. ._.- -_.---
(7.9) 1’) (I .a, @I
4’0 -
(62)
GO+ IO --
800 - 1400 1’) 0.5 x 10-3 - 0.9 x lo-J _I_ .(0,68X IO-“) (I 020)
. . __.. _.._ -__ _._.___.--
'l~i's~~'lll~ b)
(ruin)
- .- _ _.__ _ __.
l’ro~ol~s iulrnbcr llj
_.-.
.___.____..____“. _ .._. _^“. _. _ ._ -_... _. .----_.. -. .. “_“..____ ____-_____._____.__ IW Inc rc’wssor core Inc rcprcssor cure + WIG _____ _ . .___- . . .._......... ...... _,._..- ..."-....-_--. - ._.. ______.. -.. _. ..._.-_ .._.._.--._.__w-.-
(SCI: two.
nf Vslws iit ‘)~cntllc~es arc those ~~l~~~ii,t~~lfor llic best I’it. Tlw ~~~i(lcrli~~~~i VillWS give rnngc 0r ~lil~~rl;lin~y. b, T’N Ml:the /Ii2 *ISrclalcd lo the csclmgc iTIlL! Isby lllcrcfatioi~sllip: P2 = 0.693/k. c, Tllc prcswcc in this class of some mom slowly cxcli:uigin~ liydrogws c;imiot lx csclhxi sirice l'iewiswiwnts IlilVCbeen inarlc rlwiilg 700 min. only. d)TltcsO vducs corrcspotitl lo it I’our chss ;in:dysis, filling tlic cspcrirrtcnl;il curve 1~ 0.01 rchtivc ;hsorb;aicc units, bill with lack of’cspcrirncntd rcprodacibiliw
-__-_-.-.
or i=3 i-4
Fast 1=3
1-2
Slow
Very slow i= I
-_
ClflSS I
-_.__-__.
Tllblc 1
1600
155G
cm-’
1
0
For the second slow class a31d for the third class the situation is 31ot so clear, since tlte IR spcctru311 of the random regions of proteins is not well know31 and is probably different for different Proteins. The difference spectrum of amide II band. bet\vcen 20 and 50 minutes is InrgeIy accounted for by the second class and shows maxima at about 1545 a31d I535 cm -1 which may be due to ?he t-1bonded NH groups in Bsheet 22 i--21] _ IIowevcr it is possible that this class contains so3ne sofven t-zxposcd o-helical regions and buried random regions of the protein f2,3.5] _ SirniIarIy the third, fast class cannot be attributed only to random parts of the protein and probably includes some II bcr31ded exposed hydroge31s of the pepticc units.
4.
1500
IS50
cm-’
1
0
pig. 2 a) Infrared absorption spectrum of the E. roli lar xcprcssor core (saturated -4th IPTG). I in tbc initial_ bfdtozcnnted state, L?nftcr 20 mitt cschangc, _3 after 50 min, 4 after 130 min, -5 after 1330 min. b) Difference spectrum, I between 10 and 50 min eschangc, accountinS mz~inly for tb~second, slow class, E between 130 and 1330 min, accounting mainI) for the first, wry slow class. Free rcprcssor core spectra are not diffcrcnr from the above ones.
o-helix between I6 and 28%. Using a similar anafysis but with an other set of reference spectra, we cafculeted an Q:helical content of 28% (Szabo and Maurizot. to be published). The method of conformational prediction of Chou and Fasman [20] indicates a value of 29% for o-helical content when applied to the lac repressor core f I93 _AH these vaiues are in agreement with our results, assuming that the very slow class of protons corresponds to the o( helix regions of the protei31. To confirm this point, the difference spectrum between the protein in its maximum deuterated state (after 1330 min) and protein at the end of tI1e previous class (I 30 min) was drawn in the amide II region between 1500 and 1600 cm-’ _ It is in good agreement with known spectra of a-helical proteins in H70 [2 I 231 with maxima at about 154.5 and 1515 ch’ further supporting our assignment.
Conclusions
Our results show that the hydroge33 to deutcriun1 exchange methods can be applied to hydrated films in order to study the conformations of proteins. Tiie analysis of the exchange curve in terms of classes ~331 give 331 appro.ximate esti3nate of the rt helical co33te31t of the protein_ In the case of lac rcprcssor core we found an cr helical content of approximately 25% i31good aglee3ncnt with other analyses and predictions [ I9,35]_ Our results show that the cscha31ge curve does 31ot vary upon binding the inducer IPTC. This indicates that the inducer bi33ding does 31ot cause a large COIIformational change in the core part of rhe kc rcpressor. These results are in agree331ent with these of Ramstein et al. [26] obtained by the tritium hydrogen exchange using the tritium tracer method. They support the idea that the effect of the inducer bindi31g to the core is n1ainIy related to the interaction between the ‘headpiece” and the core of the Zac repressor+
Acknowledgements Thiswork was supported by a grant fro3n the DCtegation GBndraIe H la Recherche Scientifique et Technique (MRXI 79-7-O 150).
References
[ 11 K. Liiderstrom-Lanp,
CR. Trav. Lab. Carlsberg i5 (1964) 7. f?] A. fivirff nod SO. Niclscn, in: rldvzx~c~s in protein chemistry, edr. CB. Anfinscn Jr., X1-L. Anson. J.T. J:dsaJt and F.&f. Richards (Aad. Press. New-York. 1966) p_ 3S7. 131 SW. Enplander, N.W. Downer and 11. Tcitelbaum, Ann. Rev. Uiochcmistry 4 1 (1972) 903_ 141 E.R. Blotat, C. dc Lox and A. Asadourian, J. Am. Chem. sot. 63 (1961) JS9S. 151 S.W. Jlngtandcr and A. Poufscn. Biopolymcrs 7 (i969) 379. 161 S.\V. 15xgltlnder and C. Matrcl, J. Bid. Chetn. 247 (1972) 2537. 171 S.W. Engltlndcr and K, SWcy. J. Mol. Biol. 45 (1969) 177. 181. S.W. f:ngJandcr and 3.J. 15n&mdcr. in: Mcrbods in cnzymoiogy, -XXVI. part C. cds C.H.W’ Jfirs and S.N. Tim~slteff (Acad. Press. Nc\v-York, 1972) p_ 406. 191 N. Geisler and 6. Wcbcr, FEBS Letters 87 (1978) 215. 1101 T. Platt. J.G. Files and K. Wcbcr. J. Bioi. Chem. 248 (1973) 110. [ 1 I] B. Mullcr-Hill, K.T. Bcyrouthcr and W. Gilbert, in: Xlctbods in enzymology, XXI, part D, cds. L. Grossman and Zc. Motd;rvc
N. Gcislcr and K. Wcbcr, Biochemistry 16 (1977) 938. A. Hvidt and K. WaJJevik, J. Biol. CJlem. 247 (1972) 1530. S.L. L&ken and M.P. Prinz, Biochemistry 9 (1970) 1547. 5.1%‘.Provenchcr, J. Chem. Phys. 64 (1976) 2772. 3%. Nakmishi and M. Tsuboi. Biochim. Biophys. Acta 434 (I 976) 365. B.D. Hilton and C-K. Woodward, Biochemistry 7 (1978) 3325. G. Wager and K. Withrich. J. Mol. Biol. 134 (1979) 75. P.Y. Chou, A.J. Adler and G.D. J&man, J. Mol. Biol. 96 (J 975) 29. P-Y. Chou and G.D. Fasman. Biochemistry 13 (I 974) 773 ___. S-N. Timadtcff and H. Susi, J. Biol. Chem. 241 (1966) 249. JJ. S::si, S.N. Timashcff and L. Stevens, 1. Biol. Chcrn. 242 (1967) 5460. J-1‘ Susi, in: Methods in enzymology, vol. XXVI, part C, cds. C.H.W. HJrs and S.N. Timnshcff (Acad. Press, NcwYork, 1972) p. 455. T. Xliyazawa and E.R. Btout, J. Am. C&m_ Sot. 83 (1961) 712. j2.5) S. Bourgeois, R.L. Jerni,ran, SC. Szu, EA. Habat and T.T. Wu, Biopolymers 18 (1979) 3625. 126) J_ Ramstcin, hl. Charlier, J.C. Maurizot, A.G. Szabo and C. JJeJcne, Biophys. Biochem. Res. Comm. 88 (1979) 124.
HYDROGEN APPLICATION
EXCHANGE
IN HYDRATED FILMS OF PROTEINS. LAC REPRESSOR CORE
TO THE E. COLI
Received 18 March 1980
An o@inal easy method of hydrogen to deuterium cschange in hydrated films of proteins, followed by infrared absorption measurements, is described and appiicd to films of the E. c&i lac repressor core, in order to csnmine the cffcct of isopropyl-p-D-thiogalnctoside (IPTG) binding. An estimation of about 255 (1 helical structure in this protein fraemcnt is deduced from the exchange curve. The bindinp of IPTG to the core does not affect the exchange curve within the cspcrimental error limits.
I _ Introduction
The hydrogen exchange rates of peptide NH groups are known to be very sensitive to protein conformations_ They depend upon hydrogen bonding in Q helix or p sheets, and upon hydrophobic and steric environment of the !abile protons in the molecule [l-3] _ Some difficulties in hydrogen exchange techniques, when measured by infrared absorption spectroscopy on solutions arise from the necessity to dissolve quickly (which is often no possible) the lyophilized protein in D,O, and to introduce this solution into a cell which is only a few microns thick in a very short period of time. The very fast exchange of some protons at neutral pH, with half-life of the order of one second [4,5] does not allow extrapolation of the exchange curve to zero time_ Thus the protein absorbance in the initial undeuterated state is not known without indirect estimation. A similar problem exists in determining the total number of labelled protons at zero time in tritium-hydrogen exchange experiments which are measured by gel filtration and radioactivity counting [2,6-S]. Furthermore, in this latter exchange-out method, very slow exchanging hydrogens are very difficult to label except * Present address: Division of Biological Sciences, N.R.C.C.. Ottawa, Canda KlA 0R6.
if the incubation time is increased or if the conditions of incubation are more drastic (for exatnple heating the solution). We describe here a simple method to follow the hydrogen-deuterium exchange of peptide N-H groups by infrared absorption measurements of the amide 11 absorbance at 1550 cm -I. in hydrated films of proteins. when exposed to heavy water DzO vapour. This method allows one to determine the initial point of the exchange curve, and does not require any lyophilization of the protein, as well it avoids the difficulties of infrared spectroscopy on dilute solutions. In this work we describe how we have used these techniques to study the Zac repressor core and its interaction with an inducer_ Under certain conditions proreolytic cleavage of the lac repressor occurs between residues 50 and 60 yielding two fragments: the N-terminal headpiece. which contains the major portion (if not all) of the DNA binding site, and a tetrameric core with full inducer binding activity [9,10]_ We were interested in studying the effect of an inducer, in this work isopropyl-P-D-thiogalactoside (IPTG), on this core. in order to obtain further information on its mechanical action. This core has the advantage of a greater solubility than the intact protein, allowing the formation of concentrated homogeneous films of the proiein and their study by infrared absorption spectroscbpy.
2. Materials and methods E c&i lac repressor from strain E&aH 493 (a generous gift of Dr. Beyreuthcr and Dr. ~~~IIer-Hill) was purified by ammonium sulfate fractionation and column ~llro~nato~aplly on phosphocelfulose as described by Miiller-Hill et al. [ 1 I 1. Proteolysis was performed as described by Geisler and \Veber [ 121 using clostripain QVorthfngton). The core preparation was analyzed on SDS polyacrylamide gel and run as a sin& band corresponding to molecular weight about 32 000 accounting for residues 60 to 360. Films of repressor core were deposited on calcium fluoride (CaF2) IR windows by gentle drying at ambient temperature and humidity. 1 ml of a solution 4.7 X 10e5 M core (i.e. 1.5 mg/ml) as measured by W spectroscopy (using ~28~ = 16 600 per protomer). in 10T3 M phosphate buffer, pH 7.3, iOw3 dithioerytbritol was used. In _wme experiments the sample was made IO-’ h? IPTG. to insure saturation of the inducer binding site in the core. The film on the window was then put into a ceil, closed with a second window, and a l~uillidity of 98% R.H. was achieved by pIacing I ml of a saturated salt solution of lead nitrate Pb(h’0~)~ in a teflon cylinder in the bottom of the cell. This teflon cylinder could be easily and quickly exchanged without opening the cell by having a new cylinder displacing the first one. The cell was placed in an IR 12 infrared spectrophotometer_ A similar celf, without any fIItn was placed in the reference beam. The cells were thermostated by water circulation at 29 * 2OC as measured with a thermistor in close contact with the window supporting the sample. The first spectrum from 1000 to 4000 cm -I _in the initial hydrogenated state. was run when ~quilibriunl of temperature and water content (as measured using the 3400 cm-* band of water) had been obtained (usually after 1 or 2 hours). At zero time. the teflon cylinder containing the salt saturated water solution was exchanged for saturated heavy water solution previously equilibrated to the same temperature. The absorbance of the film was recorded at fled frequency of 1550 cm-‘” (Amide II band) as a function of time. From time to time spectra were run from 1500 to 1660 cm-’ during the Mchange experiment_ A complete spectrum from 1000 to 4000 cm-l was run in the final deuterated state. The absorption measurements have been corrected
for non linearity of the refernce comb and the absorbance uncertainty was only 0.003 absorbance units. The core fifms were controfled to present rather good constant thickness, without any holes in the beam. The spectral slit width used was 5 cm-’ aliowing ne&$ble noise_ The film, containing about 1.5 mg of protein, contains about an equivalent quantity of water. as shown by the water absorbance at 3400 cm-’ _ On the other hand, the internal volume of the cell is about 10 ml which contains at 79*C about 0.3 mg water vapour at 98% R-H_ Thus the total initial amount of light water is negligible in terms of the quantity of heavy water (- 1 g) which is placed in the cell with the teflon cylinder. The exchange from light to heavy water in the vapour phase is very fast (halftime about 20 s) as measured by the disappearance of the 1560 cm-l water vapour absorption in control experiments_ The half time of the eschange between the absorbed water molecule in the thin protein film and the heavy water from the vapour phase, is between 1 and 10 minutes as seen by disappearance of the 3400 cm-l water absorption band in control experiments. This is also the order of mapitude of the initially observed peptide hydrogen exchange rate at the beginning of the experiment, as calculated from the exchange curves (see results). The rate constant for the exchange of the fastest hydrogens, in the case of random polypeptides is theoretically given by [3,5,13J: kn =: 50(10-pH
+ lo~H-6.0)X
lOd.O5(e--Q
min- t _
(1) This gives with 8 = 29°C and pH 7.3 a values of k. = 2.9 X IO3 min-1 , i.e. a halftime life of 2.5 X loo4 min. Thus the exchange of this class of proton cannot be observed by this method. since it is faster than the exchange of the water molecules between the film and the vapour phase.
3. Results and discussion Fig. I shows the exchange curves for the repressor core, in the presence and in the absence of IPTG. These curves represent the absarbance ratio A(t)/d(O) at 15.50 cm-’ as a function of time where A(r) is the absorbance and A(O) the initial absorbance. This ratio is a
c
b
T (mid
Fig. 1. aj Hydrogen C~CIIJII~~ data for frc~ (s) and for IPTG saturated (-) E. coli iac rcprcssor cox Kclntivc nbsorbnncc d(f)/A(O). which is the ratio of the absorbance at 1550 cm-* at time r. ACT), to the initial absorbance, A(O), versus time T: For figure clarity we show only a few of the csperimental points we used for calculations. Full lines rcprcscnt the curves calculated from parameters in table 1 (four-class analysis). b) and c). X 5 magnified deviations of cxpcrimcntal points from cakulated curves, b) for free and c) for IPTG saturated
rcprcssor
core.
measurement of the (tro - H~)/~z~ fraction (2), 12~ being the total number of peptide NH hydrogens in
the initial state, and jr1 the number of these hydrogen atoms exchanged for deuterium at time t. We have analyzed the exchange curves as a sum of first order terms (2): “0
-
___
?It =
“0
5 Lfi j=]
exp(-kjt),
“0
(2)
where ki is the exchange rate of a class containing )ti hydorgen atoms, and N the number of such classes. We used a computer successive substraction method for
this calculation: the weigJtted least square best fit of the last experimental points. on a semi logarithmic scale, gives the size fzJ /rro and the exchange rate k 1 for the slowest hydrogen class. Then after substraction of this class from tire experimental data. we obtained sizes and rates for the N classes by iterative techniques of tlte same calculation. Since Laikcr and Prinz [14] have sltown that a same exchange curve could be fitted by several computed curves, wit11 different numbers IV of classes, we also used the more soplristicated Provenchcr’s eigenfunctions expansion method [ 151 for the same calculations. This method permits objective determination of the number JV of classes. The results given by Provencher’s “discrete“ calculation program agree very well, within the errors limits. with the four-class analysis of the successive substractions method. which is valid here because the well separated eschange rate values of the different hydrogen classes. Exchange curves are shown on fig. 1 and results of the calculation are shown in table 1_ The calculated values for the half-times of tire two fastest classes in the four-class analysis are lower than JO minutss. i.e. in the order of the D,O penetration half-time in the film. For this reason we think that the separation of these two frrstest classes is an artifact. Control esperimerits shows that the sizes of these two classes aie not reproducible. whereas the sum of their sizes is reproducible. So, we will consider them thereafter to form only one class with an unknown half-time shorter than 10 minutes. St is clear from table I that the experimental data for free and IPTG saturated repressor core are best fir by very similar parameters within the experimental error. when only three reproducible classes are considered, i.e. that binding of IPT’G to the repressor core does not affect the exchange law. As often made by several authors [7,16] _ we can try to relate the size of the various classes to the a helical, p sheet and random part fractions of the protein. However, it is necessary to make clear that this relation is only tentative since it is well known that the secondary structure is only one among the elements which determine the esclzange rate of a residue [3,17,18]. Analysis of the circular dichroism spectrum of the Iac repressor core by Chou et al. [19] gave a content of
Siac
Protoll~
7% (NW
165--
210
(7.0) (‘) (1.1)“)
(73)
7.5 f 15 --
(940)
1~~~~1~~11~~ rate! (mine’)
(0.14)
>7 x 10-2
(O&94x
_
“r2)
0.8 x 10-2 - 1.2 x 10-2 __I
(0.74 x ‘CPf
0.5 x 10-3 - 1.1 x m-3 --
. ..__ _.._..___ _
6x’ - ‘400 Cl
-. . . ...
Il~lS.~iln~ b, (ah)
-_. _ _. . . __... - ..-.- _^.___....^_.. . _. _ - . .
(0.61)
t
(491
(0.16) GiX
37 - 62 --
(71)
(0.23)
16% t 4% -
12 - 80
.__.-.-- .^
nernbcr II/
23% L 3%’ --
____.-__-“.__..
llj/llo
-
Siac
_.
-
.__..
-
._ .
-
(0. I 8) 4 (0.36) (‘1
(OS4)
54’/& ?: 5g -_-
(0.22)
22% t 3% ..__-
(024)
24% t 2% --
_.
“j//l 0
_._._“.
-.__-
“66)
‘SOp
(68) ‘80
5x _ 17 P
(74)
68 - 80 -
liXCllillt~l! mte
ftnin-‘1
(0,201
<7 x to-2 -_
(I,‘2 x ‘o-2)
I,0 x lo-” - 1.4 x 10-2 --.
_ - ..__._... .. ._.- -_.---
(7.9) 1’) (I .a, @I
4’0 -
(62)
GO+ IO --
800 - 1400 1’) 0.5 x 10-3 - 0.9 x lo-J _I_ .(0,68X IO-“) (I 020)
. . __.. _.._ -__ _._.___.--
'l~i's~~'lll~ b)
(ruin)
- .- _ _.__ _ __.
l’ro~ol~s iulrnbcr llj
_.-.
.___.____..____“. _ .._. _^“. _. _ ._ -_... _. .----_.. -. .. “_“..____ ____-_____._____.__ IW Inc rc’wssor core Inc rcprcssor cure + WIG _____ _ . .___- . . .._......... ...... _,._..- ..."-....-_--. - ._.. ______.. -.. _. ..._.-_ .._.._.--._.__w-.-
(SCI: two.
nf Vslws iit ‘)~cntllc~es arc those ~~l~~~ii,t~~lfor llic best I’it. Tlw ~~~i(lcrli~~~~i VillWS give rnngc 0r ~lil~~rl;lin~y. b, T’N Ml:the /Ii2 *ISrclalcd lo the csclmgc iTIlL! Isby lllcrcfatioi~sllip: P2 = 0.693/k. c, Tllc prcswcc in this class of some mom slowly cxcli:uigin~ liydrogws c;imiot lx csclhxi sirice l'iewiswiwnts IlilVCbeen inarlc rlwiilg 700 min. only. d)TltcsO vducs corrcspotitl lo it I’our chss ;in:dysis, filling tlic cspcrirrtcnl;il curve 1~ 0.01 rchtivc ;hsorb;aicc units, bill with lack of’cspcrirncntd rcprodacibiliw
-__-_-.-.
or i=3 i-4
Fast 1=3
1-2
Slow
Very slow i= I
-_
ClflSS I
-_.__-__.
Tllblc 1
1600
155G
cm-’
1
0
For the second slow class a31d for the third class the situation is 31ot so clear, since tlte IR spcctru311 of the random regions of proteins is not well know31 and is probably different for different Proteins. The difference spectrum of amide II band. bet\vcen 20 and 50 minutes is InrgeIy accounted for by the second class and shows maxima at about 1545 a31d I535 cm -1 which may be due to ?he t-1bonded NH groups in Bsheet 22 i--21] _ IIowevcr it is possible that this class contains so3ne sofven t-zxposcd o-helical regions and buried random regions of the protein f2,3.5] _ SirniIarIy the third, fast class cannot be attributed only to random parts of the protein and probably includes some II bcr31ded exposed hydroge31s of the pepticc units.
4.
1500
IS50
cm-’
1
0
pig. 2 a) Infrared absorption spectrum of the E. roli lar xcprcssor core (saturated -4th IPTG). I in tbc initial_ bfdtozcnnted state, L?nftcr 20 mitt cschangc, _3 after 50 min, 4 after 130 min, -5 after 1330 min. b) Difference spectrum, I between 10 and 50 min eschangc, accountinS mz~inly for tb~second, slow class, E between 130 and 1330 min, accounting mainI) for the first, wry slow class. Free rcprcssor core spectra are not diffcrcnr from the above ones.
o-helix between I6 and 28%. Using a similar anafysis but with an other set of reference spectra, we cafculeted an Q:helical content of 28% (Szabo and Maurizot. to be published). The method of conformational prediction of Chou and Fasman [20] indicates a value of 29% for o-helical content when applied to the lac repressor core f I93 _AH these vaiues are in agreement with our results, assuming that the very slow class of protons corresponds to the o( helix regions of the protei31. To confirm this point, the difference spectrum between the protein in its maximum deuterated state (after 1330 min) and protein at the end of tI1e previous class (I 30 min) was drawn in the amide II region between 1500 and 1600 cm-’ _ It is in good agreement with known spectra of a-helical proteins in H70 [2 I 231 with maxima at about 154.5 and 1515 ch’ further supporting our assignment.
Conclusions
Our results show that the hydroge33 to deutcriun1 exchange methods can be applied to hydrated films in order to study the conformations of proteins. Tiie analysis of the exchange curve in terms of classes ~331 give 331 appro.ximate esti3nate of the rt helical co33te31t of the protein_ In the case of lac rcprcssor core we found an cr helical content of approximately 25% i31good aglee3ncnt with other analyses and predictions [ I9,35]_ Our results show that the cscha31ge curve does 31ot vary upon binding the inducer IPTC. This indicates that the inducer bi33ding does 31ot cause a large COIIformational change in the core part of rhe kc rcpressor. These results are in agree331ent with these of Ramstein et al. [26] obtained by the tritium hydrogen exchange using the tritium tracer method. They support the idea that the effect of the inducer bindi31g to the core is n1ainIy related to the interaction between the ‘headpiece” and the core of the Zac repressor+
Acknowledgements Thiswork was supported by a grant fro3n the DCtegation GBndraIe H la Recherche Scientifique et Technique (MRXI 79-7-O 150).
References
[ 11 K. Liiderstrom-Lanp,
CR. Trav. Lab. Carlsberg i5 (1964) 7. f?] A. fivirff nod SO. Niclscn, in: rldvzx~c~s in protein chemistry, edr. CB. Anfinscn Jr., X1-L. Anson. J.T. J:dsaJt and F.&f. Richards (Aad. Press. New-York. 1966) p_ 3S7. 131 SW. Enplander, N.W. Downer and 11. Tcitelbaum, Ann. Rev. Uiochcmistry 4 1 (1972) 903_ 141 E.R. Blotat, C. dc Lox and A. Asadourian, J. Am. Chem. sot. 63 (1961) JS9S. 151 S.W. Jlngtandcr and A. Poufscn. Biopolymcrs 7 (i969) 379. 161 S.\V. 15xgltlnder and C. Matrcl, J. Bid. Chetn. 247 (1972) 2537. 171 S.W. Engltlndcr and K, SWcy. J. Mol. Biol. 45 (1969) 177. 181. S.W. f:ngJandcr and 3.J. 15n&mdcr. in: Mcrbods in cnzymoiogy, -XXVI. part C. cds C.H.W’ Jfirs and S.N. Tim~slteff (Acad. Press. Nc\v-York, 1972) p_ 406. 191 N. Geisler and 6. Wcbcr, FEBS Letters 87 (1978) 215. 1101 T. Platt. J.G. Files and K. Wcbcr. J. Bioi. Chem. 248 (1973) 110. [ 1 I] B. Mullcr-Hill, K.T. Bcyrouthcr and W. Gilbert, in: Xlctbods in enzymology, XXI, part D, cds. L. Grossman and Zc. Motd;rvc
N. Gcislcr and K. Wcbcr, Biochemistry 16 (1977) 938. A. Hvidt and K. WaJJevik, J. Biol. CJlem. 247 (1972) 1530. S.L. L&ken and M.P. Prinz, Biochemistry 9 (1970) 1547. 5.1%‘.Provenchcr, J. Chem. Phys. 64 (1976) 2772. 3%. Nakmishi and M. Tsuboi. Biochim. Biophys. Acta 434 (I 976) 365. B.D. Hilton and C-K. Woodward, Biochemistry 7 (1978) 3325. G. Wager and K. Withrich. J. Mol. Biol. 134 (1979) 75. P.Y. Chou, A.J. Adler and G.D. J&man, J. Mol. Biol. 96 (J 975) 29. P-Y. Chou and G.D. Fasman. Biochemistry 13 (I 974) 773 ___. S-N. Timadtcff and H. Susi, J. Biol. Chem. 241 (1966) 249. JJ. S::si, S.N. Timashcff and L. Stevens, 1. Biol. Chcrn. 242 (1967) 5460. J-1‘ Susi, in: Methods in enzymology, vol. XXVI, part C, cds. C.H.W. HJrs and S.N. Timnshcff (Acad. Press, NcwYork, 1972) p. 455. T. Xliyazawa and E.R. Btout, J. Am. C&m_ Sot. 83 (1961) 712. j2.5) S. Bourgeois, R.L. Jerni,ran, SC. Szu, EA. Habat and T.T. Wu, Biopolymers 18 (1979) 3625. 126) J_ Ramstcin, hl. Charlier, J.C. Maurizot, A.G. Szabo and C. JJeJcne, Biophys. Biochem. Res. Comm. 88 (1979) 124.