Chapter 5 - _ ___ ___ - _ _ - - - --

Rate constants and product studies of the reactions did of ozone with nucleobases, nucleotides, an re ate

d d- compoun s in aqueous me ium - --

with nitrite andlor buten-3-01 and, the indigo method (when the rate constant was < IU3 dm3 mol-' s-I). Depending on the degree of protonation, the rate constant (in units of' dm3 mol-' s") varies substantially, e.g. in the case of

k = 18 (protonated), k ;- 1.4 x 10' (neutral) and k = 1.5 x lo6 (deprotonated). A similar variation has been found with the other nucleobases. Upon deprotonation, the mechanism of the ozone reaction may also change; r.g. no singlet dioxygen {O,'%) is formed in its reaction with 5-chlorouracil, but when the 5-chloro-uracilate ion predornitlates, it becomes a major product (- 42%). In the case of adenine and its derivatives, and thus also in the case of DNA, .OH is produced (via 02*" an intermediate). For the determination of their intrinsic ozone rate constants, tert-butylalcohol was hence added as .OH scavenger. The intermediates and end products studies uf the reactions of ozone with thymine and thymidim are carried out. The instant total peroxide yield is 100% in both systems. En thymine, 25% H202 is observed h m catalave and R2S a w y and it is increassd upto 4Vh with half life -12 min. Two organic peraxides w m ident5ed as I-hydropffoxymethylen-3-~2-oxo-prg,anoyl)-u and 5-hydroperoxy-5- methyhydmtoin. At high pH, thymine glycol and singkt oxygen are formed. In the

Publications from this chapter

1 . Jacob A. Theruvathu, R. E'lyunl, C. 1'. Aravindakumar, C. von Sonntag, Rate constants of ozone reactions with DNA, i t s constituents and related compounds, J. Chem. Soc. Perkin Trans.2,200 1,269-274.

2. R. Flyunl, Jacob A. Theruvathu, A. Ixitzke, C:. von Sonntag, Ozonation of thyrninc and thymidine, (manuscript under preparation).

Rule constants undproduct studies ... . . .. --- . ,

118

5.1. General

Ozone is gaining importance in drinking-water processing, because it is not

only a good oxidant but also a powerful disinfectant.'"'~''5~''7~'75~186~257~294-296

Mechanistic details of inactivation of bacteria and viruses are as yet not fully

understood. About 10' ozone molecules are required for the inactivation of a

ba~ te r iu rn . ' ~~ . '~~ It has been suggested that this is required for the destruction of the

bacterial cell wall and subsequent leakage of cellular contents.lr2 However, ozone

also causes mutations. 1u8.297~2v8 This may he taken as evidence for damage of its

DNA (with the cell remaining adequately intact), but it can not be excluded that

ozone by-products (e.g. formed in the reaction with thc cell wall) have caused the

mutagenic effect. For example, hydroperoxidm and H,Q are typical ozonation by-

prducts, and the latter is known to be weakly m u t a g r n i ~ . ~ ' ~ ' ~ In contrast, the

inactivation of viruses is considered to be mainly caused by ozone-induced damage

of their nucleic acids, although damage of the proteins that make up the capsid also

occurs. IF' The considerable spread of inactivation half-lives among various

enterovir~ses '~~ may indicate that, depending on the structure of the virus, there is a

varying competition between ozone damage of the nucleic acid (lethal) and the

proteins (less lethal).

There is already a lot of information available on thc reaction of ozone with

nucleic acids and their constjtuents.~~'Y~2".'87.'87-19F~2X[' Ir)l However, in order to reach a

better understanding o f the inactivation of viruses and bacteria by ozone, a detailed

Rule cunstants and avoduct studies.. . . I19

study covering the rate constants and mechanism of the reactions of ozone is

required. The determination of the rate of reaction of ozone with the nucleic acids

and their constituents by e.g. the stopped-flaw technique, is difficult because of the

strong overlap of the absorption spectra of the nucleobases with that of ozone (for the

data presently available see Table 5.1 ). This has led us to use another approach, the

application of competition kinetics.*'' Deprotonation or protonation of substrates has

a dramatic effect on the rate of ozone reactions. The most striking example is

phenol, whose ozone rate constant increases six orders of magnitude upon

deprotonation.302 As a consequence of this, the reaction of ozone with phenol i s

largely a reaction of ozone with the phenolate in equilibrium [pK,(phenol) = 101,

even in slightly acidic solutions. Similarly, arnines no longer react with ozone when

protonated. 2077303 Also in the reaction of ozone with the nucleobases, one has to take

their protonation/deprotonation equilibria into account. Thus, a detailed study

covering a large pH range is required for a better understanding of the kinetics of

ozone with the nucleobases and with DNA. Such data, together with some

additional information, will be presented in the present study.

~ l t h ~ ~ ~ h ,+here is a handfUl of rewrtsl 9.20.22.36.1 87.195-197.3M on the reactions of ozone

with nucleobase derivatives, some fine mechanistic aspects are still unresolved. Some

stable end products of the ozonation of uracil, thymidine, cytosine and

2'-deoxycytidine systems were rep~rted.'"~"~'"~ In order to study the fine

mechanistic details, we have selected thyminc and thymidine systems, which could

be applied for whole pyrimidine family.

Rate constunis und product s~ucldrs ... . -. 120

5.2. ~etermination of rate constants

In the present study, many rate constants have been determined by

competition kinetics. The theoretical background of the competition kinetics is

explained in section 2.4.2. I . Consider a competitor (C) and a nucleobase derivative

(N) react with ozone [reactions (5.1) and (5.211 and, products PI and P,. The product

of the reaction o f ozone with the competitor (P,) can be monitored while the

products of the reaction of ozone with the nucleobase derivative (P,) remain

unaffected.

C + 0, +PI (5.1)

N + 0, -+ P2 (5.2)

At a given initial ozone concentration ([OJ << [C] and [N]), the final

competition equation can be written as

IP, 1, - k, EC3 + ks.1 [Nl -- k5.2 LN1 = 1 +--- [pi I ks 1 LC1 k5.1 ECI

where k, , and k,, are the rate constants of the reaction (5.1) and (5.2). [P,], is the

concentration of the measured product in the absence of N and [P,] in the p m c e of N.

Plotting ([P,]J[P,] j 1 versus [Nj/[C] yields a straight line with the slope of

k, 2/k5,1. Since the rate constant of the competitor with ozone, k, ,, is known, the rate

constant of ozone with N, k,, can be calculated.

Buten-3-01 and nitrite were used as competitors, which yield one mol of

f~rrhaldehyde'~ and nitrate2"' per mol of ozone respectively. Solutions containing

Rate conslunfs and ~roduct studies ... . 121

substrate and competitor at varying ratios were mixed with an ozone solution, and

the competitor-derived product was quantified.

Around pH 7, all of the nucleobases are present largely in their uncharged

forms (cf; Table 5.1). In the case of uracil and guanine derivatives, protonation

occurs only at very low pH. On the other hand, cytosine, adenine and their

derivatives become protonated already at around pH 4. At high pH, the uracil

derivatives get deprotonated at pH -9.8, as long as one of the nitrogens remains

unsubstituted. In thymine, for example, the second nitrogen may also deprotanate at

very high pH. Most of the reported pK, values of 6-methyluracil were around 9.8.Tns

Because of the importance of pK, value for our study, we have remeasured it by

following the change of the ratio of the absorptions at 288 nm and 270 nm as a

function of pH (c:f.' inset in Figure 5.2), and our value of 9.8 is in agreement with

the rnajori ty of the literature data.

As can be seen from Table 5.1, the rate constants span more than five orders

of maptude. When the rate constant is small (c I d d d mol-' s-I), the indigo method is

most suited. For higher rate constants, usually buten-3-01 (k = 7.9 x 1 o4 d m h o l ' ' s-' ) Iu

has been wed as competitor. At pH > 10, the competition with butm-3-01 no longer yields

fully reliable values. For unknown reasons, the formaldehyde yiclds tend to be less

reproducible; they are typically lower than expected. In addition, when

tcrt-butylalcohol is added as 'OH scavenger, buten-3-01 can no longer be used as

competitor, since tert-butylalcohol also yields formaldehyde in 'OH-induced

Rate uonstunt.~ and aroduct studies ... . 122

reactions (-30%).2R9 Thus, under these conditions we used nitrite as competitor for

which rate constants of 3.7 x lo5 dm3 mol-' s-I, 3.3 x 1O" dm' mol-' s-' (at room

tempadture) and 1.6 x 1 05 dm3 mol-' s-' (at 1 1 O C ) are reported.?" The experiments were

done at 2 1 (f 1 ) "C, and at this temperature the rate constants of ozone with butene-3-01

and nitrite were redetermined using the stopped-flow technique under the conditions of

first-order kinetics (ten-fold excess of substrate) as 9.1 x 104 dm" mol-' s"' (c$ ref. 306) and

(3.5 - 4.0) x 10' dm3 mol-I s-', respectively, i.e, in agreement with the literature values.

To assure internal consistency, the competition of butene-3-01 vs NO, has been

studied. This can be carried out by measuring either fbddehyde or nitrate yields [cj.'

reactions (2.1 2) and (2.1 3)], e.g. as a a c t i o n of the [buten-3-ol]l[NO, 7 ratio.

The formaldehyde a s s a y is sensitive to elevated NO,- concentrations which

suppress the colow formation, and therefore the competition had to be done at low

concentrations of the competing substrates. Nitrate determination by HPIC is not affected

by the presence of the reactants. These two data sets yield k(N0,-jlk(buten-3-01) ratios of

5.4 and 5.7. Taking the literature data, k(N 0, -) = 3.7 x 1 O5 dm' mol-' s-' and ybuten-3 -01)

= 7.9 x lo4 dm mol-' s-' mentioned above, one anives at a ratio of 4.7. Considering the

inherent uncertainties of all these data, there is an acceptable agreement.

In their reactions with ozone, nitrogen-contain ing compounds, e.g, arnines,3"

can give rise to the formation of 'OH, and the ensuing ' 0 ~ - i n d u d reactions can distort

Rate constants und produd studies.. . . .,--- 123 -----

the kinetics considerably. For this reason, the b~ scavenger tert-butylalmhol has been

added to test, whether its presence has an influence on the rate of reaction.264 As will

be shown below, this precaution i s only necessary in the case of adenine and its

derivatives, i.e. also in DNA. However, the addition of tert-butylalcohol is essential

here. The rate constants determined in the present study are compiled in Table 5.1.

Table 5,l. Compilation of rate constants of ozone with nucleobases and related compounds (in units of dm3 mol-' s") at different stages of protonatiotl as determined in the present study. Literature values'" in brackets

(a) pK, values taken from ref.307, 5-~hlorouraci~'~%rotic acid and iso-orotic acid:" (b) rate constant determined at pH 7, where the carboxylate group is deprolvnated, but not yct une of the nitrogens; (c) n. d. - not determined; (d) in the presence o f 0.2 mol dmd3 ~vrt-butylalcohol; (e) see text; (f) the average rnr~lecular weight of the nucfeotides in DNA is taken as 350 Da.

Deprotonated

-3 x 10"

1.2 x 10"

-

9.2 x lo5 6.0 x lo5 1.3 x 10"

11. d. (Lj

n. d. 1.5 x 10'

-

1.3 x 10' (e) ( e )

4.0 x 10' n. d.

-

Substrate

Thymine

Thymidine 5'-dTMP 1,3-Dimethyluracil Uracil 6-Methyluracil 5-Chlorouracil Orotic acid(b' Iso-nrotic acid'b' Cytosine

2'-Deoxycytidine Cytidine 5'-dCMP ~ d e n i n e ' ~ ' 2'-~eoxyadenosine'~' ~d en~sine(~' 5'-dAMP Guanosine 2'-Deoxyguanosine 5'-dGM P DNA

Protonated

-

-

-

-

-

- - 18

44

40

5 5 5

< 300 n. d.

-

pK, values'"

9.9, >12

9.8 10.0 -

9.5, >I3 9.8 8.0

2.1,9.45 4.2, 8.9

4.6, 12.2

4.3 4.15 4.6

4.15, 9.8 3.8 3.5 4.4

2.5,9.2 2.5, 9.2 2.9,9.7

Neutral

4.2 x lo4 (2 .3 x 104) 3.0 x lo4

(1.6 x lo4) 2.8 lo3

650 140

4.3 x 10' 5.9 x 1 o3 3.7 x 10' 1.4 x 10'

(930) 3.5 x lo1 3.5 x lo7

( 1 . 4 ~ 10') 12 14 16

(200) 1.6 x 1 o4 1.9 x lo4 (5 x lo4) 4 1 Otd. CI

Rule constanis and product studies ... . 124 , . . - -. - - - -. - -

5.2.1. Uracil, thymine and their derivatives

The pH dependencies of the observed bimolecular rate constants of ozone

with thymine and 6-methyluracil are shown in Figures 5.1 and 5.2. Under these

conditions, where their neutral forms predominate, the reaction is relatively sIow,

but with increasing pH the rate of reaction increases. The position of the methyl

group at the C(5)-C(6) double bond has a noticeable effect on the rate of reaction.

Comparing the values at low pH, thymine reacts more than two orders of

magnitude faster than 6-rnethyluracil. The other uracil derivatives fa1 1 in between.

In contrast, the rate constants of the corresponding deprotonated forms are very

similar; they all center at around 1 Oh dm3 mol-' s-' (c,L Table 5.1).

Figure 5.1, Observed rate constant of the reaction of thymine with ozonc as a function of pH. The solid line is calculated on the basis of the rate constants measured at pH 2 , 1 1 and 13 and the established pK, values of 9.8 and 12.5. Inset: analogous data for thymidine. Competition with buten-3-01 (I) and nitrite in the absence (0) and presence (A) of tert- butylalcohol

Rate constants and product studies.. . . . -. -- 125

Upon raising the pH, the rate of reaction increases (Figures 5.1 and 5-21, and

in the case of thymine, experiments were extended into the pH range, where the second

nitrogen deprotonates, and a further increase in rate is observed (Figure 5.1 ) . At pH 1 3,

a value of 6.5 x 10' dm3 molL' s-' is found and this is attributed to the rate constant for

doubly deprotonated thymine. The determination of rate constant for the mono-

deprotonated thymine is associated with some ermr and can only be estirnatd h r n the

inflection point near pH 1 0, h m where we obtained a value of -3 x 1 0' dm3 moP1 s- ' .

Figure 5.2. Observed rate constant o f the reaction of 6-methyluracil with ozone as a function of pH. The solid line is calculated on the basis of the rate constants measured at pH 2 and 1 1 and the pK, value of 9.8 based on the data in the upper inset. Lower inset: a typical competition plot using buten-3-01 as competitor

As can be seen from Figurc 5.1 and 5.2, there is a considerable disagreement

between the measured pH dependencies (dotted lines) and the ones calculated on

the basis the pK, values and rate constants given in Table 5.1 (solid lines). 'The

Rate conslants and product studies .... 126

experimental curves suggwt that the reactions of compounds whch are noticeably more

acidic than represented by their pK, value would indicate. In the reaction of ozone with

olefins, there is evidence at low temperatures for the formation of a charge transfer

complex, which subsequently decays into products."ofi' If in the present case such a

charge bansfer complex is sufficiently long-lived it could turn more acidic than its

parent. As a consequence of this, the rate of reaction could be higher than the one

calculated on the basis of the pK, value of the parent. We are aware, that

spontaneous deprotonation of these compounds is slow due to their high pK, values

(-0.1 - 1 dl), but the presence of buffer that had to be added to keep the pH speeds

up the rate of deprotonation. Nevertheless, the ozone charge transfer complex

would have to be more than one unit more acidic and also quite long-livd to show this

effect. In this context, it is intriguing that only a very small, if any, deviation is tbund for

thymidine (inset in Figure 5.1), uracil and 5-chlorouracil (Figure 5.4) and also no such

deviation is found for the cytosine, adenine and guanine systen~s (see below).

Figure 5.3. A typical nwe obtained by the bleaching of indigo (15 x 10.' mol dm;') by ozone (8 x 1 o*' mol dm-3) in the case of 6-methyluracil (4.5 x 1 O3 mol dm-') at pH 2.1. Inset: logarithmic plot of the curve

0.24

- 5 0 0.20 0 w - V1 II

0.16

- 1

-

- o 20 40 60 80

Time (6%)

1 1 1 I I l l 1 0 20 40 60 80 100

Time (sec)

Rate constants undprr)Juet shdies ... . 127

The formation of *OH as the reason for this unexpected deviation has been

excluded for the thymine and 6-methyluracil systems. This has been done by

adding tert-butylalcohol in excess (in the absence of a competitor) at around the pH

where the deviations are most noticeable. No formaldehyde (due to 'OH-induced

tert-butylalcohol degradation) was detected. This is in agreement with the

observation that at high pH, addition of tert-butylalcohol had no effect on the rate

of reaction (nitrite as competitor).

Figure 5.4. Obsaved rate constant of the reaction of ozone with (a) 5-chlorouracil and (b) uracil as a function of pH. The solid line is calculated on the basis of the rate constants measured at pH 2.2 and 9.2 and the pK, value of 8.02 in the case of 5-chlorouracil and that in the case of uracil are at pH 4.5 and 1 1.3 and the pK, value of 9.5. Rate constants are measured using buten-3-01 (o), indigo (A) and nitrite (m) methods

It has been mentioned above that there is an increase in rate in the range

where the deprotonation of the second nitrogen of thymine occurs. Thc rate

constant of this doubly deprotonated species might have been underestimated

(experiments could not be extended beyond pH 13; the high salt load prevented

nitrate determinations by ion chromatography). Yet, the marked deviation of the data in

Rate constants and product studies ... . 128

the pH range 9- 12.5 from a slope of unity indicates that even the assumption of a much

higher rate constant for doubly charged thymine can not be explained the observed

deviation from the expected pH dependence.

It is remarkable that the position of the methyl group, i-a when the rate of reaction

with thymine is compared with that of 6-methyl uracil, has such a dramatic effect (a factor

of 300, c.f. Table 5.1). Considerable differences in the rate constants of isomers, e.g.

1 , 1 dichloroethene (k = 1 1 0 dmhmol" s"), cix- 1,2dichloroethene (k = 540 dm3 moi- ' s-' )

and trans-] 2-dichloroethene (6500 dm3 mol-' s-' ) have been noticed before,'" but there

the rate constants differed by a factor of 60, at the most. Even more surprising is the

observation that the rate constant of 6-methyl uracil is lower than that of uracil.

Typically, m additional methyl group at the reacting C=C double bond increases

the rate of reaction by a factor of -4.'" A very marked difference between these

uracil derivatives and simple olefins is also noticed when 6-methyl uracil is

compared with 5-chloro uracil. One would have expected that the former reacts

considerably faster with ozone than the latter, but the reverse has been observed

(cJ Table 5.1).

The observed increase in rate upon deprotonation of thyrmne is connected with a

change in the reaction mechanism. While the neutral nucleobase does not give rise to the

h a t i o n of singlet dioxygen (o,'AJ, de~rotonated thymine yields Q' A, in -8% y ~ e l d . ~ ~ ) ~

Possibly, it is formed via reactions (5.4H5.6) (Scheme 5.1 ).

Rate constmts andproduct studies .... . 129

Thymine deprotonates at N(1) as well as at N(3) [equilibrium (5.4)l. Dipolar

addition of ozone to thymine deprotonated at N(1) gves rise to an isopyrimidine

hydrotiioxide [reaction (5.6)]. Since the chemistry of these short-lived intermediates is

k n ~ w n , ~ ~ . ' ' * a product study which is on the way may help to elucidate mechanistic

details. Thus far, our mechanistic proposal is supported by the observation that

thymidine which can deprotonate only at N(3) does not afford any singlet dioxygen

at high h this context, it may be worth mentioning that 5-chloro uracil

shows this change in mechanism even more strongly; no singlet dioxygen (O,'A,) is

formed at pH -3.5, but when the 5-chloro uracilate ion predominates, it becomes a

major product (-42%).205

Scheme 5.1

5.2.2. Cytosine, cytidine and 2'-deoxycytidine

As can be seen from Figure 5.5, cytidine and 2'-deoxycytidine react

somewhat faster than the free base. This is most likely due to the electron-donating

Rate constants and product studies .... 130

property of the sugar moiety. Upon protonation, their rate of reaction drops by two

orders of magnitude. Whereas the nucleosides do not show an increase in rate at

high pH, cytosine does. At high pH, the latter deprotonates (pK, = 12.2). A similar p&

value (pK, = 12.5) is reported for cytidine, but here, the deprotonation is suggested to

occur at C(2')?I3 Apparently deprotonation of the sugar moiety hsts practically no effect on

the reactivity of these compounds with ozone.

Figure 5.5. Observed rate constant of the reaction of ozone with cytosine (m), cytidine (A) and 2'-dwxycytidine (A) as a function of pH. The solid line is calculated on the basis of the rate constants and the pK, values given in Table 5.1

5.2.3. Adenine, adenosine and 2'-deoxyadenosine

In the case of adenine and its derivatives, the addition of ted-butylalcohol on the

observed rate of reaction (Fibwe 5.6) has a dramatic effect (cj.' a1 so refs. I 9, 1 94). In its

absence, the rate constant with adenine became too fast for the indigo method, i.e.

the apparent bimolecular rate constant must have been fgstcr than -10' h' mol-I s-'.

Rute constants und aroduct studies ... . 131

This points to the intermediacy of 'OH at one stage. Indeed, when m-butylalcohol is

added in large excess (0.1 mol M3), we observed the formation of considerable amounts

of formaldehyde (-0.2 mol per rnol ozone) in the case of adenine as well as

2'deoxyadenosine. Formaldehyde is an important product of the *OH-induced oxidation

of terr- b~ty ia lcohd .~ The observed formaldehyde yield of -20% relates to -60% 'OH

(the use of ted-butylalcohol in the determination of *OH yield in ozmation reactions has

been understood recently).2bJ

Figure 5.6. Observed rate constant of the reaction of ozone with adenosine (a, Cl), and 2'-deoxyadenosine (A, A) as a function of pH; the solid line is calculated on the basis on an assumed limiting rate constant of 1.3 x 1 O5 dm3 mol-' s" as in adenine and the pK, values given in Table 5.1. Inset: adenine; the solid line i s calculatecl on the basis of the rate constants and pK, values given in Table 5.1. Solid symbols refer to experiments using the indigo method, open symbols to competition kinetics with nitrite as competitor. In all experiments, tert-butylalcohol (0.1 rnol dm-3) was added as 'OH scavenger

Rate conslunl.r. and product studicLs ... . -

132

The reduction potential of adenine and its derivatives15 is too high for an

electron transfer to occur in the ozone reaction (note that in the case of guanosine

which has a much lower reduction p ~ t e n t i a l ' ~ ? * ~ ~ no 'OH is formed, see below). Thus

there must be another reason for the intermediacy of 'OH in this system. The most

likely precursor for 'OH is considered to bc 0,". Here, it suffices to mention that

tetranitromethane added as 02' scavenger, is substantially degraded (via the

intermdacy of nitro form anions). Addition of tert-butylalcohol reduces the observed

rate constant by more than a factor of two. This indicates that the radicals that are

formed upon 'OH attack on adenine/2'-deoxyadenosine must induce a chain reaction.

The rea&on of ozone with the cyanide ion also proceeds via a chain reaction.'" 'fiere,

tert-butylalcohol cannot fully suppress it, i.e. there must be an additional chain carrier.

A similar situation may prevail here, but further work will be required to elucidate this

system in more detail.

In the case of adenine, the pH dependence of rate constant can be adquatcly

fitted by the rate constants and the pK, values given in Table 5.1 (cf. solid line in the inset

of Figure 5.6). The rate of reaction with denosine and 2'-deoxyadenosine also increasa

with increasing pH, and these data can he fitted (solid line in Fibwe 5.6) if the p& of 12.5

and the limiting rate constant of adenine are taken. It has been noted above that in thc case

of the cytosine nucleosides, no further i n m e of the rate of reaction has been observed in

the high pH range (c: f.' Fib- 5.5). This has been explained by the fact that the pK. value

at very high pH is athibutad to a pK, value of the sugar moiety and that these anions do

Rute constants and product studies ... . 133 - --

not react with a rate higher than 3.5 x lo3 drnho l - ' s-' given by neutral cytidind

2'-deoxycytidine themselves. At the highest pH that can be measured with some

confidence, the adenosind 2'-deoxyadenosine rate wnstant approaches - 10" dm3 mol-I s",

not much higher than the above value. There is no indication that this value is a l i~ni ting

one, and indeed it is impossible to fit the data with such a low value. The solid line

in Fibure 5.6 has been constructed by taking the pK, of 12.5 and the limiting value

of the adenine system (1.3 x 1 O5 dms mol" s-') . This is only acceptable as long as

the pK, at -12.5 is due to the deprotonation of the adenine moiety. We have

remeasured the absorption spectrum of 2'-deoxyadenosine at pH 14 and at the long-

wavelength band, no changes, neither E nor h,,,, with respect to its spectrum at pH 7 are

observed (below 230 nm, the high absorption of OH- prevents firm measurements). This

would be in agreement with the earlier conclusion that the pK, at -12.5 is due to a

deprotonation of the sugar moiety,"'"*~" '."' however not with the observed high

reactivity. Thus, the reason for this unexpected observation remains unresolved.

The earlier determination of the rate constant of an adenine derivative,

5'-dAMP,'Y4 deserves a comment. Its rate constant has been estimated at 2 x lo2 dm3

mol-' s-' (in thc absence of terf-butyldcohol) by measuring its degradation relative to that

of dCMP for which a rate constant of 1.4 x 1 O3 dm.' mol-' s" had bcen determined by the

stopped-flow technique under close to second-order conditions. In the presence of

teut-butylalwhol, the relative rate constant dropped by nearly an order of' magnitude. In

view of this, these data agree reasonably with our value of -15 dm' mol-' s-'. It has been

Rate constunts andproduct studieLv .... 134 --

mentioned above that in the absence of tert-butyldcohol, we estimate that the observed

rate constant of adenine and its derivatives is 2 10' dmbmoI"' s". This is hgher than the

value of 2 x 1 d dm3 mol-' s" obtained in the dCMP competition. Th~s is not surprising.

When 'OH radicals play a significant role in this system, they also attack dCMP

which is subsequently consumed, i.e, the estimate of the dAMP rate constant comes

out too low,

52.4. Guanosine and 2'-deoxyguanosine

Guanosine and 2'-deoxyguanosine react about equally fast with ozone

(Table 5.1). Upon deprotonation, the rate of reaction increases 250-fold. As can be

seen h m Figure 5.7, the experimental data can be adquately fitted using the rate

mmtants and pK, values given in Table 5.1 (experiments at very low pH indicate that

upon protonation the rate constant falls below 300 dm' mol-' s-', data not shown). Addition

of twt-butylalwhol had no effect on the rate of reaction nor was any formaldehyde formed

under these conditions. This excludes any electron tmnsfkr h i n the guanine moiety to

ozone, although guanine has the lowest reduction potential of all nucleoba~es"~~ and

mpecially in basic solutions where its anion predominates, it is readily oxidised by many

otherwise only weakly oxidising agents (c$ ref: 3 1 6).

Rate constants and product studies .... 135

Figure 5.7. Observed rate constant of the reaction of ozone with guanosine as a function of pH; the solid line is calculated on the basis of the rate constants and the pK, value given in Table 5.1. Competition with 3-butenol (a) and nitnte with (A) and without (A) toy-butylalcohol(0. I mol dm5

5.2.5. DNA

As has been noticed before," 'OH plays an important role in the reaction of ozone

with DNA. From ref, 1 9 and the present study, it is clear that 'OH radical formation must

be due to the reaction of ozone with the adenine moiety. For the determination of the

intrinsic ozone rate constant with DNA, tert-butylalcohol had to be added. Under

such conditions, the rate of reaction o f DNA is only 450 dm3 moi-' s-' (in the absence of

tcrt-butylalcmhol k, = 1 . I x 10' dm.' mol-' s-'), i.e. much lower than that of the weighted

average of the nucleubases. In the case of 'OH, which reacts with the nucleobases and

their derivatives at close to diffusion-controlled rates (k - 3 x 10' dm3 mol-' s"),& the rate

of reaction of 'OH with DNA is considerably lower (k = 2.5 x I I)s dm-' rnol-' s-'),"'~ since in

Rate comtants undproducl studies ... . .. - - - - 136

th~s non-homogeneous reaction with the macromolecule DNA, two terms, a diffusion

term (k,,,) and a reaction term (k) have to be wnsidmedNR The observed overall rate

constant (k&) is the harmonic mean of these two rate constants [equation (5.711.

Since, in contrast to 'OH, ozone reacts with the nucleobases at rates much

below the diffusion-controlled limit, the second term must fall away, and the rate of

reaction of ozone with the nucleic acids is only given by the first term, ie. it should

be close to that of the weighted average of the concentndtions of the various nuclmbaqes

in the nucleic acid times their rate constants with owne. This is not observed. Structural

effects such as hydrogen bonding between the nucleobases may he a reason lbr the

strong reduction in rate.

Corresponding experiments with RNA were not carried out, because it was

not guaranteed that the commercially available RNA is of similar purity and

sufficiently double-stranded to yield complementary data. From the rate constants

given in Table 5.1, one would assume that in RNA, the guanine moiety is the most

likely one to become degraded upon ozone treatment. This has been indeed observed.""

In DNA, the situation might be somewhat diffcrcnt. Besides guanine, thymine may be

the other preferred target.

Rate constants and producl studies .... -- - 137

5.3. Product studies of the reactions of ozone with thymine and thymidine

According to a detail4 rate constant and product study involving the

characterisation of short-lived hydroxyalkylhydroperoxides, it is known that there

is a considerable influence of the elecbon-donatindwithdrawing power of substituents

(D vs. A) on the final products [reactions (5.8)-(5.12)].IW An electron-donating group

D such as a methyl substituent favours the formation of the hydroxyhydroperoxide at its

position [reactions (5.9) and (5.1 l)]. Th~s has been rationalised by a stabilisation of the

carbocation formed in d o n (5.9).

I HO-C-

H e - 0

Scheme 5.2

In thymine and thymidine, ozone will add to the C(5)-C(6) double bond.

Due to the electron-donating properties of CH, group, the Criegee intermediate

formed as in reaction (5.8) will preferentially open according to reaction (5.9). It

Rate constunts and product studies ... . 138

will be shown that the substituents at N( I ) (H in thymine, Zdeoxyribosyl in thymidine)

have a pronounced effect on the subsequent reactions of the carbocation formed in

reachon (5.9). A detailed analysis of the ozonation products and intermediates of

thymine and thymidine are presented below.

5.3.1, Thymine

Ozone reacts with thymine at a rate constant of 4.2 x lo4 dm" mol-' s-I, and

upon deprotonation (pK, = 9.9), the rate constant increases to 3 x 10' dmmolL ' s-'.'

At pH 6.5, where the majority of experiments have been carried out, the

contribution of the anion in product formation can be neglected. Under typical

experimental conditions ([thymine] = 1 x 10" mol dm", [O,] = 1 x 1 OU4 in01 dm-j),

the half-life is 0.016 s, i.e. the reaction proceeds at the time scale of mixing the two

solutions.

5.3.1 .I. Conductometry and ion chromatography

When ozonolysis is carried out in a conductometric cell, the conductance

rises with a biphasic kinetics (Figure 5.8). The first step ("prompt") is too fast to be

resolved kinetically, the second one ("slow") is of first-order kinetics and occurs

with a rate constant of 1.1 x 10" s-' at 1 8°C. When the experiinents were carried out

at 3 'C, the conductance build-up becomes slower by approximately eight times.

The rate constant of the second increase in conductance at 3 "C is 1.3 x 1 0-4 s-'. The

rate constants relevant to this study are compiled in Table 5.2.

Rlite constanis und product studies.. . . 139

*.--- - a-- - - - m - - - -m ,..-* -

/,r- -. .. .

i o f ..- ...,-., -.. I . . 3 - 7 1 "D .,,,,,,

$1 ' " 1 .mh l

Y

Timeimin Figure 5.8. Ozonolysis of thymine in aqueous solution at 18 "C. Formation of acid

(expressed as mol acid per mol ozone consumed) as a function of time followed by conductance measurements. Inset: First-order kinetic plot of the slow part of conductance increase at 3 "C ($3) and 20 OC (m)

The calibration of the conductometnc set-up has been done with sulfuric acid. The

acids that are formed in this system are weak acids, the acidic hydroperoxide

6 ( 1 -Hydropaoxymethylen-3-(2-0x0-pmpanoyl) pK, = 4.0, see below) and formic

acid (p&=3.75) and are hence partially pmtonated ([ozone] = 2 x 1 o4 In01 dm-' in these

experiments). The "prompt" yield of acid (1 -Hydroperoxymethylm-3-(2-oxo-prop1oyl)-

urea} is calculated as -34% (with respect to consumed ozone). Further acid (formic acid)

is released in the slow process whereby its yield increases to -80% (Figure 5.8).

The formic acid formed during the reaction is determined by ion chromatography

(IC) and its yield was 75% after ozonolysis. Since the eluent is basic, the release of formic

acid speeded up in the column. When ozonated thymine solutions were treated with a base

(KOH, pH 10.8,90 min, I 8 "C), f i e r formic acid is released, and the yield is being now

100%. The yelds of various species are compiled in Table 5.3.

Rute constants and product studies .... 140

Table 5.2. Compilation of rate constants relevant to the ozonolysis of thymine

Reaction Thymine + 0, -+ products

Rate constant 3.4 x l 04 dm3 mol-' s-'

Thymine anion + 0, ---b products "Prompt" release of 6 "Slow" formic acid release, 18 "C

6-8 t 1- -+ products 43 dm3 mol-' s-'

19 + I - --+ products 7.5 dm3 mol-' s-' 6-8 + Fe(CN):- --b products 0.4 dm3 rnol-' s-' 19 + F ~ ( c N ) ~ - -b products 1 d m h o l - ' s-'

4.2 x 1 04 dm3 mol-' s-'* 3 x 1 O6 dm3 mol" s-" > 70 s-' 1.1 x 10"s- '

"Slow" formic acid release, 3 "C Decay of 6, 1 8°C Slow H,O, release, 1 8 O C

* b r n 113.6; for detailY about productslintermediatm see sation 5.3. I. h

1.3 S-I

1 .O 1 u3 s“ -1 x 10" s-'

HN H N OOH

0 0-OH H

H 6 7 8 14 19

In order to test whether the fast conductance increase can be resolved hetically,

stopped-flow experiments with conductometric detection were k c d out. As can be seen

from Figure 5.9, the build-up of conductance foliows pseudo f i s t order-kinetics, and the

rate of reaction depends on the thymine concentration. Thc rate constant derived from the

data shown in the inset of Figure 5.9 ( k = 3.4 x 1 o4 dm-' mol-' s-') is in agreement with

the value obtained by competition kinetics (4.2 x lo4 dm3 mol" s" ) .V t is thus

concluded that the rate-determining step in this fast conductance increase is the

reaction of ozone with thymine, i.e. these data do not give any further hint to the

nature of the intermediate that may be involved in the fast release of a proton.

Rulr ronsttmts and product stuu'le,~. . - . - - - -. -. 141 -

Figure 5.9. Ozonolysis of thymine ( I x 10.' rnol dm-') in aqueous solution at 18°C. Increase of conductance in the reaction of ozone with thymine as followed by stopped-flow with conductometric detection. Inset: k,,, as a function of thymine concentration

0.8

> 0.6 . ?L d 0.4

0.2

Table 5.3. Compilation of yields (with respect to ozone consumed) in the ozonolysis of thymine

-.. --- . . . .. .

-

-

-

- [Thymine] I mM

I

0.2 0.3 0.4 0.5 Time / s

'measured using ion chromatography, ' from ref. 205

Process "Prompt" acid formation (conductometry) "Slow" acid release (conductometry) Formic acid release* Total formic acid rebse at high pH* Total hydroperoxide (immediate) Total hydroperoxide (after 2 h) Hydrogen peroxide (immediate, catalase assay) Hydrogen peroxide (immediate, R,S assay) Hydrogen peroxide (after I h) Organic hydroperoxides (immediate) Organic hydroperoxides (after 1 h) 1 -Hydroperoxymethylen-3-(2-0x0-propanoyl-urea 6 5-Hydroxy-5-methylhydantoin,l4 (aRer R,S treatment)

5-Hydroxy-5-methylhydantoin, 14 (after R,S and OH- treatment) Singlet dioxygen (at high p ~ ) '

.-

Yield / % 34 50 75 100 100 93 25 2 5

40 75 53 34 h7

1 00

8 --

Rule constants an J product studies.. . . -. , 142 --

5.3.1~2. Formation and decay of hydroperoxides

The yield of total hydroperoxide is 1 000h as assayed with molybdate-activated

i d d e (Figure 5.10). Upon treatment of the ozonated solution with catalase, the

hydroperoxide yield is reduced to 75%. Catalase is well known for complete1 y degrading

H,O,. Thus, the yield of H,O, present in right after ozonolysis is 25%. However, we

have observed that formic peracid, a conceivable product in the present system, is

also rapidly degraded by catalasq2" and further experiments have been carried out

to ascertain that the hydroperoxide that is eliminated by catalase is indeed H,O,. Recent

experiments demonstrated that some organic sulfides can be successfully employed to

distinguish between highly reactive organic peroxides and ~ , 0 ~ ? ~ ~ * ~ ~ ~ For example,

bis(2-hydroxyethy1)mlfide rapidly reacts with formic peracid (k = 220 d m b o l " sL')2M

yielding formic acid [i.e. leading to an incrmse in conductance, pK,(HC(O)OOH) = 7.1,

pK(HC(0)OH = 3.81, whereas its reaction with H,O, is negbgible in comparison.

However, when bis(2-hydroxyethy1)sulfide ( 1 x 10'%01 dm") has been

added to the ozonated thymine solution, no additional increase of conductance i s

observed and the slow increase in conductance shown in Figure 5.13 is suppressed.

Instead, a dec,rease of conductance (k = 50.6 dm3 mol-' s") is observed. This can be

taken as the evidence for the fast increase of conductance is due to an acidic

peroxide 6. The slow increase of conductance also must have a hydroperoxide

precursor, and formic peracid is not a product of the ozonolysis of thymine.

Rate comfunt,~ undproduct studies ... . 143 ---.--

Therefore, it was of interest to evaluate the pK, value and optical features of this

intermediate, which is described in section 5.3.1 -3.

Figure 5.10. Ozone dependence in the formation of total hydroperoxide I.) and formic acid by HPIC (0) after the ozonolysis of thymine at natural pH. Inset: Ozone dependence in H,O, formation after the desttvction of organic peroxide by bis(2-hydroxyethy1)sulfide

Thc reaction of bis(2-hydroxyethy1)sulfide cannot be used to determine the

yield of reactive hydroperoxides, since its sulfoxide is difficult to detect by HPLC

due to its low absorption coefficient in the accessible wavelength regon. However,

methionine also undergoes the corresponding reaction with reactive hydroperoxides, and

its sulfoxide can be readily detected.zM Using this assay, the yicld of sulfidereactive

hydroperoxides is 75%. A oomplenlentary value (25%) has been determined by

measuring the remaining H,O, yield a h the destru&on of sulfide-reactive

hydroperoxides by bis(2-hydroxyethy1)sulfide (Figure 5.1 0) using UV measurements.

It has been shown that H,O, does not react with bis(2-hydroxyethy1)sulfide at an

appreciable rate.'M Based on the catalase and the sulfide assays, it is hence concludd that

Hare constants und oraduct studies ... . 144

H,O, is present in 25% yield right after ozonation. Thus, at least three kinds of

hydroperoxides arc formed immediately upon ozonation, an acidic hy droperoxide

(assigned to 6 , see below), neutral hydroperoxides (assigned to 7 and 8, see below)

and H,O,. The organic hydroperoxide rmaining after one hour is a further neutral

hydroperoxide (5-Hydropaoxy-5-methy lhydantoin 19, see section 5 -3.1 .h).

Figure 5.1 1. Normalised yields of total hydroperoxides (: 1) and hydrogen peroxide (a) as a function of time after the ozonolysis of thymine. Hydrogen peroxide yi&s were determined after the destruction of organic hydroperoxides by bis(2-hydroxyethyl) sulfide

As can be seen from Figure 5.1 1, the total hydroperoxide yield is not much

(-7%) derreased over time. A k t i o n of the sulfidsreactive p x i d e s decays yielding

mainly H,O,. This reaction proceeds with a half-life of - 12 min according to a co~nputer

fit through the data (m in Figure 5. L 1 ). Thus, it follows the wnc kinetics as the "slow"

conductance i n m e reported above. i.e. a new organic hydroperoxide (assigned to 19,

see below) appears during the decay of the initial peroxides, l 'he yield of H,O, during this

decay of the primarily tbrmed hydroperoxides (6-8) i s much lower (-1 5%) than that of

Rate constants und product studie.~ . . . . , -

145

formic acid (-75%), The loss of total hydmperoxide yield (-7%) also occurs at the same

time scale (c$ Figure 5.1 1).

Hydrogen peroxide does not react with iodide at an appreciable rate unless

activated by molybdate, but the more reactive hydroperoxides do. The

hydroperoxides present right after the ozonolysis (6-8) react with iodide more

rapidly (k = 43 dm" mol-' s-', determined by stopped flow) than the remaining aRer

one hour (19, k = 7.5 dm3 mol" s-'). Accordingly, the apparent k,, of the reaction

changes with time, since the rate constants of these hydroperoxides differ only by a

factor of -5.7, i.e, these two reactions are not well separated kinetically. The

change in k,,, with time gives a half-life of - 13 min, which is close to the half-life

of the intermediate 6a (deprotonated fbrm of 6), at 255 nm (c.J Figure 5.14).

The reactivity of the organic hydroperoxides (immediate 6-8 and delayed 19)

towards Fe(CN);- is very similar. Immediately after ozonolysis and destruction of

H,O, with catalase, the rate constant of the then prevailing hydropmoxides 6-8 was

k = 0.4 dm3 mol-' s-' and after keeping the samples for one hour to let 6-8 decay into

19, the rate constant was k = 1 dm%nolK' s '.

The presence of H,O, and other organic peroxides was observd by post

column derivatisation. A 25% yield of H,O, obtained at naturai pH by post coluinn

technique, which is well agreeable with other assays. Due to the lack of authentic

samples, the yield of other peroxides could not establish by this method.

5.3.1.3. HPLC and UV-spectrometry

Immediately after ozonolysis, a very prominent product (assigned to 6a)

with h,,, = 256 nm (c.f Figure 5.1 2) was ohsewed by HPLC on a rcversed-phase

column with retention time 7.6 rnin using water as eluent. There is also a second

product ( 1 -formyI-5-hydroxy-5-methylhydantoin 1 2) eluting at 8 .3 min. This has

only an end-absorption tailing towards 220 nm. Both products must be more polar

than thymine, since they elute much earlier than thymine ( 1 5.6 rnin).

h l n m k l nrn

Figure 5.12. Ozonolysis of thymine at natural pH: UV absorption spectra of the products obtained using HPLC with diode array detector. Frame A: a, anion of 1 -hydroperoxymethylen-3-(2-0x0-propanoy1)-urea 6a (7.6 rnin); ,:I, assigned to 5-hydroperoxy-5-methylhydantoin 19 (6 rnin). Frame B : A, assigned to I -formyl-5-hydroxy-5-methylhydantoin 12 (8.3 min); A, 5-hydroxy -5-methy I hydantoin 14 (5.4 min, reference ~naterial was avail able)

As will be discussed below, we are forced to account for further organic hydropxides

(assigned to 7 and 8). These products arc expected to have only little W-absorption as

compared to 6a and m y be masked if they have retention times very close to that of 6a.

Hate consfants and ~roduct studies ... . 147

The short retention time of 6a with water as eluent is due to the fact that under this

condition 6 is dqmtonated. When the pH of the eluent is decreased, 6 elutes sorriewhat

later with a concomitant shift of its absorption maximum towards 237 nrn (c:$ inset in

Fibye 5.13). From the data shown in the main graph of Figwe 5.13, the pK, value of 6 is

calculated as 4.0.

Figure 5.13. Percentage of undissociated 1 -hydroperoxyrnethylen-3-(2-ow- propanoy1)-urea 6 as a function of pH as obtained by HPLC with diode array detector (see text). UV spectra o f 6 (solid line: h,, =

237 nrn, eluent: water at pH 2.6) and its anion (6a, dotted line: h,,,,, = 256 nm, eluent: water at pH 7).

The hydroperoxide 6 decays by the same kinetics (Figure 5.1 4, inset: k = 1.0 x 1 0-3 S-I)

as found for the release of formic acid (c$ Figure 5.8). It gives rise to the hydroperoxide

19 which elutes at 6.0 min.

Ratr constants und product sludies... -- -- -. --.-- 148 - ---

20 40 60 Time / mln

- 20 60 100 140

Time / rnin

Figure 5.14. HPLC analysis of ozonatd thymine. Decay of 1 -hydropmxyrnethylene- 3-(2-oxepropanayl)-urea 6(A) and build-up uf 5-hydroperoxy -5- methylhydantoin 19 (m) as a hnction of time. Inset: first-order kinetic plots of the data

The hydroperoxides were wiped out upon the addition of bis(2-hydroxyethyl)

sulfide which shows that these products must be strongly oxidising hydroperoxides.

The hydroperoxide 6 (together with 7 and 8, see below) is rcducd to 12, while 19 is

reduced to 14. The yield of 5-hydroxy-5-methylhydantoin 14 after the completion of the

reduction by the sulfide is 67% of omnc consumd (reference materid for the

quantification of 14 was available). When this solution was treated with NaOH at pl4 10.5

for overnight, product 12 (1 -formyl-5-hydroxy-5-methylhydantoin) also d i s a p p d ,

and the yleld of 5-hydroxy-5-methy\hydantoin 14 has been increased to 100%.

Hydantoin 14 was also confirmed by GC-MS analysis. 0;conatd samples after

rotary evapratiun were subjected to silylation to I'MS ethers fbr betier volatility and

monitored the GC-MS. Three times silylation occurrod (MW - 346) and the prominent

peaks with m/z (%) are 33 1 (loo), 21 6(7), 147(46), 73(h3).

Rute constants and aroduct studies .... 149

HPLC was used to study the acid-base behaviour of the species 6. Thymine

was ozonated at varying pHs (from pH 1-7) and immediately injected into HPLC

with water (with same pH as the solution) as eluent. pH of the thymine solutions

and eluent were varied by the addition of appropriate amount of sulfuric acid or

sodium hydroxide. The spectrum of the first eluting peak was monitored at each

expenmental pH and it was strongly pH dependent. At pH 2, this hydroperoxide

intermediate has a maxima at 237 nm and at pH above 6 is 256 nm. Measuring the ratio

of the absorbances at 237 and 256 nrn, A2371A256, it was possible to calculate the

percentage of protonated form for each experimental pH point (see Figure 5.1 3) and

determined the pK, value of hydroperoxide as 4.0.

The molar absorption coefficient of the anion of 6 at 256 nm must be higher

than that of thymine at the same wavelength. When equal volumes of thymine

(4 x 1W4 mol dm-3) and ozone (2.25 x lou4 11-101 dm-') solutions were mixed, the

absorption at 256 nm rose by a factor of 1.18. Taking the dilution by ozone solution

into account and subsequent decay again (k = 1.0 x 1 0-3 s- ' , ses above) to a level which is

compatible with the expected remaining thymine mncentmtion and no significant

absorption of the remaining products at 256 nm. Taklng the yield of h c species absorbing

at 256 nm as 34% (based on the imnediate conductance formation, c,f: Figure 5.81, its

molar absorption coefficient is calculated to he 2.5 x 1 o4 dm3 mol-' cm-' , i.c. more than

twice of that of thymine at its maximum at 265 nrn (E = 7.9 x 10' dm3 cm-' s-I).

Ku!r com~tants ur~dproduut studies. . . -- 1.50

.- - - --

5.3.1.4. Liquid Chromatography-Mass Spectrometry

At 3"C, the decay of 6 is much slower (1.3 x lo4 s-', c.j: inset it] Figure 5.8)

than at roonl temperature. This eight-fold pmlongtd 11fe time enabled us to carry out

LC-MS measurements with a sample that had been ozonated at 3OC using

acetonitrilelwater (1: 1) as eluent in the positive mode of electrospray ionisation

(Figure 5.15). Smng m/z values of 175 (M+ 1)' and 197 (M-tNa)' was observed

indicating that the molecular weight of 6 is 174 Da. Product 12 does not gve rise to a

mass spectrum under these expenmental conditions. When 6 has decayed, the mass

spectrum of 19 could be taken md this showed a pronounced i / z = 147(M+l)', and

169 (M+Na)' . Also, it gives two strong dimer peaks with m/z values 293 (2M-t 1 )'

and 315 (2M+Na)' i.e. its molecular weight is 146 Da. Similar to product 12, the

final product 5-hydroxy-5-methylhydantoin 14 (authentic reference material was

available) also did not give a mass spectrum under these conditions.

As mentioned above, 6 and 19 are hydroperoxides and are eliminated by the

addition of bis-(2-hydroxyethy1)-sulfide. As a consequence of this, after the sulfidc

beatment, compounds having an LC-MS response are no longer present except for the

resulting sulfoxide (MW = 138 Da) which shows a pronounced signal at mdz = 139,

(M-t I ) ' . The parent sulfide does not give an LC-MS signal.

Rate auastunts and product sfudics ... . 151

l o o (b)

4 A

Timelmin

Figure 5. 15. LC-ES-MS analaysis of ~zonated thymine at natural pH (a) ES-MS spectra of I - hydroperoxymethylen-3-(2-0x0-propanoy1)-urea 6 (b) ES-MS spectra of 5-hydroperoxy-5-methyl hydantoin 19 and (c) chrornat&ram at 230 nm

5.3.1.5. Assignment of the products

Product 6 has a molecular weight of' 174 Da. i . ~ . it contains three oxygen

atoms more than thymine. It is a strongly oxidising hydroperoxide, i.e. the

hydroperoxide funchon must be activated by electron-withdrawing groups. It is fairly a

strong acid IpK, = 4.0). The spectral shift k r n 237 nm to 256 nm upon deprotonation

requires a considerable conjugation of the n system. Its structure is assigned to

I -hydropemxymethylen-3-(2-0x0-propanoy1)-um 6. Product 12 is not peroxide. Upon

basic hydrolysis, it is converted into 5-hydroxy-5-methylhy dan toin 1 4 and formic acid.

So product 12 is assigned to 1 -formyl-5-hydroxy -5-methylhydantoin.

Product 19 is a hydroperoxide with the molecular weight of 146 Da. Its

precursor is 6 (and 7 plus 8). Upon reduction of 19 with sulfide, 5-hydroxy-5-

lnethylhydantoin 14 is formed. Hence 19 is assigned to S-hydropmxy-5-

methylhydantoin.

5.3.1.6. Mechanistic aspects

The release of formic acid and other conducting species at various steps has

considerable importance in the proposed reaction mechanism. It is observed that fonnic

acid is released in two steps, "slow" and "very slaw" (hydrolysis at high pH). The

"prompt" increase in condu-ce is due to the formation of an unstable species 6 , which

has a pK, 4.0, The formation and decay of this acidic organic hydroperoxide,

which shows a very strong UV absorption, have to be accounted.

Rat< conslants an Jproduut studies ... . -- 153

Ozone is a strong electrophilic agent and will hence preferentially add to the

C(5)-position of thymine 1 [reaction (5.14)). The zwitter ion 2 thus formed will

close the ring thereby forming the Criegee intermediate 3 [reaction (5.1 5)l.

Scheme 5.3

The Criegee intermediate 3 opens according to reaction (5.16) yielding

zwitter ion 4 (Scheme 5.4). The highly absorbing peroxide 6 is formed by losing a

H at N(1) position of 4 [reaction (5.18)]. Formation of this peroxidic intermediate is

responsible for the initial fast conductance ("prompt") increase. Hydroperoxides

have generally high pK, values [c..f: p&(H,O,) = 1 1.8; pK(HC(0)UOH) = 7. I], and

the hydroperoxide hnction of 6 will thus be protonated at around pH 7. In

competition with reaction (5.1 81, the zwitterion 4 may react with water yielding thc

a-hydroxyhydroperoxide 7 [reaction (5.19)]. One may also consider that the

zwitterion 4 may form 8 upon losing the acidic proton at N(3) [reaction (5.20)].

However, the strain exerted by the four-membered ring may not be much in favour

of this reaction. The hydroperoxide 6 has a conjugated x system and shows a strong

absorption at 237 nm (,c$ Figure 5.13). The proton at the N(3) is acidic

[equilibrium (5.23), pK,(6) = 4.01, and the absorption maximum is shifted to 256

mn upon deprotonation. The fast build-up of conductivity that occurs at the same

Rate consfants and product studies .... 154

time scale as the reaction of thymine 1 with ozone (Figure 5.9) is due to the

formation of 6. In the sequence of reactions that lead to 6a and H', the rate-limiting

step is the reaction of ozone with 1 [reaction (5.14)]. After correcting for some

undissociated 6 due to its pK, value of 4.0 calculated from the conductivity data

shown in Figure 5.8, the yield of 6 is close to 34% of ozone consumed.

Subsequent hydrolytic process of 6 lead to the release of formic acid

(t,:, 13 min at room temperature) monitored by the "slow" increase in

conductivity (Figure 5.8). These conductivity data suggest that the combined yields

of the formic-acid-releasing hydroperoxides (attributed to 6-8) is close to 80%, I.e.

the yield of the sum of 7 and 8 is around 46%.

The kinetics of this "slow" rise in conductivity is mirrored by a very similar

kinetics of the disappearance of the optical absorption of 6 . The fact that 6 has a

pK, value of 4.0 and that of formic acid is 3.75 warrants the presence of additional

precursors of the fbrmic acid that is released in the "slow" process. It is suggested

that 7 and 8 contribute here. The evaluation of build-up kinetics are fraught with

considerable errors, especially when more than one species is involved.720 When the

individual rate constants are not dmqtically different, they cannot be disentangled. The

plots shown in the inset of Fibwre 5.8 only indicates the marked reduction in rate of

hydrolysis upon reducing the temperature to 3OC. This enabled us to run the mass spectra

of the hydroperoxides by HPLC/MS. The molecular wei@t of 6 and 8 is 1 74 Da, agreeing

with mass spectral data. There is no mass-spectral evidence for the formation of' 7, but in

Rate constunts und product studies. .. . - . - - - 15s

the view of the absence of any mass-spectral response in the case of 12 and 14, this lack of

confirmation does not carry much weight and does not rule out the formati011 of 7. ARm

hydrolysis, a hydroperoxide 19 with the molecular weight 1 46 Da remains.

Scheme 5.4

In competition with reaction (5.1 61, the Criegee intermediate 3 can decay

according to reaction (5.17). The zwitter ion 5 may give rise to the hydroperoxides 9

and 10 [reactions (5.2 1 )-(522)]. The hydroperoxide 10 is an hydroxyhydropnnxide.

Such hydroperoxides often eliminate H,O, very rapidly and it is not possiMe to determine

their lifetime. A case in point is the h y d r ~ x y h e ~ l h ~ d ~ ~ p a ~ x i d e . ' " On the other hand,

Rule constants andproducl studies .... -- 15h .- -. -----A. --

there are also hydroxyhydropaoxides such as the hydroxym et h y 1 hydroperoxide which

have very long lifetimes in neutral solution and only release H?O1 rapidly at high pH.'"

The structural parameters that determine the stability of h ydroxyhy droperoxides is as

yet not known. We believe that hydroperoxide 10 belongs to the unstable ones and that

the H,O, observed right after ozonolysis is due to reaction (5.25). The resulting product

1 1 will convert to 1 -fomyl-5-hydroxy-5-methylhydantoin 12. This product is

rather stable and releases formic acid only upon treatment with base at high pH.

The yields of immediate H,O, and formic acid released at high pH are both -25%.

It is thus suggested that reaction (5.22) dominates over reaction (5.21) and that

reaction (5.24) is slow compared to reaction (5.25). However, the ozonolysis of

thymidine which will be discussed below takes a very different route, and acetic

acid is released from an intermediate analogous to 10. No acetic acid is formed in the

thymine system that we are concerned here. This requires that in the thymine system there

must be a much faster process competing with potentially possible acetic acid releasing

process. These two systems differ in the substitution at N( 1). The rapid release of H?02 in

the thymine system may thus involve the N(1) proton. We thus suggest that H,O, is

released in the concerted pathway (5.26) which may well proceed with a relay of water

molecules to accommodate a good transition state for the reacdiun to take place. In

thymidine, the rate of acetic acid release is 60 s-' (see below). To compete with such a

process effectively, the rate of reaction (5.26) must be considerably h t c r .

The reduction of the hydroperoxide 6 by the sulfide is depicted in reactions

(5.28)-(5.30), and since l -forrnyl-5-hydroxy-5-methylhy dantoin I 2 hydroly ses

Halt consirtnts und product studie.~. . . . . -. . . . . - - . - -. - - . -. - - .- , - - -. -. .

157

on1 y at high pH [reachon (5.31)], the mnductance observed right after the addition of

ozone disappears upon the addition of the sulfide. After hydrolysis at high pH, the only

remaining product i s 5-hydroxy-5-rnethylhyd.dntoin 1 4 ( 1 OOO!, reference material wa.,

available). The pathway depicted here is, of course, not the only one, but the other

intermediates are also hally converted into this product.

The "slow" release of formic acid with concomitant decay of the hydropero~ide 6

and the other related hydroperoxides, 7 and 8, may be rationali sod by assuming that these

hydropaoxides form hydroxyendopaoxides such as 1 5 [quilibriwn (5.3 2 )I.

Hydroperoxides readily undergo such a reaction with carbony1 compounds. Oftcm, the

equilibrium constants are quite hgh, and the hydrolysis of hydroxyhydroperoxjdes is

observed. Addition of water to 15 leads 16 [reachon (5.3311. It is noted that upon

endoperoxide formation, 7 immediately lead to 16. It is irnpurtant that in the next step no

N-formyl compound is formed. This would not hydrolyse as readily ( t , , = 13 rrlin) as

observed. Instead, we suggest reaction (5.34) which leads to the amide 17. This may then

undergo ring closure forming the hydantoin 18 [reaction (5.3 S)] which subsequently

hydrolysa [reaction (5.3611 and release formic acid and hydroperoxide hydantoin.

Mass spectral evidence for the formation of the hydroperoxide hydantoin t 9

has been given above. Its reduction by sulfide to 5-hydroxy-5-methylhydantoin

14 has also been ascertained well. It has been mentioned above that the decay

kinetics of 6 and the build-up of fbrmic acid are very close to one another and do

not deviate too much firom a first-order rate law (c,f.' inset in Figure 5.14). 'l'his

Rate c o n s t ~ n l ~ ~ ondproducl studies .... -. - --. 158

could be accounted for if the rate determining step in this sequence of reactions is

the formation of the endoperoxide 15 (or 16 fiom 7).

Scheme 5.5

The observation that no singlet dioxygen i s formed in ncutral solution

requires that ring closure of the zwitterion 2 [reaction (5.15)] is much faster than its

deprntonation at N(1). This is not surprising since recombination of oppositc

charges can be very fast (at every encounter) but the deprotonation is comparatively

slow (the rate depending on the pK;, value).

Rate constants and produel sludies.. . . -. 159

5.3.1.7. Ozonolysis at high p H

Thymine deprotonates at high pH CpK, = 9.9), and the rate of' reaction with

ozone becomes considerably faster (k = 3 x 10' dm3 mol-' s-'). Concomitantly, the

reaction mechanism must be partially altered, since now the formation of singlet

dioxygen, o,('A,) is observed in 8% yield.205 Thymine glycol is a conceivable

product at high pH. Its presence is characterised by the GC-MS spectrum after

silylating the dried sample obtained by rotary evaporating the ozonated thymine

solution (Figure 5.16). Four times silylated thymine glycol obtained with MW 448.

The major m/z (%) peaks are 433 (lo), 405(8), 33 1 (75), 147(25), 73(100).

i 0 - N' ."

I TMS

- m.w. 448

Figure 5.16. Mass spectrum of the thymine glycol obtained at pH 11.6 aRm silylation of the ozonated thymine solution

Since the acidity of N(1) and N(3) i s very similar, the anions l a and I b are

both present at comparable concentrations. The higher electron density of la in the

C(5)-C(6) double bond will favour reaction (5.5). The hydrotrioxide anion 20b has

then no positive charge at C ( 6 ) for a rapid ring closure to the Criegce intermediate.

Rate constants and product studies ... . .- -

160

This will enhance the life time of this hydrotnoxide, i.e. i t stands a chance to

decompose into 21 and singlet dioxygen, O?('AJ [reaction (5.6)]. The yield of singlet

dioxygen is a lower limit of this process. Spin conversion and elimination of

ground-state triplet dioxygen, which is energetically favoured by 105 kJ mol-' can

compete. 'This effect is most prominent in the case of the reaction of I- and B r with

ozone.2w In these systems, spin-orbit coupling via the heavy-atom effect favours the

singlet triplet conversion, and singlet dioxygen formation is dramatically reduced.

Also, when a hydrotrioxide anion has a long lifetime and the energetics of singlet

dioxygen elimination disfavour the reaction to proceed ei'fectively, loss of ground-

state triplet dioxygen may proceed with quite some efiiciency.

Scheme 5.6

Rute constclnts undproduet studies ... . . . I h l

Isopynmidines are well-established intermediates in the fiesradical chemistry of

pynmidines.m"2 They react quite readily with water. In the present system, this reaction

leads to the formation of thymine glycol 22. Reactions (5.37) and (5.38) followed by a

reprotonation at N(3) would yield the Criegee intermediate 3. This has been shown above

to yield the strongly UV-absorbing hydroperoxide 616a. This is not formed under present

conditions. Although 6a is not very stable in basic solutions, its formation should not have

escaped the detection. This poses the qumtion as to whether in basic solution the (Jriegee

intermediate 3 is formed at all. If the formation of 20b were not the only prominent but the

only primary process, further reactions of this hydrohioxide would have to be considered.

A detailed product study is still missing.

5.3.2. Thymidine

Ozone reacts with thymidine at a rate constant of 3 x 1 O4 dm7 mol-' s-' and with its

anion at a rate constant of 1 -2 x I o6 dm' mol-' s-' . In contrast to thymine, no singlet

dioxygen is formed at high pH?'' f i e products that have been reported in an carlier

study1* are also markedly different h r n those reported above for the thyrninc reaction. So

it is worthwhile to have a closer look on the mechanistic details.

5.3.2.1. Conductometry and ion chromatography

Similar to thymine, there is a fast and slow build-up o f conductance upon

the addition of omne to the thymidine solution. The ratc of the slow component is four

times faster (k = 0.005 1 s-') than the corresponding process in the m e of thyrninc (Figure

5.17). The fast component in th~s system was resolved by stopped-flow conductance

Huge constants und product studies. ,. . - - 162 . ---

(Figure 5.17, inset). Using ion chromatography, 75% yield of formate was determined.

This yield was increased to 100% by keeping the solution at high pH (base hydrolysis).

L- L

0 3 Time /set

Figure 5.1 7. Ozonolysis of thymidine in aqueous solution. Furmation of acids (proposal as acetic and formic: acid) in the reaction of omne with thymidine as followed by steady state conductance. Inset: Stopped flow conductance analysis of the formation of prompt release of acetic acid

When the reaction between thymidine and ozone has been followed by stop@

flow technique with conductivity detection, a conduc~ance increase was observd, which

was completed within 10 seconds. The rate of h s proton release step (k = 0.55 s-I) is two

orders of magnitude slower than the calculated k,, for the reaction of thymidine with

ozone (at 2 mM thymidine, k = 60 s- ' ) .~ This indicates that protons are released due to

decomposition of primary intermediate. Regarding the nature of acidic product

responsible for the fast conductance increase, k, h r this increase is much lowm than thc

first conductance step for thymine. And most importwtly, such a reaction sequence is

leading to the loss of total peroxide. The yield of the acetic acid determined by ion

chromatogaphy was equal to 18%, which fits well with the percentage o f decoimposed

Rulc L.on.srants nnd prndur,/ studies. .. . - . -- lh3

. - - -

peroxide. The yeld af acetic acid remains same at high pH. 'l'hmefore, we suggest that acctic

acid is responsible fbr the observed condu~;tancc pic~re. Such kind of tramfirmation tor

hy~~xyhydroperoxide was recently reported for the reaction of glyoxylic acid with ~ ~ 0 ~ ~ ' ~

w h ~ h prOcOedS at a rate consmt of the same order (k > 1 s-'1.

Table 5.4. Compilation of yields (with respect to ozone consumed) in the ozonolysis of thyrnidine - -

* measured using ion chromatography

Species

Acetic acid*

Slow acid release (acetic and formic acids, conductoinetry)

Formic acid release'

Total formic acid release at high p ~ *

Total hydroperoxide (immediate) Total hydroperoxide(afier 25 s - 1.5h)

Hydrogen peroxide (immediate, catalase assay)

Hydrogen peroxide (immediate, R,S assay)

Hydrogen peroxide (after 1 h) Organic hydropcroxides (after 1 h)

1 -(2'-deoxyribosy1)-5-hydroxy-5-methylhydtin 35 (after R,S treatment)

1 -(2'-deoxyribosy1)-5-Hydroxy-5-methyl hydantoin 35 (after R,S and OH ' treatment)

Singlet dioxygen (at high pH) A- - . - - -

Simultaneously with acetic acid, carbon dioxide and N-(2'-dmxyribosyl)

fbrmylurea are expect& as by-products. The latter one (N-(2'deoxynbosyl) formylurea)

was detectad by us and determined by Cadet et a1 with 1 8% yield.'" Its f~rmzition as

the product in the ozonolysis o f thyrnidine is indirectly proved by its hydrolysis

(N-(2'-deoxyribosyl) furmylurea undergoes hydrolysis leading to an increase of

Yield I %

I8 -40-45

76

1 00

100

78

8

8

8

70

45

7 5

0 - - . .

Rule constants find product studies.. . . 164

formic acid for about 20%), when ozonated thymidine solution was kept for 2 hours at

pH 11.6. Exactly, at the same experimental conditions it was found that reference

N-formylurea (without sugar moiety) is fully hydro1 ysed giving the stoichiometric

amounts of formic acid. The yields of acidic and peroxidic materials are given in

Table 5.4.

5.3.2.2. Formation and decay of hydroperoxides

To investigate the nature of the peroxide, the Allen's reagent was added to

the ozonated thymidine solution at 2, 10 and 25 s after the ozonation. The yield of

instant total peroxide (at 2 s) was 100% while in the other cases, it was only 80%

(20% loss of total peroxide). Hence there must he 20% very reactive organic

hydroperoxide. The addition of catalase reduces this yield by 8%, 1. e. only 8% H,O,

is formed. The total hydroperoxide yield remains practically stable over time.

However, no further increase of H,O, was observed.

Since, catalase datroys not only H,O,, but some reactive organic peroxides as

well, experiments with non-activated Allen's reagent was carried out to analyse the nature

of the peroxides in the ozonated thymidine solution. Under these conditions, reactive

organic peroxides react with non-activated 1- as fast as with Mo-activated iodide

(k = 100 dm3 mol-' s-' as measured by stopped tlow technique), but H,O, reacts much

slower (2.1 dm7 mol-' s-' for activated iodide and 1 .# x 1 0"dm3 mol-' s-' for non-

activated iodide, r e ~ ~ e c t i v e l ~ ) . ~ " ~ SO, the formation of 1, in this case i s biphasic:

fast initial step (could not be resolved by conventional UV-spectroscopy) and the

Rate cunstants and product studies .... 165

slow one, which is nicely resolved. The kinetics of slow component is matching

with experimentally determined rate constant for H,O, reaction with non-activated

Allen's reagent.

To investigate the nature of long-lived intemediates that can account for

more stable organic peroxides formed in 70% yield, the reaction W e e n thymidine

and ozone has been followed by conventional UV-spectroscopy. As can be seen from the

Figure 5.18, 30 set afia the ozonation, an intermediate is f o m d with absorption

maximum at 2110 run with a small absorption at 270 nm. Within 12 min, this

intermediate has been totally disappeared and a product with h,, = 227 run is formed.

Figure 5.18, OzonoIysis of thymidine at natural pH. UV-absorption spectra obtained at 0.5 (m), 1.5 (A) and 7.5 min (A) after omolysis. Inset: Decay of 240 species (A) and build up of the 220 species (a)

The kinetics of the intermediates at different wavelengths has been analysed using

conventional UV spectrophotometer. The decay of the spedcs at 240 nm and the

formation at 220 nrn resulted an identical rate constant of 0.0062 s". These rate constants

art: close to the rate constant of the slow increasc of conductance (0.005 1 s-I j.

K t r ~ r constunts undprodurt studies. .. . . --- - . - -- --- -.

Ihh -

5.3.3.3. HPLC and LC-MS analysis

In contrast to thymine, the HPLC chromatogram of thy~nidine contains a nuinbcr

of product peaks. The retention times of the product peaks are 2.9, 1 1 37 , 12.83, 14.93 and

24.96 min. All of these species have h,, around 227 nrn except in the case of the first

eluting peak, which absorb at 256 and a tailing to~vards 220. All of these initial peks were

disappeared by the duction with bis(2-hydroxyethy1)sulfide. Upon post column

derivatisation with molybdate activated iodide, five hydroperoxides were detected. The

hydantoins were eluted at 12.74 and 14.25 min with A,, i 220. Because of its low yield

and the presence of peroxide peaks at close prox~mity, it is not well separated in the

immediate injection. Except the first unstable peroxide, other peroxides are stable with

time. Thymidine was eluted at 60 min at this condition of the reversed phase column with

water as eluent. The larger number of products as compared to thymine is largely due to

the fact that now rneso/d,l isomers are formed.

Figure 5.1 9. ES-MS spectrum (positive mode) of the separated I 42'- deox~~ribosyl)-5-hydroxy-5-methy I hydantoin 35 Iiaction obtained after the ozonation of thymidine

257 I I I I . tP I l l n r ' " ' = ' ~ ' ' ' ' ' ' ~ ~ ' ~ ~ ~ ' ~ ~ ' ~ ~ ~ ~ r ~ ~ ~ ~ ~ c l ~ ~ ~ ~ ~ ~ ~

a o o 253 300 3 50 uao 451: son 5 5 0 609 1; 5 0

miz

A few mg of hydantoins were isolated by piAeparative HPLC as reference

material. The material was confirmed by ES-MS (Figtrre 5.19). Afta rotary evaporation,

thts material was remained oily and HPLC chromatogram still showed some trace amount

o f impurities. Using h s material for calibration, the total hydantoin yeld was determined

to he 45%. 'The yield of hydantoins was increased to 75% by base hydrolysis.

Considering that the reference material contained some trace mounts of impurities, we

have also calculated their yield based on the assumption that they have the same molar

absorption coefficient like their parent, 5-hydroxy-5-methylhydantoin for which retiable

material was available and obtained a close value. In any case, their yields wert much

higher than the reported value of only 1 9%. In the latter study, the hydropaoxides

were subjected without prior reduction to a laborious work-up, and possibly only a

fraction underwent H,O, elimination, while the rest was degraded.

Upon LC-ES-MS analysis, four products were detected in the experimental

condition. The ozonation was carried out at 3". The first peak gave a molecular

weight 176 Da with m/z 1 77 (M+1) and 353 (2Mt- 1). The next three peaks gave the

molecular weight 1 6 1,262 and 290 Da with m/z 1 62,263 and 29 1 (M+ 1 ) respectively in

positive mode of electrospray ionisation (Figure 5.20). The pmks with molecular weights

262 and 290 are considered to be peroxide peaks. For these experimcnts, the

conversion of thymidine had to be very high, and it as yet not ascertained whether

the first two decomposition products (M W 16 1 and 1 76) are really primary ones.

'I'he peroxide 26 is considered to be the precurser of 290 pcak with MW 308 (see

section 5.3.3 -4). After the treatment with sulfidc the two hydroperoxide peaks werc

Kart, r.on,stun/e and product stud~es ... - - - - -. . - - - - - -- IhX

-. -

disappeared. The sulfoxide (138 Da) that is formed in [his reaction was also

detected. So it is assigned that product with M W 262 i s 34 and product with M W

1 76 is 29 (see Scheme 5 .8 ) . At present, other products are not fully assigned.

Figure 5.20, LC-ES-MS spectra of (a) 1 -(2'-deoxyribosyl)-5-hydroperoxy-5- methyl hydantoin 34 and (b) N-(2'-deoxyribosyl) urea 29 obtained by the ozonolysis of thymidine

5.3.3.4. Mechanistic aspects

Similar to thymine, the mechanism of thymidine ozonation is bascd on the

release of acidic materials, acetic and formic acid and the reduction of peroxides

with sulfide. Thc initial fast increase of conductance is because uf the I-elease of

acetic acid. The initial formation of acetic acid can be represented as in Scheme 5.7.

We propuse that initially formed hydroxyhydroperoxide 25 is undergoing C-C: bond

cleavage and release acetic acid [reaction (5.42)] to form 27.

27 Scheme 5.7

Hydrolysis of the urea derivative 28 can bc considered as the very slow

release of formic acid af ia base hydrolysis. ' lhs initially t'ormed hydroxy

hydroperoxide 25 can eliminate formic acid by the hydrolysis of formyl group

attached to the N(1) position and cyclisation of 33 leads to the formation of'

N-(2'-deoxyribosyl)-5-hydroperoxy-5-hydntin 34. The R,S reduction of 34 yield

N-(2'-deoxynbosyl)-5-hydroxy-5-methylhydntoin 35. The slow release of formic

acid can be schematised as follows.

Rate constants and product studies.. -- . . .------ 170

-%Q R2Sl- R2S0 OOH _I__)

15.49)

1 0 0

Scheme 5.8

The hydropemxide 26 would have a molecular weight of 308 Da, but the highest

observed molecular weight in LC-MS analysis is 290 Da, i.e the hydroperoxide either has

lost one molecule of water upon detection by LCMS or effectively contains one water

less. The latter could be accounted for if reactions (5.51) and (5.52) would have taken

place prior to the LC-MS analysis (MW(37) = 290 Da). A hydroperoxide that has a

noticeable absorption at h > 260 m has been detected (see below), and this could

be due to the extended chromophore in 37.

Rate conslants and u ~ ~ d u c t studies .... 17 1

Scheme 5.9

5.4. Summary and conclusions

In this chapter, a detailed study covering the determination of the rate

constants of the reactions of ozone with nucleobases and their related compounds

and the pH dependence on the rate constant have been carried out. A detailed

investigation of the products and intermediate species formed by the reaction of

ozone with thymine and thymidine is alm presented. The mechanism and product

analysis of the ozonolysis of thymine and thymidine have been undertaken as

model systems for pyrimidine nucleobases.

The pH dependent rate constants of the reaction of ozone with DNA, i t s

constituents and related compounds have been determined using competition kinetics

(nitrite andlor butcn-3-01} and the indigo bleaching method. The rate constants vary

drastically with the degree of protonation of the nucleobase derivatives (in units of

dm' mol" s-'), c.g. in the case of cytosine, k = 1 8 (protonated), k - 1 -4 x 1 o3 (neutral) and

k = 1.5 x 1 Oh (deprotonatd). All the of selected nucleubases have shown similar results.

Upon deprotonation, the mechanism of the ozone reaction may also change; e.g. no

singlet dioxygen ( 0 , ' ~ ~ ) is formed in its reaction with 5-chlorouracil, but when the

Kate consrunt.7 rrnd product studre.5 . 172 -- -.- - - .-- ---- -

5-chlorouracilate ion predominates it becomes a major product (-42%). Rate

constants f i r the neutral compounds are: thymine (4.2 x 1 07, thy rnidinu (3 -0 x 1 04),

1,3-dimethyluracil (2.8 x 10'1, uracil (6501, 6-mcthyluraci l ( 1 40), 5-chlorouracil

(4.3 x I#), orotic acid (5.9 x 10'), isoorotic acid (3.7 x 1 O'), 2'-deoxycytidine

(3 -5 x 1 0->), cytidine (3.5 x 1 O'), adenine (1 2), 2 '-deoxyadenosine ( 1 4), adenosine

(1 6), guanosine (1.6 x 1 04), 2'-deoxyguanosine (1.9 x 1 0 ~ ) and DNA (4 10). In the

case of adenine and its derivatives, and thus also in the case of DNA, 'OH is

produced (via 02*- as an intermediate). For the determination of their intrinsic

ozone rate constants, tert-butylalcohol was hence added as 'OH scavenger.

The yield of instant total peroxide in thymine and thymidine is 100% with

respect to the consumed ozone. In the case of thymine, a very little decay of the

total peroxide is observed (-7% loss), while in the case of thymidine, a 20% loss of

peroxide is observed very rapidly. In thymine, a 25% H,O, is observed by catalase and

R,S assay and it was increased up to 40% with time with a half life -1 2min. The total

organic peroxide has been decreased from 75% to 53 % with the same rate of increase of

H,O,. It is tbund that a highiy UV absorbing acidic peroxide (pKa = 4.0) is released

promptly with v a y high rate (3.4x104 dm3molL1s-'1. The subsequent hydrolysis of the

above paoxide releases formic acid along with other formic acid rcleaqing precursors. The

two organic peroxide identified are 1 -hydmperoxymethyl en-3-(2-oxo-propanoy 1 f-urm and

5-hybpxy-5-methyIhydantoin. After the reduction with bis(2-hydroxyethyl)suIfidc,

67% 5-hydroxy -5-methyhydantoin was ohserverl and after base treatment its yield

H L ~ ~ P cuns!unn and product studies .. . - --A .

17.3 -- -

went up to 100%. The very slow release of formic acid is considered to be fi-oin the

hydrolysis of N(1)-fomyl-5-hydroxy-5methylhydantoin. At high pH, thymine

glycol and singlet oxygen are formed during ozonation.

In the case of thymidine, only a very little H,O, was observed (-8%) ,and

there was no change in the concentration of H,O, with time. The fast conductance

increase is due to the release of acetic acid (18% was measured using ion

chromatography). HPIC measurements showed a 75 % formic acid yield which has

been increased to 100% after base hydrolysis. TotaI yields of the two isomers of

1-(2'-deoxyribosyl)-5-rnethyl-5-hydroxy hydantoin were observed as 45% after

reduction with R,S, while the yield was increased upto 75% after base hydrolysis.

A tentative mechanism to satisfy ail these features is proposed.

In brief, the reactions of ozone with nucleobases proceeds via a cyclic

ozonide which cleaved into peroxyl and carbonyl compounds. The formation of

peroxidic materials gives much impact in biological systems due to the presence of

trace amounts of transition metal ions in physiological condition which can produce

highly reactive radicals. The formation of acidic materials in the present systems clearly

demonstrate that the reaction proceeds via the release of acids and H,O, against the simple

H,02 elimination of ozonide as proposed in some earlier reprts.'"~'"S