Synthesis and interaction of some dipeptide amides with deoxyribonucleic acid

MISSING IMAGE

Material Information

Title:
Synthesis and interaction of some dipeptide amides with deoxyribonucleic acid
Physical Description:
xi, 116 leaves : ill. ; 28 cm.
Language:
English
Creator:
Sheardy, Richard Dean
Publication Date:

Subjects

Subjects / Keywords:
DNA   ( lcsh )
Amides   ( lcsh )
Chemistry thesis Ph. D
Dissertations, Academic -- Chemistry -- UF
Genre:
bibliography   ( marcgt )
non-fiction   ( marcgt )

Notes

Thesis:
Thesis--University of Florida.
Bibliography:
Bibliography: leaves 109-115.
Statement of Responsibility:
by Richard Dean Sheardy.
General Note:
Typescript.
General Note:
Vita.

Record Information

Source Institution:
University of Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
aleph - 000094680
oclc - 06202783
notis - AAL0105
sobekcm - AA00004912_00001
System ID:
AA00004912:00001

Full Text











SYNTHESIS AND INTERACTION OF
SOME DIPEPTIDE AMIDES
WITH DEOXYRIBONUCLEIC ACID




BY


RICHARD DEAN SHEARDY


A DISSERTATION PRESENTED TO THE GRADUATE COUNCIL
OF THE UNIVERSITY OF FLORIDA IN
PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE
DEGREE OF DOCTOR OF PHILOSOPHY


UNIVERSITY OF FLORIDA


1979
























TO MOM AND DAD.

AND ARTIE.















ACKNOWLEDGEMENTS


The author extends his appreciation to Professor

Edmond J. "Eddie" Gabbay for his guidance, continuous

support and friendship during the past five years, until

the time of his death. He was a good friend and excellent

scientist -- and he will be missed very much.

Thanks are also extended to all those who have

helped in the author's graduate studies, especially

Mr. M. Adkins, Mr. B. Arvan, Mr. T. Baugh, Mr. C. Cromwell,

Mr. B. Griggs, Mr. D. King, Dr. R. King, Dr. D. McRitchie,

Dr. S. Pearce and Mr. T. Rigl. Special thanks also go to

Dr. M. Battiste, whose friendship and support during our

time of sadness this past summer is greatly appreciated.

Thanks also go to Dr. W. D. Wilson for being an outside

reader. Very special thanks to Ms. Joanne Upham, whose

kindness, patience, and love are immeasurable, for her

excellent typing.

The completion of this course of study would not have

been possible without the support and encouragement of the

author's family.







For Eddie


There was a man-
I knew him well.
He lived his life
With dreams to sell.
I know that some
Can't understand,
But dreams can oft
Turn into sand.

And now my eyes
Fill with sorrow,
For he will see
No tomorrow.

He was a friend-
I know that well;
A helpful hand-
A joke to tell.
He had to go-
As all things must;
To fly away
And turn to dust.

And though my tears
May dry so fast,
My mind will not
Forget the past.

What is a life-
A game to play
Until it's time
To fade away?
Or is it more-
A heart to give
To those we love
That they may live?

And though there's fear
Throughout the night,
The morning sun
Is shining bright.

Richard Dean Sheardy
October 17,1979















TABLE OF CONTENTS


Page

ACKNOWLEDGEMENTS iii

LIST OF TABLES vi

LIST OF FIGURES vii

ABSTRACT x

CHAPTER I INTRODUCTION 1

Structure of DNA 2
Physical Properties of DNA 9
Evidence for Hydrogen Bonding 13
Stacking Forces 17
Electrostatic Effects 20
Electronic Effects 20
Dynamic Structure of Nucleic Acids 23
Chromatin 24
Small Molecule-Nucleic Acid Interactions 25
Statement of Problem 30

CHAPTER II RESULTS AND DISCUSSION 31

H NMR Studies 31
Viscosity Studies 40
Ultraviolet Absorption Studies 46
Binding Studies 46
Circular Dichroism Studies 52
Equilibrium Dialysis 52
Melting Temperature Studies 56
Discussion 58

CHAPTER III EXPERIMENTAL 69

Synthesis 70
Analytical Methods 101

REFERENCES 109

BIOGRAPHICAL SKETCH 116














LIST OF TABLES


Table Page

1. The Effect of Increasing Length of the
Polynucleotide (Ap) A on the UV Absorp-
tion Spectrum. 20

2. Absorption and Hypochromicity Data on
the Dipeptide Amides. 47

3. Typical Output for McGhee-von Hippel
Binding Isotherms for Various Values
of n. 49

4. Binding Isotherms for DNP and Nitrotolyl
Containing Dipeptides. 51

5. Apparent Binding Affinity (Ka) of the
Dipeptides to Salmon Sperm DNA. 55

6. The Effect of the Dipeptides on the ATm
on the Helix-Coil Transition of Salmon
Sperm DNA, Poly I Poly C, and Poly
d(A-T). 57















LIST OF FIGURES


Figure Page

1. Schematic Representation of the Watson-
Crick Double Helix of DNA. 3

2. Watson-Crick Base Pairs. 4

3. Keto-Enol Tautomers for the Bases Guanine
and Cytosine. 4

4. Structure of a Section of a DNA Chain. 6

5. Absorption-Temperature Profile for DNA. 10

6. Intrinsic Viscosity-pH Profile for DNA. 10

7. Acid-Base Titration Curve for DNA. 11

8. a) Hoogsteen Adenine-Thymine Base Pair.
b) Anit-Hoogsteen Adenine-Thymine Base Pair. 14

9. Shielding of the Aromatic Protons Caused
by Vertical Stacking. 18

10. Geometry of Stacked Nucleosides. 19

11. Exciton Splitting of Energy Levels. 21

12. Intensity Interchange Between Two Inter-
acting Transition Moments. 22

13. Reporter Molecule I. 28

14. Schematic illustration of a segment of DNA
double helix which can either partially
intercalate a molecule or fully intercalate
a molecule. 28

15. Schematic illustration of a DNA segment
showing a possible mechanism whereby the E
and a amino groups of the N-terminal L-lysyl
residue are stereospecifically anchored and
thus dictating the positioning of the aro-
matic ring of the diastereomeric dipeptides
in the DNA-peptide complex. 29


vii









Figure


16. The Dipeptide Amides Synthesized and
Studied. 32

17. The Partial NMR Spectra of DNA at
Various Temperatures. 33

18. H NMR Signal of the Aromatic Protons of
Dipeptides 1 and 2 in the Presence and
Absence of DNA at Various Temperatures. 34

19. 1H NMR Signal of the Aromatic Protons of
Dipeptides 3 and 4 in the Presence and
Absence of DNA at Various Temperatures. 35

20. 1H NMR Signal of the Aromatic Protons of
Dipeptides 5 and 6 in the Presence and
Absence of DNA at Various Temperatures. 36

21. 1H NMR Signal of the Aromatic Protons of
Dipeptides 7 and 8 in the Presence and
Absence of DNA at Various Temperatures. 37

22. H NMR Signal of the Aromatic Protons of
Dipeptides 9 and 10 in the Presence and
Absence of DNA at Various Temperatures. 38

23. The Effect of Dipeptides 1 and 2 on the
Relative Specific Viscosity of Salmon
Sperm DNA. 42

24. The Effect of Dipeptides 3, 4, 7 and 8
on the Relative Specific Viscosity of
Salmon Sperm DNA. 43

25. The Effect of Dipeptides 5 and 6 on the
Relative Specific Viscosity of Salmon
Sperm DNA. 44

26. The Effect of Dipeptides 9 and 10 on the
Relative Specific Viscosity of Salmon
Sperm DNA. 45

27. A Plot of v/L Versus v From Data in
Table 3. 50

28. The CD Spectra of Dipeptides 5, 6, 9
and 10 in the Absence and Presence of DNA. 53

29. The CD Spectra of Dipeptides 3 and 4 in
the Absence and Presence of DNA. 54


viii


Page









Figure Page

30. Liquid HF Apparatus. 74

31. Diagram of the Viscometer with the Photo-
electric Device in Place. 103














Abstract of Dissertation Presented to the
Graduate Council of the University of Florida
in Partial Fulfillment of the Requirements
for the Degree of Doctor of Philosophy


SYNTHESIS AND INTERACTION OF
SOME DIPEPTIDE AMIDES
WITH DEOXYRIBONUCLEIC ACID

By

Richard Dean Sheardy

December 1979


Chairman: Edmond J. Gabbay (deceased), Merle A. Battiste
Major Department: Chemistry


This thesis discusses the synthesis and interaction

specificities of ten dipeptide amides with salmon sperm

DNA. For each dipeptide, the N-terminal amino acid was

an L-lysyl residue, while the C-terminal amino acid had

an a carbon of defined chirality (either L or D) and an

aromatic moiety a certain distance from the a carbon. The

rationale for this series was to determine if intercalation

or partial intercalation is a function of both the chirality

of the C-terminal amino acid a carbon as well as the dis-

tance of the aromatic moiety from the a carbon. The ex-

tent of intercalation was determined by a variety of tech-

niques: 1) 1H NMR studies; 2) viscometric titration; and

3) UV absorption and CD studies. It was found that partial









intercalation occurs with the peptide containing a C-ter-

minal L-pNO2-phenylalanine residue with no intercalation

for its diastereomeric dipeptide amide, indicating a

dependence on the chirality of the C-terminal amino acid.

Six of the remaining dipeptides fully intercalated, sug-

gesting the necessity to evaluate electronic and hydro-

phobic effects on intercalation. The two diastereomeric

dipeptide amides containing an S-p-nitrobenzylcysteinyl

moiety gave conflicting results. Possible explanations

for the above phenomenon are provided.














CHAPTER I
INTRODUCTION


Undoubtedly, one of the greatest advances of biological

science in this century was the determination of deoxyribo-

nucleic acid fiber structure by Watson and Crick in 1953.1,2

Since that time, volumes of work have been published at the

expense of many man hours to either confirm or disprove

certain theoretical aspects of the Watson-Crick structure.

Certainly, knowledge of DNA structure may lead to the

treatment of inherited diseases, or those diseases caused

by viruses and bacteria, and ultimately to find cures for

cancer. This knowledge is not easy to obtain due to the

immense complexity of the genetic code.

It has been suggested that a single chromosome may be

just one DNA molecule having a molecular weight of up to

4 x 10 daltons and an overall length of over 2 centi-

meters.3 Unfortunately, it is difficult to isolate a com-

plete DNA molecule since its structure is: 1) sensitive to

changes in temperature4'5 and pH;6 and 2) subject to breakage

under minimal shearing forces.7

Commercially available DNA is usually derived from calf

thymus or salmon testes. Also, numerous synthetic DNA's are

available with definite sequence and molecular weight. The

concentration of a nucleic acid solution is determined from









the absorption of the bases at 260 nm using, for example, a

molar extinction coefficient of E = 6,500 for salmon sperm

DNA,8 and has the dimensions of moles of phosphate per liter.


Structure of DNA

Watson and Crick proposed a right-handed double helix

(Figure 1), comprised of two anti-parallel sugar-phosphate

backbones held together by specific hydrogen bonds between

the complementary bases and by hydrophobic forces that

favor stacking of the bases.1,2 Their model was based on

X-ray diffraction data on DNA fibers as well as the chemical

evidence of Chargaff and Lipschitz.9 Due to the low resolu-

tion of X-ray analysis of fibers, much of the evidence came

from the chemical approach.

Chargaff and Lipschitz used a variety of naturally

occurring DNA to determine the relative number of purines

and pyrimidines.9 They found that: 1) the amount of

guanine (G) is equal to the amount of cytosine (C); 2) the

amount of adenine (A) is equal to the amount of thymine (T);

and 3) the A-T/G-C ratio is constant for a particular species.

The base pairing scheme of Watson and Crick, seen in Figure

2, accounts for these observations.

The free bases can exist in one of two tautomeric

forms as shown in Figure 3 for guanine and cytosine.

However, in the Watson-Crick scheme, the bases are in their

keto forms which allow for two hydrogen bonds between the

A-T base pair and three hydrogen bonds between the G-C base















0
11 A


maj or
groove


0
7A .o


minor
groove















Figure 1. Schematic Representation of the Watson-Crick
Double Helix of DNA. The outer helical strands
represent the sugar-phosphate backbone, while
the horizontal lines represent the base pairs
and the vertical line is the helical axis.













major groove


*CH3


minor groove


G-C


minor groove

A-T


Figure 2.


Watson-Crick Base Pairs.


H2N N


NH




I
N |

O^


Figure 3.


Keto-Enol Tautomers for the Bases Guanine and
Cytosine.


major groove









pair. Watson and Crick used molecular models to fit the

base pairing scheme to the dimensions of the molecule as

obtained from X-ray analysis. They found that an A-G base

pair was too large to fit these dimensions, whereas a T-C

pairing was too small. In their scheme (i.e., A-T and G-C

base pairing), it was found that the distance between the

sugar phosphate backbone for each set of base pairs was

identical, lending further support to their base pairing

scheme.

The individual strands of DNA are enzymatically joined

monomer nucleoside phosphates resulting in an alternating

sugar-phosphate-sugar backbone in which the bases are

stacked on top of one another (Figure 4). The D-deoxyribose

sugar, in the furanoside form, is numbered as shown in

Figure 4 and has two hydroxyl groups at the 3' and 5' posi-

tions, respectively, and the base at the 1' position. The

two complementary strands are placed anti-parallel in order

to attain maximum symmetry. In other words, one strand has

its sugar-phosphate backbone directed 3' 5', while the

other backbone is directed 5' 3'. This aspect of the

double helix was proven correct by Josse et al. in 1961 using

a nearest neighbor analysis.10

The linkage between two successive sugars is formally

a phosphate diester. At neutral pH, the phosphate is mono-

anionic and thus, the oxygens are directed away from the

helix into solution. Futhermore, the two oxygens are not

equivalent since one lies parallel to the helical axis

















o-dHO/ 'O

3" 2 NH2



-- --0 -CHN

oO N


-0-P
11
0


-0--P -0
11
0


0


Figure 4. Structure of a Section of a DNA Chain.









(axial) whereas the other oxygen lies perpendicular to the

helical axis (equatorial).

Langridge and coworkers used X-ray diffraction data to

determine that the double helix makes one complete turn
o 11
every 34 A; this is known as the pitch. With ten base
0
pairs per pitch, there is a translation of 3.4 A between

successive base pairs. Since each turn of the double helix

requires 360 of rotation, the angle between successive

base pairs must be 36.

Inspection of Figures 1 and 2 reveals that two grooves,

one large (major) and one small (minor), are formed as a

consequence of the twist of the double helix. As a result

of the N-9 purine glycosidic bond and the N-l pyrimidine

glycosidic bond, the sugar phosphate backbones on each base

lie on the same side of each base pair, thus giving rise to

two distances between the backbone. As seen in Figure 2,

the third hydrogen bond of the G-C base pairs is located in

the minor groove, while the methyl group of thymine is lo-

cated in the major groove. These two structural features

affect the binding of small molecules to DNA.

X-ray work with DNA requires the use of DNA in its

fiber state, which does not allow enough definable data

points to give a definite arrangement of the atoms. Usually,

molecular models are built and fitted to the data, allowing

for elimination of those models which are inconsistent with

the data or stereochemically unfeasible. This process even-

tually leads to a model that is consistent with the data.









Conclusions from this type of analysis must be taken

with caution. For example, X-ray analysis of DNA fibers

drawn from different media (i.e., Li+ K+ Na+ or Mg salts)

have yielded slightly different structures for the same DNA
11
molecule. Also, at relative humidity above 80%, X-ray

data suggest that the DNA is in the B form, which has a
O
pitch of 34 A with 10 residues per turn with the base pairs

perpendicular to the helical axis. However, the X-ray

studies of this fiber cannot distinguish between a left- or

right-handed helix.2 Below 80% relative humidity, the DNA

structure changes to the A form, in which the base pairs are

tilted 15-20 from the perpendicular and the helical pitch
O
is lowered to 28 A with 11 bases per turn. X-ray analysis

of the A form shows it to be a right-handed helix, and from

this, it can be assumed that the B form is also right-

handed. It has been suggested that there are at least eight

different structural forms of natural DNA (i.e., A, B, C,

P, P2' J, J2 and S),13 all varying in the angle of the tilt

of the base pairs from the perpendicular and the pitch of

the helix. Furthermore, it has been shown that the A-T/G-C

ratio will cause a change in DNA structure, suggesting that

the secondary structure of DNA is a function of the primary

structure.1

It is apparent that the structure of DNA in the fiber

state is sensitive to many variables and thus the conclu-

sions of X-ray analysis can be questioned. In fact,

Donahuel4'15 and Arnottl6 have suggested that X-ray








techniques not be used. Furthermore, extrapolation from

the fiber structure to that of solution can only be done if

one assumes that there is no structural change of the DNA

molecule upon solvation. In fact, Bram used the low angle

X-ray scattering technique on DNA in solution to support

the contention that the fiber and solution structures of
17 18
DNA are different. 17,


Physical Properties of DNA

It was mentioned earlier that DNA is a delicate mole-

cule whose structure is subject to breakage due to shearing

forces and is sensitive to changes in temperature and pH.

For example, if the absorption at 260 nm is monitored as the

temperature of a DNA solution is gradually increased, a
4,5
sigmoidal curve, as shown in Figure 5, will be observed.4'

Concomitantly, the molar ellipticity and viscosity

versus temperature profile of the DNA solution also shows

a dramatic change. These observations can be explained in

terms of a molecular transformation from a double helix to

a random coil. This helix-coil transition is known as de-

naturation (and/or "melting out") and is usually irrever-

sible, indicating that permanent alteration of the native

DNA structure has occurred upon melting.

The structure of DNA is affected by hydrogen-ion concen-

tration as depicted in the viscosity-pH profile of a DNA

solution in Figure 6.6 In the pH range of 4.5 to 11.5, the

viscosity of the solution remains relatively constant,


















260
I
I



I

Tm
T

Figure 5. Absorption-Temperature Profile for DNA.





40



30



[n] 20



10




2 4 6 8 10 12
pH

Figure 6. Intrinsic Viscosity-pH Profile for DNA.









indicating no major structural changes. However, at pH

below 4.5 or above 11.5, the viscosity decreases rapidly,

indicative of a "melting out" of the DNA structure. There-

fore, DNA solutions are always prepared in a buffer whose

pH lies between 6 and 8.

A titration curve for DNA is shown in Figure 7.19

Curve A represents the titration starting at pH 6.9 and

titrating with either acid or alkali. Curve B is the back

titration starting from either pH 2.5 or 12.5 to neutral.

It is apparent that: 1) DNA has both acidic and basic ti-

tratable groups; and 2) the titration curves are not rever-

sible. The back titrations indicate that weakly acidic

and basic groups not available in the first titration are

available in the second (denatured DNA).



3
0 /
o 2
1 Cd B /A

1
0

0
A B


1 1 A / B




3 / I

2 4 6 8 10 12
pH

Figure 7. Acid-Base Titration Curve for DNA.









The DNA structure (i.e., native or denatured) is also

dependent upon salt concentration. Thomson et al.20 have

shown that DNA in solution is irreversibly denatured when
-3
the sodium chloride concentration falls below 10 M or when

the magnesium chloride concentration falls below 105 M.

Due to the irreversibility of the denaturation, care must

be taken in preparing DNA solutions to keep the ionic

strength at a minimum of 1 mM.

The hydrogen bonding capacity and the dielectric con-

stant of the solution medium are also important for main-

taining the DNA structure. Lower alcohols (i.e., methanol

and ethanol) cause reversible denaturation of DNA, while

organic solvents such as formamide, DMF or DMSO cause
21
irreversible denaturation.21 Since DNA is not soluble in

the above solvents, these experiments are performed by

adding the organic solvent to an aqueous DNA solution.

It is obvious from the above studies that the stability

of the DNA double helix is subject to many types of forces.

An understanding of how these forces influence the structure

of DNA in solution is prerequisite to gain insight into the

mechanisms by which other molecules interact with DNA.

Therefore, analysis of these systems must be at the molecular

level. Since DNA is a very large molecule, classical chem-

ical approaches, which are successful for small molecules,

are not always feasible when studying macromolecules. The

usual practice is to construct model systems using monomer

or oligomer units and to extrapolate the data to the larger






13


molecules. Some of the more revealing studies in this re-

gard are now examined with the intention of specifying the

forces that control the structure of nucleic acid in solu-

tion.


Evidence for Hydrogen Bonding

Watson and Crick proposed that the hydrogen bonding

scheme depicted in Figure 2 is mandatory to account for

the base pairing interactions, as well as to help maintain

the double helical structure. '2 It has been demonstrated

that, based on the geometrical restraints of colinear

hydrogen bonds (i.e., A-H B) and H-bonding distances
0
of 2.80 3.00 A, 29 different base pairs connected by two

or three hydrogen bonds could be formed between the four

nucleosides found in nucleic acids.22,23 However, only

three distinct schemes have been found in the crystalline

state. Calculations based on dipole-dipole interactions by

Nash and Bradley have shown that only the Watson-Crick scheme

is favored for guanine-cytosine base pairing, but adenosine-

uracil pairs can exist in three forms with two of these forms
24
of approximately equal energy.24 In fact, for adenine and

thymine derivatives, two different hydrogen bonding schemes

have been found, neither of which is the Watson-Crick type.

Hoogsteen postulated the hydrogen bonding scheme in Figure 8a
25,26
for cocrystals of 9-methyladenine and 1-methylthymine.2

Since the sugar-phosphate backbones would be closer together

in a Hoogsteen type base pairing than in the Watson-Crick











/"----- ,~
H,

-H CH 3




R
(a)



0 CH
R

R -H N

H-
N N/ (b)
\H

Figure 8. a) Hoogsteen Adenine-Thymine Base Pair.
b) Anti-Hoogsteen Adenine-Thymine Base Pair.


type, the authors suggested that the Watson-Crick scheme

is favored due to less electronic repulsions between the

strands. This type of hydrogen bonding has also been ob-

served for 9-ethyladenine-l-methylthymine cocrystals.27

Moreover, cocrystals of adenosine and 5-bromouridine have

been found to assume a different hydrogen bonding scheme:

the anti-Hoogsteen type (Figure 8b).28,29 This type of

hydrogen bonding scheme has also been observed for other

systems.30,31 On the other hand, X-ray analysis of single

crystals of derivatives of guanine and cytosine has shown

that only the expected Watson-Crick base pairing is oper-
ative in these base pairs.32-34 This is presumably due to
ative in these base pairs. This is presumably due to









the fact that the pairing scheme for guanine-cytosine has

three hydrogen bonds, whereas the others (Hoogsteen and anti-

Hoogsteen) have only two. Thus, in the crystalline state,

only three out of a possible twenty-nine base pairing com-

binations have been observed.

The question naturally arises as to what happens in

solution. Tuppy and Kueblercovalently attached a specific

nucleoside to an Amberlite support and used this modified
35
support in a column. It was found that when a solution

of the other three nucleosides was passed through this

column, only the complement to the covalently bound nucleo-

side was retarded. Furthermore, this association was

abolished when the column was eluted with an aqueous-urea

solution, strongly suggesting that the retention of the

complementary nucleoside was due to hydrogen bonding for-

mation.

Infrared spectroscopy has been used to determine

hydrogen bonding between derivatized purines and pyrimidines

by monitoring the appearance of new absorption bands be-
-I
tween 3500 and 3000 cm due to N-H stretching frequency

changes when this group is involved in hydrogen bonding.36

When solutions of 9-ethyladenine and 1-cyclohexyl-uracil

were mixed, two new bands at 3490 and 3330 cm-1 appeared

and were maximized when the components were mixed at a

ratio of 1:1. Similar results have been obtained with de-

rivativesof guanine and cytosine.37-39









The 1H NMR has also been used to investigate hydrogen

bonding between bases in nonaqueous solution by monitoring

the downfield chemical shift of the hydrogen bonded N-H pro-

ton due to a decrease in the electron density about the pro-

ton nucleus upon hydrogen bonding. Katz and Penman found

that: 1) base pairing decreases as the hydrogen bonding

ability of the solvent increases; and 2) in nonhydroxylic

solvents, the base pairing is quite specific.40 For

example, the downfield chemical shift of the N-l proton of

guanosine was -134.7 Hz in the presence of the complemen-

tary cytosine, whereas, there was relatively little or no

downfield shift in the presence of the other nucleosides.

A similar observation was seen for the N-3 proton of uridine

in the presence of adenine.41 Shoup et al. have also em-

ployed an NMR study of the four purine and pyrimidine com-

ponents of DNA and have obtained similar results.42

The above studies reveal that hydrogen bonding between

base pairs is quite specific in nonhydroxylic solvents.

However, in aqueous media, competition for the hydrogen

bonding sites on the bases by water would minimize the im-

portance of these forces. Thus, one might conclude that

hydrogen bonding forces are not significant in terms of

stabilizing the double helical structure. This dilemma is

resolved by the realization that the interior of the double

helix, in which the base pairs are located, is very non-

polar. The exclusion of water from this region due to the

hydrophobic nature of the bases provides an excellent








environment for the formation of hydrogen bonds between the

bases.


Stacking Forces

It was suggested as early as 1958 that hydrogen

bonding between bases was probably not the sole source of

DNA stability.43,44 This was based on the observation that

at a pH low enough to break hydrogen bonds, the DNA helix

could be kept intact,providing a sufficiently low temper-

ature was maintained. Further work has proposed that a

major stabilizing factor for DNA is a hydrophobic inter-

action between the bases. Since the bases are essentially

nonpolar, they tend to stack in aqueous media. Since the

stacked bases are less hydrated than the individual bases,

the phenomenon of stacking is thought to be entropy driven.45

Ts'O has determined the osmotic coefficients for

several purine and pyrimidine systems and calculated

activity coefficients by the Gibb-Duhem relationship. The

data clearly indicated that the properties of the bases and

nucleosides in solution were far from ideal, indicating

extensive interaction of the nucleosides in aqueous solu-

tion. Hydrogen bonding was ruled out for two reasons:

1) the nucleosides associate much more extensively than

urea, which is known to be one of the best hydrogen bonding

agents in water; and 2) methylation of hydrogen bond donor

sites on the nucleosides enhances association.

A more direct method to study the association of bases

and nucleosides in solution is through NMR. Chan et al.








found that as the concentration of purine increased, the

resonances of the three protons of purine became in-

creasingly shifted to higher fields.46 From the H NMR

theory on ring current anisotropy (Figure 9), it is

possible to predict that the association results from

vertical stacking of the purine bases.47







H0




H


Figure 9. Shielding of the Aromatic Protons Caused by
Vertical Stacking.


From the magnitude of the 1H NMR chemical shifts for

the various protons of adenosine, it was found that the

five-membered ring experiences less shielding than the six-

membered ring. Broom et al. proposed two models (Figure 10)

for the preferred average orientation of the stacked

nucleoside bases to account for these observations.48

It has been found that substitution with an amino

group on the 6 position of purine to give adenosine en-

hances association. Furthermore, 5-bromo uridine asso-

ciates to a greater extent than thymidine. These obser-

vations lead to the conclusion that a hydrophobic force








R




(a) Alternate
R







(b) Straight R
R
Figure 10. Geometry of Stacked Nucleosides.


is not solely responsible for the stacking interaction.4

It has been suggested that there is a correlation between

the stacking and the polarizability of the bases. The con-

clusion by Hanlon that London dispersion forces are respon-

sible for the stability of the DNA helix is in line with
49
this concept.4

Further insight into stacking interactions has been

obtained from studies of dinucleoside phosphates. Using

H NMR, Chan and Nelson showed that ApA exists in a 3'-

anti-5'-anti right-handed stack.50 It should be noted

that the Watson-Crick model requires that the bases be in

the anti conformation. It has also been observed that

different dinucleoside phosphates differ in their stacking

ability.51 For example, ApA's structure persisted upon

elevation of temperature, whereas UpU lost its ordered

structure. This has been attributed to different stacking

interactions for the two dinucleoside phosphates.









Electrostatic Effects

At neutral pH, the backbone phosphate groups of the

DNA double helix are negatively charged, creating a poly-

anion. Therefore, at low ionic strength, there is a large

interstrand Coulombic repulsion between the "naked" anionic

phosphates. There are also intrastrand phosphate inter-

actions. Thus, at low ionic strength, there is an increased

propensity for the DNA to denature. Gabbay has determined

that the distance between adjacent phosphate groups along

the same chain of the Watson-Crick double helix is approxi-
o 52
mately 7 A. As the ionic strength of a DNA solution in-

creases, the intrinsic viscosity decreases (the intrinsic

viscosity is related to the length of rod-like DNA). It

has been suggested that the decrease in viscosity is re-

lated to a decrease in the length of rod-like DNA due to

less inter- and intrastrand repulsions of the phosphates

as they are neutralized by added electrolyte.53


Electronic Effects

The fact that the bases interact electronically is

well known. Examination of Table 1 reveals that as the


Table 1. The Effect of Increasing Length of the Pol nucleo-
tide (Ap)nA on the UV Absorption Spectrum.r4

Smax/monomer
n max max

0 260 15,000
1 257 13,600
2 257 12,600
3 257 11,300
4 257 10,800
300 (poly A) 256 9,000








length of the molecule increases, there is a concomitant

decrease in the extinction coefficient of the adenine bases

as well as a blue shift in the wave length maximum.54'55

A theoretical treatment, employing the exciton theory,

was used to explain the results. 58 The theory states

that for a molecular crystal, the absorption of light is

not limited to a single molecule, but rather, it is dis-

tributed over many. Since a polynucleotide can be con-

sidered a one-dimensional crystal, the exciton theory

can be applied to nucleic acids. According to the theory,

the excited states of the bases would be split into, say,

two new levels (Figure 11). The higher energy level has

the electronic vectors of the bases parallel, whereas the

lower energy level has the vectors anti-parallel. Quantum

mechanical selection rules forbid transitions to the lower


Figure 11. Exciton Splitting of Energy Levels.








excited state. Therefore, the allowed transition to the

higher energy level is one of energy higher than the origi-

nal uncoupled transition and results in a blue shift for

the absorption maximum.

The observed hypochromism is a result of an intensity

interchange of coupled transition moments as illustrated in

Figure 12. Since the transition moments are in a card

stacked orientation in a DNA molecule, the higher energy

transition is hyperchromic, while the lower energy transi-

tion is hypochromic.


(a) Card Stack Arrangement

VA < B
A is hypochromic

B is hyperchromic

A B
(b) Head-to-Tail Arrangement


A B
S-) vA < VB

A is hyperchromic

B is hypochromic
(c) Herringbone Arrangement



A
No intensity interchange

B


Figure 12. Intensity Interchange Between Two Interacting
Transition Moments.









Studies utilizing circular dichroism have also shown

the theory to be correct. The selection rules for cir-

cular dichroism allow both transitions to occur, resulting

in both a red shifted band and a blue shifted band relative

to the absorption maximum of the monomer. However, without

more complicated quantum mechanical considerations, the

signs of the CD bands cannot be predicted.60 It should be

noted, though, that the two bands are of opposite signs,

resulting in a distinctive double cotton effect for the

polynucleotide.


Dynamic Structure of Nucleic Acids

All of the above factors (i.e., hydrogen bonding,

stacking forces, electronic and electrostatic forces) con-

tribute to various extents to the overall stability of the

double helical structure. However, the double helical

structure is not a static one, but rather a dynamic one.

For example, von Hippel and coworkers incubated nucleic

acids in tritiated water, and after separation of the poly-

nucleotide from radioactive solvent via gel filtration,

found that tritium was incorporated into the nucleic acids

with kinetics of exchange much faster than with protein

systems.61-63 It was also shown that the observable ex-

change took place with those protons involved in the base-

base hydrogen bonding of the double helix. This led

von Hippel and McConnel to propose three models to account

for the phonemenon: 1) unstacking without hydrogen bond









breakage; 2) hydrogen bond breakage without unstacking; and

3) hydrogen bond breakage with partial or complete unstacking

and strand separation.6263 Based on tritium exchange and

pH studies, it was concluded that the "breathing model" (3)

is correct. Kinetic studies of intercalation by Muller and

Crothers64 and Gabbay et al.65 agree with this hypothesis.


Chromatin

As stated before, one DNA molecule may be up to 2 cm

long if fully extended.3 However, a typical animal cell
3 66
nucleus has a diameter of 5 pm and a volume of %65 pm .

The question obviously arises as to how such a long mole-

cule fits into such a small package. The answer lies in

the substituents of the nuclear material, chromatin.

Chromatin contains 15% DNA, 10% RNA and 75% protein. Some

of this protein contains enzymes or repressors, but a por-

tion is what is known as histones. Histones are rich in

basic amino acid and are tightly associated with DNA pre-

sumably by neutralization of the negatively charged phos-

phates. There are 5 classes of histones with molecular

weights in the range of "11,000 21,500.67-69 They are:

1) H1 (lysine rich); 2) H2a and H2b (moderately lysine

rich); and 3) H3 and H4 (arginine rich).

Electron micrographs of chromatin reveal a regular

repeating structure resembling a string of beads. Each

bead is 7-10 nm in diameter and there is a spacer section

of DNA 2-14 nm between each bead.68'70-73 Each bead









contains approximately 200 base pairs of DNA folded around

a histone octamer containing H2a, H2b, H3 and H4. This

would effectively condense a 68 nm DNA chain into a 10 nm

bead. It has thus been concluded that histones act to fold
74
DNA so that it will fit into the nucleus.7

Of course, histones are not the only proteins which

interact with DNA. Repressor proteins bind to specific

sections of DNA to prevent transcription of a particular

gene. It is still not clear how these interactions take

place.


Small Molecule-Nucleic Acid Interactions

The recognition process between protein systems and

nucleic acids has generated much research in the past ten

years. However, it is not feasible to study the interac-

tion specificity between two macromolecules due to the

immense complexity of these interactions. A more sim-

plified approach involves the use of model systems whereby

the interaction specificity of small molecules with DNA

is considered. With the results of these studies, one can

extrapolate to more complex systems.

There are several forces operative when a small mole-

cule binds to DNA. These interactions involve electrostatic

interaction between positively charged groups on the mole-

cule and the anionic backbone phosphates, hydrogen bonding

between available hydrogen bonding sites on the molecule and

the functional groups of the purine and pyrimidine bases,









and hydrophobic forces. Three types of hydrophobic-type

interactions have been noted: 1) intercalation between
75-79
base pairs of DNA by aromatic cations; 75 2) "partial"

insertion between base pairs by sterically restricted com-

pounds containing an aromatic moiety;78'8084 and 3) ex-

ternal hydrophobic-type binding as exemplified by steroidal

amine-nucleic acid complexes.8586 It has been proposed

that the nucleic acids may use all or any combination of

the above forces to specifically bind polypeptide chains.87

Intercalation of aromatic amino acids with double and

single stranded nucleic acids was observed with L-phenyl-

alanine, L-tryptophan, and L-histidine using poly A. The 1H

NMR studies showed that the H-2 and H-8 protons of adenine

were downfield shifted due to the destacking of the adenine

bases as a result of intercalation of the aromatic ring of

the amino acids.88 Poly A is also destacked and inter-

calated by derivatives of tyramide, tyrosine and tryptamine

since it was found that the aromatic protons of the latter

were upfield shifted in the poly A complex due to ring
89-91
current anisotropy. 91

Brown observed a very large stabilizing effect on the

Tm of DNA in the presence of arginyltryptophan methyl ester

when compared to other similarly charged dipeptide deriva-

tives.92 He therefore suggested that the aromatic rings of

amino acids behave as "bookmarks" which anchor the proteins

to specific sequences of nucleic acids via an intercalation

mode of binding.









In order to investigate the extent of insertion of an

aromatic ring between base pairs of DNA, Gabbay and Kapicak

examined the interaction specificity of reporter molecule I
78
(Figure 13) with DNA.78 They found that when n = 1, there

is a decrease in viscosity of the DNA-reporter molecule (due

to shortening of the DNA helix), whereas there was an in-

creased viscosity of the DNA-reporter complex for n = 2, 3

and 4 (due to an increase in the effective length of the

DNA helix). Furthermore, 1H NMR studies reveal that the

aromatic protons HA and HB (in DNA-1 complex when n = 1)

experience upfield chemical shifts of 10 and 6 Hz, respec-

tively, but that the aromatic protons of the DNA-1 complex

when n = 2, 3 and 4 are completely broadened and indis-

tinguishable from baseline noise.

It was therefore proposed that the p-nitrophenyl ring

of I (when n = 1) is partially inserted between base pairs
78
of DNA (Figure 14).78 This model accounts for: 1) the

decrease in the effective length of DNA-I caused by bending

at the point of insertion; and 2) the different upfield

chemical shifts of the aromatic protons HA and HB (i.e., HA

experiences a larger ring current anisotropy than HB as a

result of "partial" insertion. Presumably, the single

methylene between the aromatic moiety and the quaternary

ammonium group on the side chain is not sufficient to allow

for "full" insertion and lengthening of the helix.

Further evidence for "partial" intercalation came from

experiments using the diastereomeric dipeptides













+F

O


0 0


U)o X

0


'0 'O\1 'O '0

I I I I -


t)


C)
I-I









U
u


C)
Cd
4-J

nrl
U cH
*HU





.i-I






QO



co








0
o su




















rlCd U
4-) -H



C) 0









(/) 0Uc
C/) U C






~
E cb
b0k
0 a),
<^ i) -
cf r





0~









Er-lP,

^-i (U
U < r-
*H ^!


H
(-"3

(N




U
u

+Z
b0



U
CO




+ ;:
t^
="




SN!
V-5


0 0 CD ,









H H
+ I I
NH -C-CONH-C-CONH?

(+H2)4 C
+NH3


L-lys-L-pheA


Figure 15.


H CH2
+ I
NH -C-CONH-C-CONH2


+NI.3


L-lys-D-pheA


Schematic illustration of a DNA segment showing
a possible mechanism whereby the E and a amino
groups of the N-terminal L-lysyl residue are
stereospecifically anchored and thus dictating
the positioning of the aromatic ring of the
diastereomeric dipeptides in the DNA-peptide
complex.








L-lysyl-L-phenylalanine amide (L-lys-L-pheA) and L-lysyl-

D-phenylalanine amide (L-lys-D-pheA). 84 In particular,

the NMR, viscometric, and flow dichroism data suggest that

the aromatic ring of L-lys-L-pheA is partially inserted

between the base pairs of DNA, whereas the aromatic ring

of L-lys-D-pheA points out toward the solvent (Figure 15).

It was concluded that this specificity arises from a

stereospecific interaction of the e- and a-amino groups of

the N-terminal L-lysyl residue which dictates the posi-

tioning of the aromatic ring of the C-terminal phenyl-

alanine residue of the L-lys-L-pheA and L-lys-D-pheA. The

specificity observed with L-lys-L-pheA and L-lys-D-pheA

cannot be attributed to the chirality of the phenylalanine

moiety itself since IH NMR studies reveal no significant

differences in the binding of L- and D-phenylalanine amides

to DNA.


Statement of Problem

In this dissertation, the synthesis of several dipep-

tide amides was undertaken to determine the effect of the

chirality of the C-terminal amino acid a carbon and the

length of the "arm" between the a carbon and the aromatic

ring on the extent of intercalation of the aromatic ring

(used in this case as a probe) between the base pairs of

DNA.














CHAPTER II
RESULTS AND DISCUSSION


The interactions of the dipeptide amides (Figure 16)

with DNA has been examined by pulsed Fourier transform 1H

NMR, viscosity, UV absorption and binding studies, circular

dichroism, equilibrium dialysis, and melting temperature

(Tm of helix-coil transition) studies. The results are pre-

sented below.

H NMR Studies

The 1H NMR studies conducted in this work were used

to determine the effect of binding of the peptide to DNA

on: 1) the line broadening of the aromatic proton NMR

signals of the peptide; and 2) the upfield chemical shift

of the aromatic protons of the peptide. The change in

chemical shifts is related to the proximity of the aromatic

protons to the ring current of the DNA base pairs. The 1H

NMR line broadening can be explained by several mechanisms:

1) slow rate of exchange between various DNA binding sites;

2) restricted tumbling of the aromatic ring; 3) larger dif-

ferences in the chemical shifts experienced by the ortho,

para, and meta protons; or 4) a combination of all three

mechanisms. In these experiments, the peptides are fully

bound as determined by equilibrium dialysis. The results

















Na
C







M Z
2:



C =
I Ci

u--5-

O z
L


0
N
O









oZ
O
=--t;--U


Z t
0

:---UL--Z


+2z


UZ
N










C,
S



2:


=--6--=--z +
=2

z o
= -- u- -


Or








-r-


:Z =
o u




U -
0 -I




+ z


z




Or




z
u O




Z


=-U---z +


+


N0
Z




0

o2


"r-- -U-0 r.-
Z -





+ 2




















1-1
s-; T


0




eq LI



z 9
O

':-L.-LI--z +













+Z




0.
0;

>>


0 o



Z 2








+Z + 2
2 : CM UIfl



2Z :2


z C
H H(




U 4-

o | ci
C r




d U
N C















*H*







4J 0
02 :l

























z<



z *


---
VI "
8-I
S-U 0 ,- (
= 6x)
g 2






P< !


*-I N|








of these studies are shown in Figures 17 22, and lead to

the following observations.

First, the extent of 1H NMR signal line broadening of

the aromatic protons depends in general on the chirality of

the C-terminal amino acid a carbon and the length of the

"arm" between the aromatic ring and the a carbon. For

example, the 1H signal of the aromatic protons of L-lys-

L-pNO2-pheA (1) are broadened more than those of the



















cc



b





i I I I i ___ __--- -1
8.0 7.0 ppm 2.5 1.5 ppm

Figure 17. The Partial NMR Spectra of DNA at Various
Temperatures. The temperatures are: a) 370C;
b) 70"C; and c) 900C.












































SI I I I i I I


8.0


7.0 ppm


8.0


7.0 ppm


L-lys-L-pNO2-pheA (1)


Figure 18.


L-lys-D-pNO2-pheA (2)


H NMR Signal of the Aromatic Protons of Dipep-
tides 1 and 2 in the Presence and Absence of
DNA at Various Temperatures. The DNA base pair
to dipeptide ratio is: a) 0 (370C); b) 10 (370C);
c) 7 (370C); d 5.5 (370C); e) 5.5 (700C); and
f) 5.5 (900C).











































a

7.0 ppm


L-lys-L-S-NT-cysA (3)


Figure 19.


I I l I
8.0 7.0 ppm

L-lys-D-S-NT-cysA (4)


1H NMR Signal of the Aromatic Protons of Dipep-
tides 3 and 4 in the Presence and Absence of
DNA at Various Temperatures. The DNA base pair
to dipeptide ratio is: a) 0 (370C); b) 15 (370C);
c) 10 (370C); d) 7 (370C); e) 7 (700C); and
f) 7 (900C).


8.0


b




















, fs


f



e






C


b


8.5 7.5 ppm 8.5 7.5 ppm


L-lys-L-DNP-DABA (5)


Figure 20.


L-lys-D,L-DNP-DABA (6)


H NMR Signal of the Aromatic Protons of Dipep-
tides 5 and 6 in the Presence and Absence of
DNA at Various Temperatures. The DNA base pair
to dipeptide ratio is: a) 0 (370C); b) 15 (370C);
c) 10 (370C); d) 7 (370C); e) 7 (700C); and
f) 7 (900C).


e













































7.0 ppm


8.0


L-lys-L-S-pNO2bz-cysA (7)


Figure 21.


L-lys-D-S-pNO2bz-cysA (8)


H NMR Signal of the Aromatic Protons of Dipep-
tides 7 and 8 in the Presence and Absence of
DNA at Various Temperatures. The DNA base pair
to dipeptide ratio is: a) 0 (370C); b) 15 (370C);
c) 10 (370C); d) 7 (370C); e) 7 (700C); and
f) 7 (900C).


8.0


7.0 ppm











f
f

e



....... ...


d

~---~I-c _c


Sb
b........


8.5 7.5 ppm 8.5 7.5 ppm


L-lys-L-DNP-ornA (9)


Figure 22.


L-lys-D-DNP-ornA (10)


H NMR Signal of the Aromatic Protons of Dipep-
tides 9 and 10 in the Presence and Absence of
DNA at Various Temperatures. The DNA base pair
to dipeptide ratio is: a) 0 (370C); b) 15 (370C);
c) 10 (370C); d) 7 (370C); e) 7 (700C); and
f) 7 (900C).


_ s-----"-"


............








diastereomeric L-lys-D-pNO2-pheA (2). The H NMR signal for

the aromatic protons of L-lys-L-S-NT-cysA (3), L-lys-D-S-NT-

cysA (4), L-lys-L-DNP-DABA (5), L-lys-D,L-DNP-DABA (6),

L-lys-L-DNP-ornA (9) and L-lys-D-DNP-ornA (10) are com-

pletely broadened into the base line and are indistinguish-

able from base line noise at all DNA base pair to peptide

ratios. For compounds L-lys-L-S-pNO2bz-cysA (7) and L-lys-

D-S-pNO2bz-cysA (8), there is considerable line broadening,

with perhaps slightly more broadening for peptide 7.

It should be noted that the line broadening decreases

at higher temperatures as the DNA begins to "melt-out,"

freeing bound peptide. Furthermore, the effect of the vis-

cosity of the DNA solution on the line broadening is found

to be small [e.g., the signal line width of internal stan-

dard 2,2,3,3-d4-3-trimethylsilylproprionate (TSP) is found

to be 1.2 + 0.2 and 2.1 + 0.3 Hz in the absence and presence

of DNA, respectively, at 370C]. In addition, examination

of Figure 17 reveals that some of the structure of the NMR

spectra in the region of 6 7.0 8.5 ppm is due to DNA base

protons.

Secondly, the extent of the upfield chemical shift of

the 1H NMR signal of the aromatic protons also depends on

the chirality of the C-terminal amino acid a carbon and the

distance from it to the aromatic ring. The spectra for

L-lys-L-pNO2-pheA (1) is considerably more complicated in

the DNA complex as opposed to free peptide. Examination of

Figure 18 shows two peaks at approximately45 Hz upfield








from the original peaks. Interpretation of this spectra is

difficult due to the complexity of the splitting. However,

for L-lys-D-pNO2-pheA (2), there is little or no upfield

shifting of the aromatic proton signals. The 1H NMR spectra

of L-lys-L-S-pNO2bz-cysA (7) and L-lys-D-S-pNO2bz-cysA (8)

also become more complex in the presence of DNA (Figure 21).

There is considerable upfield shifting of the aromatic sig-

nals for both compounds, with more shifting for compound 7

than 8. In fact, both spectra resemble that of L-lys-L-

pNO2-pheA (1).

Since the 1H NMR signals for the aromatic protons

of compounds L-lys-L-S-NT-cysA (3), L-lys-D-S-NT-cysA (4),

L-lys-L-DNP-DABA (5), L-lys-D,L-DNP-DABA (6), L-lys-L-DNP-

ornA (9), and L-lys-D-DNP-ornA (10) are completely broadened

into the baseline, nothing can be concluded about upfield

shifting of these protons in the DNA complex. However,

in all cases, the resolution of the signals increases as

the temperature increases, indicative of "release" of the

dipeptide from the DNA.complex upon denaturation of the

DNA.


Viscosity Studies

It has been shown by a number of investigators that

planar aromatic compounds (e.g., ethidium bromide, dauno-

rubicin, and the acridine dyes), intercalate between the

base pairs of DNA. The DNA complex increases in viscosity

relative to free DNA due to an increase in the effective








length of the DNA molecule. To prove that the DNA length

increases, the determination of the intrinsic viscosity,

(n), of the solution is necessary. Intrinsic viscosity

is defined by

SLimit Tsp
c 0 c


where nsp is the specific viscosity of a DNA solution and

c is the concentration of DNA. Unfortunately, these experi-

ments are uninformative since the peptide dissociates from

the DNA complex at high dilution (probably due to the low

binding constant of peptide to DNA) and thus, the intrinsic

viscosity of the complex approaches that of free DNA.

Therefore, the effect of increasing the concentration

of the dipeptide amides on the relative specific viscosity,

nsp /spo (where nsp and nspo are the specific viscosities

of the DNA solution in the presence and absence of dipeptide

amide, respectively) at 3.0 x 10-4 DNA phosphate/liter (in

10 mM MES, 5 mM Na+, pH 6.2) was studied at 370C using the

low shear Zimm viscometer. Since the study was carried out

at low DNA concentration, the relative values of n /n
sp spo
are close approximations of the relative values of the

intrinsic viscosity of a DNA-dipeptide complex to free DNA

([n]/[n]o). The results are shown in Figures 23 through 26,
and indicate that peptides 1, 2, 7 and 8 decrease the spe-

cific viscosity of a DNA solution, with peptide 1 decreasing

the viscosity to a much greater extent than 2, 7 and 8.

Furthermore, peptides 3, 4, 5, 6, 9 and 10 increase the












































0 0





< <
SC)

I I
o t
z Z
ft ft



0 0
^-1 1-


C X






--I














Lof

o


U
.r-j
.-1

a)

u)




--e




CO
rl








0














4-4
04-4

NI





0

6 LH




C)
,-l









t0 H










*H
rCi


0
Cl f
1/1 1































































000
OO


) M

u u
I I

I I

I I

I I

I I
in v)
>N >
r-l rl
l- I


00 o
co Co
O O
.


L-

'1-










o 4

4 a
13
*i-


C m
i- oC


0

na
P4
SA


0
pp
C:7
(T C


CS
'-1
Ln



1-4

Q


u
*H
'*H
U


U)












4-J













*H
in








Lr
























a, -
U
0

















































00
4)

P.CH





U>





^*H=










&0
*l-































































OO


\o|


. i











IZ
-Z I


1-1 1-
.-a ^-


00 \o f N oC 00 \
r-4 A rC O O

i- -


O
I to
tnl in
171- 7


Ln

N














N
4





-4

o
C--
In X











4-)
C) -
r4




T3

+->
0 CL


u
*.H
t4-4

u
U


-)










4
C)
r-i












LO


-I)




a)



Ul








LnH

u
O)










U3 *
r-












U

4-
0 4-
0,
























































0


< <

0 0
O OI

Z 2
I I

I I
tn to

I "!


N


LO






N




N



*
<
*0
'-l




Z




C)


*:14

'-4 PL


0
4-4
rH

U
u)

Cr
>
*I


ra
41-
ci






0

O












04Q
"r


So
ro








4-)
O


UK

tH o
O *

S3tZ

>
vrk
M f-i

Tl f
*ilCO














(U t/


to
'A i=









specific viscosity of a DNA solution, the effect being

greatest for 5, 9 and 10 with only a slight increase for

peptides 3 and 4.


Ultraviolet Absorption Studies

Absorption data for the dipeptides are shown in Table

2. All peptides have a well defined Amax in the range of

260 400 nm with extinction coefficients on the order of

8.5 x 103 for compounds 1, 2, 3, 4, 7 and 8 and 15.5 x 103

for compounds 5, 6, 9 and 10. For those compounds with

Xmax greater than 300 nm, two effects can be seen on the

absorption spectra upon binding to DNA. First, large

hypochromic effects are seen for all DNP derivatives

(namely 5, 6, 9 and 10), with a considerable hypochromic

effect for one of the nitrotolyl derivatives, peptide 3.

Furthermore, there is a large bathochromic shift for max

in 3 and 4 in the DNA complex (i.e., from 340 nm to 355 nm)

and a small bathochromic shift for the DNP derivatives 5,

6, 9 and 10 (i.e., from 360 nm to 365 nm).


Binding Studies

The binding of those peptides with max greater than

300 nm to salmon sperm DNA was determined by spectral

titration. The effect of increasing concentration of DNA

on the absorption of peptides 3, 4, 5, 6, 9 and 10 at 340

nm for the nitrotolyl containing peptides or at 360 nm for

the DNP containing peptides was studied in 6.6 mM phosphate

buffer (pH 7.2) at three ionic strengths (5 mM, 10 mM and












-0 1 n n n m c Lni Ln
L1 Z Z L n i Z Z Ln Ln







C) C) C) C) C) C)
r4Z-l i-I ri- r4

Z z x x x x z z x x






t ) L C) C) Ct t ) OC)





LO Ln L O 0 L L L )
If \D LI \.D \.O0
S 0 Z 0 0 V) t0)








C) C) CD CD CD C CD CD CD C
x -l r4 w-q r-4 r-- 4 i-l v- r1 rAl

X 8 x x x< x x x x x

:. 00 00 .0 L LI
IO) LI) if) I)
COO j. r- -i 00 00 r-4 r-4







Cd LO If) CD C) C) LI) L) C)
SLn Ln o 0o 0o N N. o o
e' (M C4 LI) VI) V) r1 tn t


Sn




0 l 4 I I
0 Z 1Z P

U z ) Qz i



.U. U) C)V (
t r!
M r -I
I s i ss- ^


* > co o o
01 01










z z zoI|
4 4 Z


>. >, s >,
! U U) U) U
ft o o, to to







H MH1 ^- -


*








tJ4-


> 0









(U


LI)a
4--

0
..0 P




bif



2H
U)

u
o



N ,





41I
i-H










4 r:9 0


0 0
m=o




















0lU) + U

cAuI II
0 ) .
C odP
Ud 4 .0

*M U C
t3k ft Q
o nt U
e i-L








and 50 mM Na+) in a 1 or 5 cm cell which was thermostated

at 210C.

The spectral data were analyzed by the McGhee-von Hippel

technique93 according to the following equation:


SK ( n) ( 1 nv
Sa (1 (n 1)v


where v is the number of moles of peptide bound per DNA base

pair, L is the concentration of free peptide, K is the
a
apparent binding constant, and n is the number of base pairs

per binding site. To calculate the above parameters, the

raw absorption data are fed to a Pet 2001 microcomputer

which then calculates v and v/L (see experimental chapter).

For each pair of v and v/L, a value of K is calculated as

a function of N. The programmer inputs various values of n

until the smallest deviation in the set of K's is obtained.

The average value of the K's is then taken as Ka (e.g., see

Table 3).

A plot of v/L versus v is shown in Figure 27. It can

be seen that only a small range of v's was obtained from

the spectral titrations. Furthermore, most of the points

are on the straight part of the graph, with only a few

points on the curved part. Therefore, the values of Ka and

n should be taken with caution. More accurate values of

Ka and n could only be obtained with a broader range of

v's. However, the values of K in Table 4 do reflect rela-
tive binding strengths.
tive binding strengths.









Table 3. Typical Output for McGhee-von Hippel Binding Iso-
therms for Various Values of na.



K


a

.405

.421

.445

.465

.488

.508

.524

.541

.561

.581

.601

.617

.631

.657

.681


V

.052

.051

.050

.049

.049

.048

.048

.047

.046

.046

.046

.045

.044

.043

.042


v/L

7379

7384

7620

7782

8077

8298

8435

8599

8908

9254

9643

9939

10126

10585

11001


n=3

9894

9814

10096

10261

10628

10875

10997

11157

11527

11948

12423

12760

12942

13420

13835


n= 4

11347

11201

11505

11662

12065

12321

12424

12573

12973

13430

13948

14302

14471

14944

15343


n=5

13203

12958

13283

13423

13868

14127

14198

14325

14757

15254

15823

16190

16337

16790

17155


performed by a Commodore Pet 2001 microcomputer for L-lys-
L-DNP-ornA (9) (10 mM Na+) where a = ratio of bound peptide,
v = moles of bound peptide/mole of DNA base pairs, L =
concentration of free peptide, n = number of base pairs per
binding site, and K = apparent binding constant for that set
of v and v/L.










11.0





10.0




7 .















4.1 4.3 4.5 4.7 4.9 5.1 5.3
vx 12

Figure 27. A Plot of v/L Versus v From Data in Table 3.


The values of Ka and n for the dipeptides 3, 4, 5, 6,
9 and 10 are found in Table 4. A number of observations

can be made.

1) In general, Ka and n decrease with increasing

ionic strength.
2) Peptides 5, 9 and 10 bind very strongly to DNA,

while compounds 4 and 6 exhibit fairly weak binding.

3) Peptide 3 binds stronger than its diastereomer,

peptide 4, and peptide 5 binds stronger than the diastereo-

meric mixture 6.











-l r-l Ll Ln U)
I *



*C *


Z '


r-

Q t-1 0 cn C) to i-





o a e a C)

+ 0



4-J 4J 0
Pl>-













.4-jf
C-1 C) ( 0





Z 0








o) +
C*











0I0
c, z0 Cu










0 o c \, O M H
a ZU

9 2 uo 0 E- L L ;R; Io I ~ 4 9

P-. X r C)- ( P -4 (N c!



0 0 0O
= > Q *H


*H tI a r-

H H*1 0



2 C3 t-) |r-I
S'-' 0
0 < < F 4-





I I k 2 Z U <
I II I I

U-) Q) ) U ) Q U 4


r-l r-4 r-l -i- r-4 r-4 0 r4
-i -i i -i i CU 2O u
^ ^i ^i c









4) All binding constants are higher than those ob-

tained from equilibrium dialysis, but these constants were

calculated at different peptide and DNA concentrations.


Circular Dichroism Studies

The effect of increasing concentration of salmon

sperm DNA on the circular dichroism spectra of compounds

3, 4, 5, 6, 9 and 10 in the range of 300 400 nm was

studied. The results, shown in Figures 28 and 29, show

that there is an induced CD spectra for all compounds in

the range of 300 400 nm. The extent of the induced CD

in the presence of excess DNA (i.e., at a DNA base pair to

peptide ratio of 12) follows the following order:

1) L-lys-L-DNP-ornA (9) = L-lys-D-DNP-ornA (10) = L-lys-L-

DNP-DABA (5) > L-lys-D,L-DNP-DABA (6) > L-lys-L-S-NT-cysA (3)

> L-lys-D-S-NT-cysA (4).


Equilibrium Dialysis

Equilibrium dialysis was utilized to perform direct

binding studies of the dipeptide amides with salmon sperm

DNA. The quantitative analysis of the dipeptide amide was

accomplished by UV absorption data at the Xmax of the

various dipeptides. These studies were carried out in

duplicate at a single peptide and DNA concentration (i.e.,
-4 -4
at 3.0 x 10 M DNA phosphate/liter and either 2.5 x 10 M
-4 +
or 5.0 x 10 M peptide) using a 10 mM MES buffer (5 mM Na ,

pH 6.2). The apparent binding constant, Ka, was evaluated

from












































* *C
Nr o cN a \o ( o
+ i i I I r-


SOT x (c)


0 o 00 00


+ I xI r

O-0 x (e)


+ 1 I I I


x (e)


So

N
+


o o o


OT x (e)


*o
.r-0


0





m Cd
P4











kcdO







41
Cr
















Ca

SN
**Hl















U r4
E- a4
Sr-
























o *


H U
4 0
C)


















co



10
*rto






0 *r-
< :r








^3 ?-
[_CL^

*1=i
00,







3r
b0
*i-t








1 L-lys-L-S-NT-cysA (3)
+2.0
+2.32-5 350 375 400 nm
S0.0



C -4.0 -

-6.01

L-lys-D-S-NT-cysA (4)

+2.0 325 3 0 375 4 0 nm
: 0.0

-2.0 c

S-4.0

-6.0


Figure 29. The CD Spectra of Dipeptides 3 and 4 in the
Absence and Presence of DNA. The DNA base
pair to dipeptide ratio is: a) 0; b) 1.5;
and c) 12.

Rb
K = R
a (Pt Rb)Rf


where Rf and Rb is the concentration of free and bound

dipeptide, respectively, and Pt is the total DNA phosphate

concentration. This equation can be used assuming that

each DNA phosphate binds independently to a peptide mole-

cule and that the maximum number of binding sites per DNA

phosphate is one. Even though this may not be true, the

binding data from the interactions of peptides to nucleic

acids are evaluated in this way and thus the K values in
a













Table 5. Apparent Binding Affinity (K ) of the Dipeptides
to Salmon Sperm DNA a


Peptide Systema

L-lys-L-pNO2-pheA (1)

L-lys-D-pNO2-pheA (2)

L-lys-L-S-NT-cysA (3)

L-lys-D-S-NT-cysA (4)

L-lys-L-DNP-DABA (5)

L-lys-D,L-DNP-DABA (6)

L-lys-L-S-pNO2bz-cysA (7)

L-lys-D-S-pNO2bz-cysA (8)

L-lys-L-DNP-ornA (9)

L-lys-D-DNP-ornA (10)


K1

5.3 x 102

6.8 x 102

3.6 x 103

2.7 x 10

3.0 x 103

1.3 x 103

4.8 x 102

2.2 x 102

2.7 x 103

3.1 x 103


K2

5.4 x 102

4.9 x 102

1.3 x 103

8.7 x 103

1.3 x 103

1.3 x 103

3.2 x 102

6.6 x 101

7.9 x 102

1.6 x 103


a -4
Equilibrium dialysis was carried out with 3.0 x 10-4 M
salmon sperm DNA in 10 mM MES buffer (5 mM Na+ pH 6.2)
with peptide concentration of either 2.5 x 10-4 M (K1)
or 5.0 x 10-4 M (K2).








Table 5 do reflect differences in binding strength.

Examination of Table 3 reveals that, in general, dipeptides

containing a nitrotolyl moiety (i.e., 3 and 4) or a DNP

moiety (i.e., 5, 6, 9 and 10) bind more strongly than

those compounds containing a p-nitrophenyl moiety (i.e., 1,

2, 7 and 8).


Melting Temperature Studies

The effect of increasing concentration of peptide on

the Tm (melting temperature) of the helix-coil transition

for three different nucleic acids was determined. Several

observations can be made from examination of the Tm data

in Table 4.

1) All peptides stabilize the helix relative to the

random coil and the transition exhibits a monophasic

melting behavior. Furthermore, the DNP derivatives 5, 6,

9 and 10 and the nitrotolyl containing peptide 3 stabilize

the helix to a greater extent than the other peptides.

2) All peptides, in general, stabilize poly d(A-T)

to a greater extent than poly I poly C, indicative of

preferential binding to A-T sites.

3) For a given set of diastereomeric dipeptides,

the compound having the L configuration for the C-terminal

amino acid a carbon stabilizes the helix to a greater ex-

tent than the peptide with the D configuration at the C-

terminal amino acid a carbon (e.g., 3 > 4, 9 > 10, etc.).


















0

OH
H E-

.,-
*H *H
vn Bu





r- <3 U

U
S E-,
,H a<
'-4 I


:It. 0 E-
0 <
U )-
E n




4-) 0



S-4 4'd
6HS I n3

<

Q 0 ()
9, 4-y j "H

0 W a

0 *- 4-
io -
O2

0*
















J r =) 0)
<1 -a U







L 0 -C )
010




<*H *



.4- 4 -1



4-1 r E
a O F



csl ) c
4-4 cx z=
0 0
04004
SC 9


H o Ea






DC a l



U 0 0

Sr- 4-
W ^ +->






^,
P

s 1I


AI O t ,-i Nr tN
CD







Lr)
1 0 H l- I Cl i


U


0


I




0
<





p.:
z





p.



0


;1
r- 00 t

O 0 O
Ln





00 t '.0










A K9 0 C
1-0









c 4 o i-q
Ln






*1 0
C r-4 C Ln






















S I?
CM4



























a a I
C?4
1 i
.A P4 ..
c^-) Q ^-


* 00- Vo ) Lin

4 r4 t_: t)
'-4





om Co t1 r-










M9 0 1

a) in 00 CM






* *
N b <0


ir-
M M






* -) *










4 m

S r-4
l-


CN t4 0 -H
CM \O i-f


* *
r-4 C
tn1
A




O O

S 0
A


t--, e.



-| '-' -) U) ,-> o|C
>, > -I 1
u U





< i i a
V i I I I
tn < < N N <
> p C.
U < I CN C 1-
I A 0 0 0 0
E- z 2 Z I
Z P.. Q P.. p. P. P.






-4 .r-q r-4 -4 v--i v-I r-4
E l-1 ^- >-1 ^) -] ^1 _


9L4




N




to






\o






te4
r1-
0


4)
'*-0





C1)
(M




v0






-
cl
va

o-H



4-o
Ln
















00
0)S












a) 'd
m o

S o











'-ia)


U) .3









The interpretation of these data is complicated by

the fact that the helix-coil transition involves the inter-

action of the dipeptide (P), not only with the helix (H),

but also with the coil (C) as shown in the following

equation:

KH KC
H + P H P C P C + P.


It can be assumed that KH is greater than KC since T is
SC m
increased in the presence of the dipeptide system. How-

ever, Gabbay and Kleinman have pointed out that this con-

clusion is valid only when KH and KC are determined

directly.4


Discussion

Much research has been devoted to the study of the

recognition specificity of proteins for nucleic acids.

Due to the several types of forces operating at several

sites along the nucleic acid and protein, this type of

study is of immense complexity. Furthermore, recent

studies on chromatin reveal that protein-DNA binding

specificity is not only a dynamic process, continuously

changing during the cell cycle, but may also involve spe-

cific protein aggregates-DNA recognition.95-

The problem can be simplified by studying the inter-

action specificities of small dipeptides with DNA as under-

taken in this work. As stated before, small molecules may

bind to DNA by electrostatic, hydrogen bonding, hydrophobic









forces, or any combination of these. Electrostatic

forces are not expected to lead to specific recognition

between a small molecule and DNA, since this type of

interaction may occur at each of the DNA phosphate groups.

Furthermore, hydrogen bonding may occur at any base pair;

so, it can be assumed that hydrogen bonding forces would

not lead to specific recognition.

It has therefore been suggested that hydrophobic-type

interaction (which has been noted in the binding of aro-

matic and hydrophobic amino acids to DNA preferentially at

A-T sites) is primarily responsible for the recognition
87
process. Gabbay has shown that the "partial" intercala-

tion of the aromatic ring of the side chain of phenylalanine

with native DNA (which has been shown to possess A-T

clusters 05- ) is consistent with this interpretation.

From the work with reporter molecule I and the dia-

stereomeric dipeptides L-lys-L-pheA and L-lys-D-pheA, two

criteria were established to dictate whether full, partial

or no intercalation will operate.78,82,84 From the re-

porter molecule work, it was shown that there is a depend-

ence of length between the aromatic ring and the quaternary

ammonium group, i.e., if the length was long enough, full

intercalation occurs; if the length was short enough, only

partial intercalation occurs. The dependence on the

chirality of the C-terminal amino acid a carbon for partial

intercalation was shown by the work with L-lys-L-pheA and

L-lys-D-pheA. The results suggest that the dipeptide is









stereospecifically anchored to the DNA by an electrostatic

interaction between the N-terminal lysyl a and E amino

groups and that as a result of this binding, the orienta-

tion of the aromatic ring of the C-terminal amino acid is

established. When the C-terminal amino acid is of the L

configuration, the aromatic ring is pointed toward the in-

terior of the helix; whereas, in the case of the D config-

uration, the aromatic ring is pointed out toward the sol-

vent (Figure 15).

In an attempt to provide additional experimental

evidence for (or against) these two criteria, the synthesis

and study of the interaction specificities of a series of

dipeptide amides were undertaken. The series consisted of

pairs of diastereomeric dipeptides (differing in chirality

of the C-terminal amino acid a carbon) with an aromatic

ring of various distances from the C-terminal amino acid

a carbon.

From the above considerations, it was assumed that

the aromatic ring of L-lys-L-pNO2-pheA (1) would partially

intercalate while that of L-lys-L-S-NT-cysA (3) may par-

tially or fully intercalate and that there would be full

intercalation for L-lys-L-DNP-DABA (5), L-lys-L-S-pNObz-

cysA (7) and L-lys-L-DNP-ornA (9). On the other hand, no

intercalation of the aromatic ring was expected for those

peptides whose C-terminal amino acid a carbon was of the

D configuration [i.e., L-lys-D-pNO2-pheA (2), L-lys-D-S-NT-

cysA (4), L-lys-D-S-pNO2bz-cysA (8) and L-lys-D-DNP-ornA (10)].








A "mixed" behavior was expected for the diastereomeric mix-

ture L-lys-D,L-DNP-DABA (6).

Several types of experimental data can be obtained to

determine the extent of intercalation. From 1H NMR studies,

the extent of line broadening and upfield shifting of the

aromatic proton signals is indicative of the extent of in-

tercalation. Full intercalation results in total line

broadening, while partial intercalation results in broad-

ening and upfield shifting of the aromatic peaks. No up-

field shifting is observed in the absence of intercalation.

Inspection of the NMR spectra of the peptides in the

presence of DNA (Figures 18 22) reveals some interesting

points. First, the NMR signals of aromatic protons of com-

pounds 3, 4, 5, 6, 9 and 10 are completely broadened,

indicative of full intercalation. The aromatic signals for

peptides 1, 7 and 8 are upfield shifted and broadened,

which would suggest partial intercalation. On the other

hand, the NMR signals for the aromatic protons on compound

2 are only slightly shifted and broadened. This can be in-

terpreted as the absence of intercalation. The greater

extent of line broadening for 1 as compared to 2 can be

explained by several mechanisms: 1) slower tumbling rates

of the aromatic ring of 1 in the DNA complex as compared

to 2; 2) slower exchange between the various DNA binding

sites for DNA-I as compared to the DNA-2 complex; 3) larger

differences in the ortho and meta protons of the aromatic

ring of 1 in the DNA complex as compared to 2; or 4) a








combination of all three mechanisms. Since the value of

the spin lattice relaxation time (Tl) is dependent (among

other things) on the correlation time (T ) and mean resi-

dence time (Tm), determination of T1 for 1 and 2 in the

presence and absence of DNA could help discriminate be-

tween the three mechanisms. Unfortunately, due to the

complexity of the spectra and inability to accurately meas-

ure peak intensities in the T1 spectra, T1 values could

not be obtained for the DNA-1 complex. However, previous

work by Gabbay87 on L-lys-L-pheA-DNA and L-lys-D-pheA-DNA

complex revealed that the T1 of the aromatic protons of

both peptides in the DNA complex were nearly identical

(T1 = 0.65 sec.), suggesting that the tumbling rate (l/Tc)

and the chemical exchange rate (I/T ) of the aromatic pro-

tons of the peptides in the DNA complex were very similar

in magnitude.47,109'110 It was concluded that the greater

signal line broadening observed for the aromatic protons

of the L-lys-L-pheA-DNA complex as compared to the L-lys-

D-pheA-DNA complex could only be due to large differences

in the chemical shifts experienced by the ortho, meta, and

para protons. It can therefore be assumed that the greater

line signal broadening of the aromatic protons of L-lys-L-

pNO2-pheA in the DNA complex as compared to that of the

L-lys-D-pNO2-pheA-DNA complex can only be due to differences

in the chemical shifts experienced by the ortho and meta

protons.








The magnitude and the sign of the change in the vis-

cosity of a DNA-peptide complex can also be informative as

to the extent of intercalation. An increase in the vis-

cosity of a DNA-small molecule solution is noted for full

intercalation of the aromatic ring of the small molecule

between the base pairs of DNA, while a decrease in viscos-

ity is observed for "partial" intercalation. It should be

noted that electrostatic binding (without intercalation)

of cationic molecules to DNA also decreases the viscosity

of a DNA-small molecule solution due to neutralization of

the anionic phosphates and subsequent shortening of the

DNA chain. However, the degree of decrease in the viscos-

ity of the DNA solution is not as great as when intercala-

tion also takes place.

Consistent with the NMR data, peptides 3, 4, 5, 6, 9

and 10 increase the viscosity of the DNA-peptide solution

(i.e., full intercalation), while peptides 1, 2, 7 and 8

decrease the viscosity of the DNA-peptide solution, the de-

crease greatest with L-lys-L-pNO2-pheA (1).

Further evidence for full intercalation of the aromatic

rings of compounds 3, 4, 5, 6, 9 and 10 comes from the UV

and CD studies. From the UV studies, a red shift in X
max
and hypochromism is observed for these peptides in the

presence of DNA. The red shift arises from an electronic

interaction between the aromatic ring and the purine and

pyrimidine base resulting in a splitting of the excited

state energy levels of the peptide aromatic ring producing









a lower excited state energy level and the hypochromism

arises from an intensity exchange between the coupled

transition moments of the aromatic ring and the purine

and pyrimidine bases. There is also an induced CD in the

region of ax of the peptides in the DNA complex. This

can be explained by intercalation of the aromatic ring

into the interior of the DNA double helix, which is

assymetric (thus, an induced CD).

When considering the binding of these peptides to DNA,

it must be kept in mind that two types of interaction are

operative, namely electrostatic binding of the N-terminal

amino acid lysyl residue via the a and E amino groups and

intercalation of the aromatic ring of the C-terminal amino

acid. Furthermore, one type of binding may predominate

depending on the relative strengths of the two possible

interactions. In general, electrostatic binding is favored

at low ionic strength, while intercalation is favored at

higher ionic strength. The apparent binding constant, Ka,

is reflective of this phenomenon.

Those compounds which have been shown to intercalate

from NMR and viscosity studies also have a higher K than
a
those compounds which only partially intercalate or do not

intercalate at all (Table 5). This is also borne out from

the Tm studies. The compounds that fully intercalate

stabilize the DNA double helix to a greater extent than

those that only partially intercalate. The Tm data also

show that poly d(A-T) is stabilized more than poly I -

poly C, indicating the peptides show preference for A-T sites.








The results from these studies were pretty much in

line with what was predicted. However, a few unexpected

results were obtained.

1) The aromatic ring of L-lys-D-S-NT-cysA (4) inter-

calates. It was predicted that this would not be the case,

since the aromatic ring of 4 should point out toward the

solvent. However, if a strong electronic interaction

exists between the aromatic ring of 4 and the base pairs

of DNA, intercalation would be favored even at low salt.

Hydrophobic-type interactions would also favor intercala-

tion. One or both of these factors is presumably opera-

tive for this molecule.

2) The aromatic ring of L-lys-D-DNP-ornA (10) also

fully intercalates. This result can be explained by two

mechanisms: 1) intercalation is favored for the same

reasons as noted above (i.e., electronic or hydrophobic);

or 2) electrostatic binding and intercalation as a result

of the flexibility of the "arm" between the a carbon and

aromatic moiety. In other words, the arm is long enough,

after electrostatic binding, to allow the aromatic ring,

which is seeking a nonpolar environment, to "swing" around

and intercalate. This is not unreasonable in light of the

results from L-lys-D,L-DNP-DABA (6). The extent of in-

duced CD and increase in viscosity for this peptide is not

as great as for L-lys-L-DNP-DABA (5). However, L-lys-D-

DNP-ornA (10) behaves exactly like L-lys-L-DNP-ornA (9).

With regard to mechanism 2, it would be tempting to assume,









in the case of the L-lys-D,L-DNP-DABA (6) peptide, that

electrostatic binding is taking place, and the aromatic

ring of the L-lys-L-DNP-DABA diastereomer is fully inter-

calating (as in peptide 5), while that of the L-lys-D-

DNP-DABA is only partially intercalating since the arm

is not long enough (as opposed to the case of L-lys-D-

DNP-ornA, 10) to allow full intercalation; thus, the

"mixed" results.

Model building studies can be used to support

mechanism 2. For example, for reporter molecule I with

n = 1, the distance between the quaternary ammonium

nitrogen and the center of the aromatic ring is on the
0 0
order of 4 A, while this distance is on the order of 5 A

for reporter molecule I with n = 2. Furthermore, the

distance between the center of the aromatic ring and the

assymetric carbon of the phenylalanyl residue of L-lys-L-
0
pheA is also approximately 4 A. Therefore, the critical

distance between the aromatic ring and the a carbon to
O
which it is attached must be between 4 and 5 A to allow

for full intercalation. For L-lys-L-DNP-DABA (5) and

L-lys-L-DNP-ornA (9), this distance is approximately 6
0
and 7.5 A, respectively, which is certainly long enough

to allow full intercalation of the aromatic ring for

each compound. For L-lys-D-DNP-ornA (10), the distance

between the aromatic ring and the a carbon to which it is
O
attached is about 4.5 A after swinging the aromatic ring

around and pointing it toward the interior of the helix.








This distance is apparently long enough to allow for full

intercalation. However, for L-lys-D-DNP-DABA, this dis-
O
tance is only on the order of 3 A, which is insufficient

for full intercalation. It should be noted that the models

were built assuming a certain rigidity in the peptides

after binding. More experimental evidence is needed to

support this mechanism, however.

3) Finally, the results from the studies with L-lys-

L-S-pNO bz-cysA (7) and L-lys-D-S-pNO bz-cysA (8) are

very difficult to interpret. It was hoped that studies

with these two compounds would resolve any anomolies

created by L-lys-D,L-DNP-DABA (6). Instead, the entire

situation becomes more complicated. It is evident from

these studies that full intercalation is not taking place

even for L-lys-L-S-pNO2bz-cysA (7). The NMR suggests that

there is at least partial intercalation for both 7 and 8;

yet neither of these peptides significantly decreases the

viscosity of a DNA-peptide solution [relative to L-lys-

L-pNO2-pheA (1)]. The fact that there is no full intercala-

tion for L-lys-L-S-pNO2bz-cysA (7) can be explained by two

possible effects: 1) the sulfur in the middle of the side

chain is creating a steric effect by positioning the aro-

matic ring away from the interior of the helix; or 2) there

is insufficient electronic interaction between the aro-

matic ring and the base pairs of DNA. This last explana-

tion is unlikely in light of the fact that the nitrophenyl

ring of reporter molecule I (where n = 2 4) fully









intercalates. In order to resolve this dilemma, a more

in-depth study of these compounds and derivatives is re-

quired. Furthermore, with regard to the first explana-

tion (i.e., steric effects), model building studies were

uninformative.

In conclusion, this work has shown the partial depend-

ence of chirality of the N-terminal amino acid a carbon and

the distance between the aromatic ring and this center of

chirality for intercalation. It has also been suggested,

though, that electronic and hydrophobic effects can also

predominate in terms of intercalation, and must be con-

sidered in any study of the interaction of small molecules

with DNA.














CHAPTER III
EXPERIMENTAL


All starting amino acids were purchased from either

Sigma Chemical Co. or Vega Biochemicals, and were used with-

out further purification. Salmon sperm DNA was purchased

from Worthington Biochemical Corp. Elemental analysis of

all products was performed by Atlantic Microlab, Inc.,

Atlanta, Georgia. TLC plates with fluorescent indicator

were purchased from either Kodak Chemical Co. or Scientific

Products, Inc. Melting points were taken on a Mel-Temp appa-

ratus and the measurements are uncorrected. All solutions

containing peptides or DNA were prepared in buffers made

with deionized water.

The 1H NMR spectra were recorded with either a Varian

A-60A or a Jeol-JNM-FX100 Fourier Transform NMR spectro-

meter. Chemical shifts were determined relative to the

internal standard trimethylsilane (TMS) for spectra taken

with organic solvents or 2,2,3,3-d4-3-trimethylsilylpro-

prionate (TSP) for those determined in deuterium oxide.

Ultraviolet and visible absorption spectra were recorded

with a Cary-17D UV/Vis spectrophotometer. A Jasco J-20

spectropolarimeter was used to record circular dichroism

measurements. Viscosity measurements were performed with

a low shear Zimm viscometer (Beckman Instrument Co.).









Synthesis

All of the prepared dipeptides, and the required inter-

mediates leading up to the final products, were analyzed for

purity by two or more of the following: TLC, NMR or ele-

mental analysis. All reactions are known to proceed without

racemization. The various abbreviations used in this sec-

tion are: 1) THF (tetrahydrofuran); 2) CBZ (carbobenzoxy);

3) t-Boc (t-butyloxycarbonyl); 4) NT (nitrotolyl); 5) DNP

(dinitrophenyl); and 6) TFA (triflouroacetic acid). In some

cases, a molecule or two of water of hydration has to be in-

cluded for the elemental analysis. This is reasonable since

,the final compounds, and some of the precursors, are quite

hygroscopic.

Preparation of L-lysyl-L-pNO -phenylalanine amide dihydro-
bromide, 1.

L-pNO2-phenylalanine ethyl ester hydrochloride was syn-

thesized by the Curtius and Goebel esterification proce-

dure:11 L-pNO2-phenylalanine (1.0 g, 4.7 mmol) was placed

in a 3 neck flask equipped with a magnetic stirrer, a drying

tube and a gas dispersing tube. Then, 1u00 mL of ethanol

was added and a stream of dry hydrogen chloride was passed

through the solution at room temperature until all the solid

dissolved. Upon dissolution, the flask was placed in an ice/

acetone bath and the bubbling continued to saturation. The

tube was removed and the stoppered flask was allowed to stand

at room temperature for 4 hours. The solution was then evap-

orated by a stream of dry nitrogen. The solid obtained was








recrystallized from ethanol/ethyl ether to yield 1.1 g (4.0

mmol, 85% yield) of L-pNO2-phenylalanine ethyl ester hydro-

chloride. The 1H NMR in D20 showed a 3H methyl triplet at

6 1.3 ppm, a 2H methylene doublet at 6 3.5 ppm, a 3H multi-

plet at 6 4.4 ppm and two 2H doublets at 6 7.6 ppm and 6.

8.3 ppm, respectively, for the aromatic protons.

Anal. calculated for C11H15N204C1; C, 48.00; H, 5.46.

Found: C, 48.06; H, 5.51.

The ester was then coupled to DiCBZ-L-lysine by the
112
general method of Anderson et al. DiCBZ-L-lysine

(0.60 g, 1.4 mmol) in 20 mL freshly distilled THF was placed

in a 3 neck flask equipped with a magnetic stirrer and a

calcium sulfate drying tube. A stream of dry nitrogen was

passed over the solution to keep the atmosphere free of

moisture. The flask was then placed in an ice/acetone bath

and 1.5 mmol of triethylamine was added to the solution,

followed by 1.5 mmol of isobutylchloroformate. The solution

was allowed to stir for 15 minutes, after which time a pre-

cooled solution of L-pNO2-phenylalanine ethyl ester hydro-

chloride (0.40 g, 1.4 mmol) in 10 mL of 1:1 THF/DMF was

added. An additional 1.5 mmol of triethylamine was added

and the reaction was allowed to stir at -100C for 2 hours.

Upon completion of the reaction, the ice bath was removed

and the solvent evaporated with a stream of dry nitrogen.

The resulting solid was dissolved in ethyl acetate/saturated

NaCI solution, and the organic phase was washed 3 times with

1 M HC1, 3 times with saturated sodium bicarbonate, and 2








times'with saturated sodium chloride. The organic phase

was dried over sodium sulfate, filtered and evaporated.

The resultant solid was recrystallized from ethyl acetate/

hexane to yield 0.74 g (1.2 mmol, 86% yield) of DiCBZ-L-

lysyl-L-pNO2-phenylalanine ethyl ester. The 1H NMR in

CDC13 showed a broad 9H multiple at 6 1.6 ppm, a broad 4H

multiple at 6 3.2 ppm, a 2H methylene quartet at 6 4.2 ppm,

a broad 6H multiple at 6 5.0 ppm, a broad IH (NH) doublet

at 6 5.5 ppm, a broad 1H (NH) doublet at 6 6.8 ppm, a 13H

multiple at 6 7.3 ppm, and a 2H doublet at 6 8.1 ppm.

Anal. calculated for C33H38N409.H20; C, 60.71; H, 6.18.

Found: C, 60.80; H, 6.19.

The diprotected dipeptide ester (0.69 g, 1.1 mmol) was

dissolved in 70 mL methanol and saturated with ammonia gas

at 0C in a high pressure bottle. The bottle was securely

corked and allowed to stand at room temperature for 48 hours.

The cork was removed and the solvent evaporated by a stream

of dry nitrogen. The solid obtained was recrystallized from

ethyl acetate/hexane to yield 0.53 g (0.89 mmol, 81% yield)

of DiCBZ-L-lysyl-L-pNO2-phenylalanine amide. The 1H NMR in

D6-DMSO showed a broad 6H multiple at 6 1.4 ppm, a broad 2H

multiple at 6 3.0 ppm, a very broad 2H multiple centered

at 6 4.4 ppm, a 4H singlet at 6 5.1 ppm, a broad 17H multi-

plet at 6 7.4 ppm, and a 2H doublet at 6 8.2 ppm.

Anal. calculated for C31N508H33 1.5H20; C, 59.97; H, 5.71.

Found: C, 59.97; H, 5.71.








To cleave the CBZ group, 0.60 g (1.02 mmol) of the di-

protected dipeptide amide was placed in a reaction reservoir

of the HF generator (Figure 30) with a small stir bar and

0.5 mL of anisole. Liquid HF was placed into the main reser-

voir and distilled, using a water jacket at 900C, into the

reaction reservoir which had been placed in a dewar of dry

ice/acetone. After 20 mL of HF had been distilled over, the

reaction vessel was allowed to warm to 0C and stirred in an

ice water bath for an additional 1 hour. The HF was evapo-

rated and the residue dissolved in 0.1 M HCl/ethyl ether.

The organic layer was discarded and the aqueous layer washed

2 more times with ethyl ether. The aqueous layer was

lypholyzed.

The residue was then dissolved in 2 mL of 0.1 M NH OAc

and placed on a CM-Sepharose (CL-6B) cation exchange column.

The column was eluted with 0.1 M NH4OAc and fractions col-

lected. Those fractions absorbing at 295 nm were pooled and

lypholyzed. The residue was placed in a drying oven (400C)

for 12 hours, lypholyzed, and dried in an oven for an addi-

tional 12 hours.

The dipeptide amide dihydroacetate was dissolved in

2 mL H20 and placed on anAmberliteCG-400 anion exchange

column (bromide form) and eluted with H20. Fractions

absorbing at 275 nm were pooled and lypholyzed.

The residue was dissolved in MeOH, transferred to a

tared, labeled vial, evaporated and placed in a P205 dessi-

cator for 48 hours. The solid was broken into a fine powder

































N --
2


Figure 30.


A








D__3_ D





S2 D2B
B
4







c




Liquid HF Apparatus. With valve D1 open and all
others closed, liquid HF is drained from the HF
tank (A) into the distillation reservoir (B).
Valve D1 is closed and valves D2 are opened,
and liquid HF is distilled over into reaction
reservoirs (C). The reaction is stirred by mag-
netic stirrers (E) until completion. Upon com-
pletion, valves D2, D3, and D4 are opened and
the HF is evaporated.








and the vial placed in a drying pistol for 12 hours to

yield 133 mg (0.27 mmol, 26% yield) of L-lysyl-L-pNO2-

phenylalanine amide dihydrobromide. The 1H NMR in D20

showed a 6H multiple at 6 1.6 ppm, a 2H methylene triplet

at 6 3.0 ppm, a 2H methylene doublet at 6 3.3 ppm, a 1H

methine triplet at 6 3.9 ppm, a 2H doublet at 6 7.5 ppm and

6 8.2 ppm, respectively, for the aromatic protons.

Anal. calculated for.Cs H25N5O4Br1..5H20; C, 34.22; H, 5.32.

Found: C, 34.26; H, 5.34. M.P. 189 1920C.

Preparation of L-lysyl-D-pNO2-phenylalanine amide dihydro-
bromide, 2.

The dipeptide amide was prepared in the same manner as

compound 1, but starting with D-pNO2-phenylalanine. D-pNO2-

phenylalanine (1.0 g, 4.7 mmol) was converted to 1.1 g of

D-pNO2-phenylalanine ethyl ester hydrochloride (4.0 mmol,

85% yield) by the action of absolute ethanol and dry hydrogen

chloride. The dipeptide ester, DiCBZ-L-lysyl-D-pNO2-phenyl-

alanine ethyl ester was prepared by reacting DiCBZ-L-lysine

(0.60 g, 1.4 mmol) with 1.5 mmol triethylamine and 1.5 mmol

of isobutylchloroformate followed by addition of D-pNO2-

phenylalanine ethyl ester hydrochloride (0.40 g, 1.4 mmol)

and 1.5 mmol of triethylamine. Workup and recrystallization

yielded 0.70 g (1.1 mmol, 79% yield) of DiCBZ-L-lysyl-D-pNO2-

phenylalanine ethyl ester. The 1H NMR in CDC13 showed a

broad 9H multiple at 6 1.6 ppm, a broad 4H multiple at 6

3.2 ppm, a 2H quartet at 6 4.2 ppm, a broad 6H multiple at

6 5.0 ppm, a broad 1H doublet at 6 5.5 ppm, a broad 1H doublet




/









at 6 6.8 ppm, a 13H multiple at 6 7.3 ppm, and a 2H doublet

at 6 8.1 ppm.

Anal. calculated for C33H 38N409 5H20; C, 61.58; H, 6.11.

Found: C, 61.56; H, 6.11.

The diprotected dipeptide ester (0.59 g, 0.93 mmol) was

converted to 0.40 g (.67 mmol, 72% yield) of DiCBZ-L-lysyl-

D-pNO2-phenylalanine amide by the action of MeOH saturated

with NH3 and subsequent workup and recrystallization. The

H NMR in D6-DMSO showed a broad 6H multiple at 6 1.4 ppm,

a broad 2H multiple at 6 3.0 ppm, a very broad 2H multiple

centered at 6 4.4 ppm, a 4H singlet at 6 5.1 ppm, a broad 17H

multiple at 6 7.4 ppm, and a 2H doublet at 6 8.2 ppm.

Anal. calculated for C31N50 H33; C, 60.88; H, 5.63.

Found: C, 60.67; H, 5.68.

The diprotected dipeptide amide (0.60 g, 1.02 mmol) was

cleaved with liquid HF (20 mL) as previously described. The

lypholyzed residue obtained was passed through a CM-Sepharose

(CL-6B) column, eluting with 0.1 M NH4OAc and collecting

fractions absorbing at 275 nm as before. After lypholyza-

tion of the appropriate fractions and removal of all residual

NH4OAc, the solid was dissolved in 2 mL H20 and passed through

an Amberlite CG-400 anion exchange column (bromide form).

The fraction absorbing at 275 nm was then lypholyzed. The

resultant residue was dissolved in MeOH, transferred to a

tared vial, evaporated and placed in a P205 dessicator and

drying pistol for 48 and 24 hours, respectively. The resul-

tant powder was 375 mg of L-lysyl-D-pNO2-phenylalanine amide








dihydrobromide (0.75 mmol, 73.5%). The 1H NMR in D20

showed a 6H multiple at 6 1.6 ppm, a 4H multiple at 6 3.2

ppm, a 1H methine triplet at 6 3.9 ppm, and a 2H doublet at

6 7.5 ppm and 6 8.2 ppm, respectively, for the aromatic pro-

tons.

Anal. calculated for C15H25N5O4Br2 4H20; C, 31.52; H, 5.78.

Found: C, 31.50; H, 5.83. M.P. 194 1970C.

Preparation of L-lysyl-L-S-nitrotolyl-cysteine amide dihydro-
bromide, 3.

L-S-nitrotolyl-cysteine hydrochloride was prepared by

dissolving 4.0 g (2.55 mmol) of L-cysteine hydrochloride

in 60 mL of 1 M NaOH. The solution was cooled to 0C and

a solution of 2.0 g (12.7 mmol) of 2-flouro-5-nitrotoluene

in 40 mL of dioxane was added over a period of 2 hours.

Stirring at 0C was continued for 2 more hours after which

time the reaction was stirred at room temperature for 3 4

hours. The reaction mixture was washed 3 times with ethyl

ether, acidified with concentrated HC1 and evaporated. The

resultant solid was recrystallized from water to yield 2.84

g (9.7 mmol, 76.4% yield) of L-S-nitrotolyl-cysteine hydro-

chloride. The 1H NMR in D6-DMSO showed a sharp 3H methyl

singlet at 6 2.4 ppm, a broad 3H multiple at 6 3.8 ppm, a

1H doublet at 6 7.6 ppm, and a 2H multiple at 6 8.1 ppm.

Anal. calculated for C10H13N204SC1*H20; C, 38.64; H, 4.83.

Found: C, 38.70; H, 4.89.

L-S-NT-cysteine hydrochloride (1.5 g, 5.1 mmol) was

esterified with ethanol saturated with dry HC1. Subsequent








workup and recrystallization from ethanol/ethyl ether yielded

1.3 g (4.1 mmol, 80.4% yield) of L-S-NT-cysteine ethyl ester

hydrochloride.

The 1H NMR in D20 showed a 3H methyl triplet at 6 1.3

ppm, a sharp 3H methyl singlet at 6 2.5 ppm, a 2H methylene

doublet at 6 3.9 ppm, a 3H multiple at 6 4.4 ppm, a 1H

doublet at 6 7.6 ppm, and a 2H multiple at 6 8.1 ppm.

Anal. calculated for C12H17N204SC1-H20; C, 42.60; H, 5.62.

Found: C, 42.71; H, 5.70.

Di-t-Boc-L-lysine (1.3 g, 3.7 mmol) was dissolved in

dry THF and treated with 4.0 mmol of Et3N and isobutyl-

chloroformate, respectively, followed by 1.2 g of L-S-NT-

cysteine ethyl ester (3.7 mmol) and an additional 4.0 mmol

of Et3N. Subsequent workup and recrystallization from THF/

hexane yielded 1.2 g (2.0 mmol, 54.1% yield) of Di-t-Boc-

L-lysyl-L-S-NT-cysteine ethyl ester. The IH NMR in CDC13

showed a broad 27H multiple at 6 1.4 ppm, a 3H methyl

singlet at 6 2.4 ppm, a broad 2H multiple at 6 3.1 ppm,

a broad 2H multiple at 6 3.5 ppm, a 3H multiple at 6 4.2

ppm, a broad 3H multiple at 6 5.0 ppm, a 1H (NH) doublet

at 6 7.1 ppm, a 1H doublet at 6 7.4 ppm, and a 2H multiple

at 6 8.0 ppm.

Anal. calculated for C28H44N409S; C, 54.90; H, 7.19.

Found: C, 54.79; H, 7.27.

The diprotected dipeptide ethyl ester was converted

to the amide by ammoniolysis of 1.0 g (1.6 mmol) of Di-t-

Boc-lysyl-L-S-NT-cysteine ethyl ester in methanol.








Subsequent workup and recrystallization from ethyl acetate/

hexane yielded 0.37 g (0.63 mmol, 39.4% yield) of Di-t-Boc-

L-lysyl-L-S-NT-cysteine amide. The 1H NMR in CDC13 showed

a 24H multiple at 6 1.4 ppm, a 3H methyl singlet at 6 2.4

ppm, a 2H multiple at 6 3.1 ppm, a 2H multiple at 6 3.5

ppm, a 1H methine multiple at 6 4.1 ppm, a 1H methine multi-

plet at 6 4.8 ppm, a 2H multiple at 6 5.8 ppm, a 4H multi-

plet at 6 7.3 ppm, and a 2H multiple at 6 8.0 ppm.

Anal. calculated for C26H41N508S; C, 53.52; H, 7.03.

Found: C, 52.19; H, 6.96.

In the deprotection reaction, 0.23 g (0.39 mmol) of

Di-t-Boc-L-lysyl-L-S-NT-cysteine amide was treated with 30

mL of a solution of 25% TFA in CH2C12. The reaction was

allowed to proceed for 45 minutes at room temperature, and

then evaporated. The resultant residue was dissolved in

0.1 M HC1 and washed 3 times with ethyl ether. The aqueous

layer was lypholyzed to yield a solid which was subsequently

dissolved in 2 mL H20 and passed through an Amberlite CG-400

anion exchange column (bromide form). The fraction absorbing

at 275 nm was lypholyzed to yield a solid that was pure by

TLC. The solid was dissolved in methanol, and the methanol

solution was transferred to a tared vial and evaporated.

The vial was placed in a P205 dessicator for 48 hours, and

the resultant solid broken into a fine powder. The solid

was identified by NMR and elemental analysis as 0.17 g (0.31

mmol, 79.5% yield) of L-lysyl-L-S-NT-cysteine amide dihydro-

bromide. The 1H NMR in D20 showed a 6H multiple at 6 1.5









ppm, a 3H methyl singlet at 6 2.4 ppm, a 2H multiple at 6

2.9 ppm, a 2H multiple at 6 3.6 ppm, a 1H methine triplet

at 6 4.1 ppm, a 1H doublet at 6 7.5 ppm, and a 2H multiple

at 6 8.1 ppm.

Anal. calculated for C16H27N504SBr2 1.5H2O; C, 32.43; H, 5.07.

Found: C, 32.63; H, 5.48. M.P. 209 2110C.

Preparation of L-lysyl-D-S-nitrotolyl-cysteine amide dihydro-
bromide, 4.

D-S-nitrotolyl-cysteine hydrochloride was prepared

from 4.0 g (25.5 mmol) of D-cysteine hydrochloride and 2.0

g (12.7 mmol) of 2-flouro-5-nitrotoluene as previously des-

cribed. The resultant solid was recrystallized from water

to yield 1.9 g (6.5 mmol, 51.2% yield) of D-S-NT-cysteine

hydrochloride. The 1H NMR in D6-DMSO showed a sharp 3H

methyl singlet at 6 2.4 ppm, a broad 3H multiple at 6 3.8

ppm, a 1H doublet at 6 7.6 ppm, and a 2H multiple at 6 8.1

ppm.

Anal. calculated for C10H 3N204SC1H20; C, 38.64; H, 4.83.

Found: C, 38.66; H, 4.89.

D-S-NT-cysteine hydrochloride (1.7 g, 5.8 mmol) was

esterified with ethanol/HCl to yield, upon recrystallization

from ethanol/ethyl ether, 1.4 g (4.4 mmol, 75.9% yield) of

D-S-NT-cysteine ethyl ester hydrochloride. The 1H NMR in

D20 showed a 3H methyl triplet at 6 1.3 ppm, a 3H methyl

singlet at 6 2.5 ppm, a 2H methylene doublet at 6 3.9 ppm,

a 3H multiple at 6 4.4 ppm, a 1H doublet at 6 7.6 ppm, and

a 2H multiple at 6 8.1 ppm.









Anal. calculated for C 12H17N204SC.H20; C, 42.60; H, 5.62.

Found: C, 42.53; H, 5.68.

D-S-NT-cysteine ethyl ester hydrochloride (1.2 g, 3.7

mmol) was coupled to Di-t-Boc-L-lysine (1.3 g, 3.8 mmol)

using 3.9 mmol of Et3N and isobutylchloroformate, respec-

tively, in the mixed anhydride reaction. Subsequent workup

and recrystallization from ethyl acetate/hexane yielded 1.5

g (2.5 mmol, 67.6% yield) of Di-t-Boc-L-lysyl-D-S-NT-

cysteine ethyl ester. The 1H NMR in CDC13 showed a broad

27H multiple at 6 1.4 ppm, a 3H methyl singlet at 6 2.4

ppm, a broad 2H multiple at 6 3.1 ppm, a broad 2H multiple

at 6 3.5 ppm, a 3H multiple at 6 4.2 ppm, a broad 3H

multiple at 6 5.0 ppm, a 1H (NH) doublet at 6 7.1 ppm, a

1H doublet at 6 7.6 ppm, and a 2H multiple at 6 8.0

ppm.

Anal. calculated for C28H44N409S; C, 54.90; H, 7.19.

Found: C, 54.65; H, 7.27.

Ammoniolysis of 1.3 g (2.1 mmol) of Di-t-Boc-L-lysyl-

D-S-NT-cysteine ethyl ester, and subsequent recrystalliza-

tion, yielded 300 mg (0.51 mmol, 24.3% yield) of Di-t-Boc-

L-lysyl-D-S-NT-cysteine amide. The 1H NMR in CDC13 showed

a 24H multiple at 6 1.4 ppm, a 3H methyl singlet at 6 2.4

ppm, a 2H multiple at 6 3.1 ppm, a 2H multiple at 6 3.5

ppm, a 1H methine multiple at 6 4.1 ppm, a 1H methine

multiple at 6 4.8 ppm, a 2H multiple at 6 5.8 ppm, a 4H

multiple at 6 7.3 ppm, and a 2H multiple at 6 8.0 ppm.

The TLC on this compound indicated that it was only on the








order of 95% pure; therefore, no elemental analysis was

obtained, and the compound was cleaved without further

purification.

For the cleavage reaction, 200 mg (0.34 mmol) of Di-

t-Boc-L-lysyl-D-S-NT-cysteine amide was treated with TFA/

CH2C12 as before. After the reaction was completed, the

solvent was evaporated and the resultant residue was dis-

solved in 0.1 M HC1, washed 3 times with ethyl ether, and

the aqueous layer lypholyzed. The solid obtained was dis-

solved in 1 mL of H20 and passed through an Amberlite

CG-400 anion exchange column (bromide form). The fraction

absorbing at 275 nm was lypholyzed to yield a solid that

was pure by TLC, and was identified by NMR as 75 mg (0.13

mmol, 38.2% yield) of L-lysyl-D-S-NT-cysteine amide di-

hydrobromide. The 1H NMR in D20 showed a 6H multiple at

6 1.6 ppm, a sharp 3H methyl singlet at 6 2.4 ppm, a 2H

multiple at 6 2.9 ppm, a 2H multiple at 6 3.6 ppm, a 1H

methine triplet at 6 4.1 ppm, a 1H doublet at 6 7.5 ppm,

and a 2H multiple at 6 8.1 ppm.

Anal. calculated for C6 H 27N 0SBr2-.5H20; C, 33.45; H, 4.88.

Found: C, 33.46; H, 5.36. M.P. 214 2170C.

Preparation of L-lysyl-L-DNP-diaminobutyric acid amide
dihydrobromide, 5.

L-N,y-DNP-diaminobutyric acid was synthesized according
113
to the general method of Sanger. L-diaminobutyric acid

dihydrochloride (2.0 g, 10.5 mmol) was dissolved in 75 mL

hot water and excess CuCO3 (5.0 g, 40.7 mmol) was added.








The solution was filtered and the volume reduced to 30 mL;

NaHCO3 (5.7 g, 68.7 mmol) was added followed by dinitro-

flourobenzene (6.0 g, 32.2 mmol) in 20 mL ethanol. The

solution was stirred for 2 hours at room temperature, during

which time a green precipitate formed, which was filtered

and dried. The solid was subsequently dissolved in 100 mL

of 1 M HC1, followed by addition of thioacetamide (7.8 g,

10.4 mmol) dissolved in water. The solution was heated

for 1/2 hour on a steam bath to precipitate CuS. Activated

charcoal was added and the solution filtered. The filtrate

was evaporated and the solid obtained recrystallized from

1 M HC1 to yield 1.4 g (4.4 mmol, 42% yield) of L-N,y-DNP-

diaminobutyric acid hydrochloride. The 1H NMR in D6-DMSO

showed a broad 2H multiple at 6 2.3 ppm, a broad 3H multi-

plet at 6 3.9 ppm, a 1H doublet at 6 7.5 ppm, and a broad 4H

multiple at 6 9.0 ppm.

Anal. calculated for C10H13N406C1; C, 37.44; H, 4.08;

N, 17.47. Found: C, 37.40; H, 4.12; N, 17.45.

L-N,y-DNP-diaminobutyric acid (1.0 g, 3.1 mmol) was

converted to the ester via the normal procedure utilizing

absolute ethanol saturated with dry hydrogen chloride.

After evaporating the solvent, the resultant solid was re-

crystallized from ethanol/ethyl ether to yield 0.82 g

(2.4 mmol, 77.4% yield) of L-N,y-DNP-diaminobutyric acid

ethyl ester hydrochloride. The 1H NMR in D20 showed a

3H methyl triplet at 6 1.3 ppm, a 2H methylene quartet at

6 2.4 ppm, a 2H methylene triplet at 6 3.8 ppm, a 3H








multiple at 6 4.4 ppm, a 1H doublet at 6 7.3 ppm, a 1H

doublet of doublets at 6 8.4 ppm, and a broad 1H doublet

at 6 9.0 ppm.

Since the addition of the DNP group could go to either

the a amine nitrogen or the y amine nitrogen, it was neces-

sary to prove that only the y amine group reacted. The

ester hydrochloride was treated with aqueous NaHCO3 and

extracted into CDC13. The 1H NMR revealed a 3H methyl

triplet at 6 1.3 ppm, a 4H multiple at 6 2.1 ppm, a 3H

multiple at 6 3.7 ppm, a 2H methylene quartet at 6 4.2

ppm, a 1H doublet at 6 7.0 ppm, a 1H doublet of doublets

at 6 8.3 ppm, and a broad 2H doublet at 6 9.0 ppm.

In the protonated ester, the a proton was buried under

the 3H multiple at 6 4.4 ppm, with the other two protons

coming from the ester ethyl methylene. The y methylene was

a 2H triplet at 6 3.8 ppm in the protonated ester. Upon

neutralization of the positive charge, only one proton,

namely the a proton, shifted (from 4.4 ppm to 3.7 ppm),

thus indicating that the addition of the DNP group did not

occur at the a amine group. Furthermore, the y methylene

did not shift significantly, indicating that addition did

occur on the y amine group.

Anal. calculated for C12H17N406C1; C, 41.32; H, 4.88.

Found: C, 39.46; H, 4.83.

The TLC and elemental analysis indicated that this

compound was only on the order of 95% pure; however, it was

used without further purification. L-N,y-DNP-diaminobutyric








acid ethyl ester (0.82 g, 2.4 mmol) was converted to the

amide through the use of methanol saturated with ammonia.

Workup and recrystallization gave 0.52 g (1.6 mmol, 67%

yield) of L-N,y-DNP-diaminobutyric acid amide hydrochloride.

The 1H NMR in D6-DMSO/D20 showed a broad 2H multiple at 6

2.1 ppm, a broad 3H multiple at 6 3.8 ppm, a broad 1H

doublet at 6 7.3 ppm, a 1H doublet of doublets at 6 8.3

ppm, and a 1H doublet at 6 9.1 ppm.

Anal. calculated for C10H14N505C1.H20; C, 35.55; H, 4.74.

Found: C, 35.47; H, 4.63.

The L-N,y-DNP-diaminobutyric acid amide hydrochloride

was then coupled to DiCBZ-L-lysine by the normal procedure.

DiCBZ-L-lysine (0.32 g, 0.77 mmol) was dissolved in THF

and 0.78 mmol of Et3N and 0.78 mmol of isobutylchloroformate

added. A precooled solution of L-N,y-DNP-diaminobutyric

acid amide hydrochloride (0.25 g, 0.77 mmol) in DMF/THF was

added followed by an additional 0.78 mmol of Et3N. Subse-

quent workup and recrystallization afforded 0.25 g (0.37

mmol, 48% yield) of DiCBZ-L-lysyl-L-DNP-diaminobutyric acid

amide. The 1H NMR in D6-DMSO showed a broad 8H multiple

at 6 1.4 ppm, a broad 4H multiple at 6 3.0 ppm, a very

broad 2H multiple at 6 4.1 ppm, a 4H singlet at 6 5.1 ppm,

a broad 15H multiple at 6 7.4 ppm, a very broad 2H multi-

plet at 6 8.2 ppm, and a broad 2H doublet at 6 9.1 ppm.

Anal. calculated for C32H37N7010; C, 56.55; H, 5.49.

Found: C, 56.43; H, 5.53.








The diprotected dipeptide amide (200 mg, 29 mmol) was

stirred in 50 mL acetic acid that had been saturated with

dry hydrogen bromide (35% HBr by weight) for 2 hours. The

solution was poured into 150 mL of dry ethyl ether and

placed in the refrigerator overnight. The precipitate was

filtered and recrystallized from MeOH/ethyl ether to yield

100 mg (0.17 mmol, 58.9% yield) of L-lysyl-L-DNP-diamino-

butyric acid amide dihydrobromide. The 1H NMR in D20

showed a broad 8H multiple at 6 1.6 ppm, a 2H methylene

triplet at 6 3.0 ppm, a 2H methylene triplet at 6 3.6 ppm,

a 1H methine triplet at 6 4.0 ppm, a 1H doublet at 6 7.1

ppm, a 1H doublet of doublets at 6 8.3 ppm, and a 1H doublet

at 6 9.1 ppm.

Anal. calculated for C H 27N 70Br 23H20; C, 30.62; H, 5.31.

Found: C, 30.62; H, 5.33. M.P. 170 1730C.

Preparation of L-lysyl-D,L-DNP-diaminobutyric acid amide
dihydroacetate, 6.

The DNP derivative of D,L-diaminobutyric acid was pre-

pared as described previously. D,L-diaminobutyric acid di-

hydrochloride (5.0 g, 26.2 mmol) was dissolved in hot water

and treated with excess CuCO3 and NaHCO3. Then, 2,4-dinitro-

flourobenzene (15.0 g, 80.6 mmol) in ethanol was added.

The greenish solid obtained was dissolved in HC1 and thio-

acetamide (2.0 g, 26.6 mmol) was added. After filtering

and evaporation, the yellow solid obtained was recrystal-

lized from 1 M HC1 to yield 5.32 g (16.6 mmol, 63.4% yield)

of D,L-N,y-DNP-diaminobutyric acid. The IH NMR in D6-DMSO








showed a broad 2H multiple at 6 2.3 ppm, a broad 3H multi-

plet at 6 3.9 ppm, a 1H doublet at 6 7.5 ppm, and a broad

4H multiple at 6 9.0 ppm.

Anal. calculated for C10H13N406C1; C, 37.44; H, 4.08.

Found: C, 37.40; H, 4.12.

The D,L-N,y-DNP-diaminobutyric acid hydrochloride

(4.5 g, 14.0 mmol) was treated with absolute ethanol satu-

rated with dry hydrogen chloride. Workup and recrystalliza-

tion from ethanol/ethyl ether yielded 4.8 g (13.2 mmol,

94% yield) of D,L-N,y-DNP-diaminobutyric acid ethyl ester

hydrochloride. The 1H NMR in D20 showed a 3H triplet at

6 1.3 ppm, a 2H quartet at 6 2.4 ppm, a 2H triplet at 6

3.8 ppm, a 3H multiple at 6 4.4 ppm, a 1H doublet at 6

7.3 ppm, a 1H doublet of doublets at 6 8.4 ppm, and a broad

1H doublet at 6 9.0 ppm.

Anal. calculated for C12H17N406Cl; C, 41.38; H, 4.92.

Found: C, 41.39; H, 4.94.

The D,L-N,y-DNP-diaminobutyric acid ethyl ester hydro-

chloride (0.30 g, 0.86 mmol) was converted to 0.21 g (0.66

mmol, 77% yield) of D,L-N,y-DNP-diaminobutyric acid amide

hydrochloride by the action of methanol saturated with

ammonia and subsequent workup and recrystallization from

methanol/ethyl ether. The 1H NMR in D6-DMSO/D20 showed a

broad 2H multiple at 6 2.1 ppm, a broad 3H multiple at 6

3.8 ppm, a broad 1H doublet at 6 7.3 ppm, a 1H doublet of

doublets at 6 8.3 ppm, and a 1H doublet at 6 9.1 ppm.









Anal. calculated for C10H14N505C-12H20; C, 33.73; H, 5.10.

Found: C, 33.55; H, 4.70.

For the coupling reaction, DiCBZ-L-lysine (0.39 g,

0.94 mmol) was dissolved in THF and 1.0 mmol of Et3N and

1.0 mmol of isobutylchloroformate added. D,L-N,y-DNP-

diaminobutyric acid amide hydrochloride (0.30 g, 0.94

mmol) in DMF/THF was added followed by an additional 1.0

mmol of Et3N. Subsequent workup and recrystallization

yielded 0.36 g (0.53 mmol, 56% yield) of DiCBZ-D,L-DNP-

diaminobutyric acid amide. The 1H NMR in D6-DMSO showed

a broad 8H multiple at 6 1.4 ppm, a broad 4H multiple at

6 3.0 ppm, a very broad 2H multiple at 6 4.1 ppm, a 4H

singlet at 6 5.1 ppm, a broad 15H multiple at 6 7.4 ppm,

a very broad 2H multiple at 6 8.2 ppm, and a broad 2H

doublet at 6 9.1 ppm.

Anal. calculated for C32H37N7010; C, 56.55; H, 5.45.

Found: C, 57.29; H, 5.74.

The diprotected dipeptide amide was deprotected by

the action of 20 mL of liquid HF on DiCBZ-L-lysyl-D,L-DNP-

diaminobutyric acid amide (0.90 g, 1.3 mmol). After

lypholyzation, the residue was dissolved in 2 mL of 0.2 M

NH4OAc and placed on a CM-Sepharose (CL-6B) cation exchange

column and eluted with 0.2 M NH4OAc. The fractions ab-

sorbing at 480 nm were pooled and lypholyzed. After re-

peated drying and lypholyzation, 150 mg of L-lysyl-D,L-DNP-

diaminobutyric acid amide dihydroacetate (0.28 mmol, 21.5%

yield) was obtained. The 1H NMR in D20 showed a broad 8H








multiple at 6 1.6 ppm, a 2H methylene triplet at 6 3.0 ppm,

a 2H methylene triplet at 6 3.6 ppm, a 1H methine triplet at

6 4.0 ppm, a 1H doublet at 6 7.1 ppm, a 1H doublet of doub-

lets at 6 8.3 ppm, and a 1H doublet at 6 9.1 ppm.

Anal. calculated for C20H33N 010 .5H20; C, 44.44; H, 6.30.

Found: C, 44.30; H, 6.38.

Preparation of L-lysyl-L-S-p-nitrobenzylcysteine amide di-
hydroacetate, 7.
114
The method developed by Berse et al. was used to

prepare L-S-p-nitrobenzylcysteine. L-cysteine hydrochloride

(3.14 g, 20 mmol) was dissolved in 60 mL of 1 M NaOH at

0C. Over a period of 30 minutes, a-chloro-p-nitrotoluene

(1.71 g, 10 mmol) in 30 mL dioxane was slowly added. The

reaction was then stirred for an additional 30 minutes at

room temperature. The reaction mixture was then washed 3

times with ethyl ether, acidified with concentrated HC1

and then the solvent evaporated. The resultant solid was

recrystallized from water to yield 2.0 g (8.0 mmol, 80%

yield) of L-S-p-nitrobenzylcysteine. The 1H NMR in D6-DMSO

showed a 2H multiple at 6 3.0 ppm, a 3H multiple at 6 4.0

ppm, and a 2H doublet at 6 7.7 ppm and 6 8.2 ppm, respec-

tively, for the aromatic protons.

Anal. calculated for C10H12N204S.I.5H20; C, 42.40; H, 5.30.

Found: C, 42.40; H, 5.35.

L-S-p-nitrobenzylcysteine (2.3 g, 9 mmol) was dissolved

in 10 mL H20 and 19 mmol of Et3N was added. To this solution,

Boc-On115 (2.1 g, 9 mmol) in 15 mL dioxane was added and the