The photochemistry of thymidylyl-(3′-5′)-5-methyl-2′- deoxycytidine in aqueous solution

The Photochemistry of Thymidylyl-(3′-5′)-5-methyl-2′-deoxycytidine in Aqueous

Solution

Author(s): Lech Celewicz, Moriz Mayer, Martin D. Shetlar

Source: Photochemistry and Photobiology, 81(2):404-418.

Published By: American Society for Photobiology

DOI: http://dx.doi.org/10.1562/2004-06-15-RA-201.1

URL: http://www.bioone.org/doi/full/10.1562/2004-06-15-RA-201.1

BioOne (www.bioone.org) is a nonprofit, online aggregation of core research in the biological, ecological, and

environmental sciences. BioOne provides a sustainable online platform for over 170 journals and books published

by nonprofit societies, associations, museums, institutions, and presses.

Your use of this PDF, the BioOne Web site, and all posted and associated content indicates your acceptance of

BioOne’s Terms of Use, available at www.bioone.org/page/terms_of_use.

Usage of BioOne content is strictly limited to personal, educational, and non-commercial use. Commercial inquiries

or rights and permissions requests should be directed to the individual publisher as copyright holder.

BioOne sees sustainable scholarly publishing as an inherently collaborative enterprise connecting authors, nonprofit publishers, academic institutions, research

libraries, and research funders in the common goal of maximizing access to critical research.

Photochemistry and Photobiology, 2005, 81: 404–418

The Photochemistry of Thymidylyl-(39-59)-5-methyl-29-deoxycytidine in

Aqueous Solution^{

Lech Celewicz^1,2, Moriz Mayer¹and Martin D. Shetlar*¹

¹Department of Pharmaceutical Chemistry, School of Pharmacy, University of California, San Francisco, CA and

²Faculty of Chemistry, Adam Mickiewicz University, Poznan, Poland

Received 15 June 2004; accepted 8 October 2004

about 5% of the total Cyt residues in calf thymus DNA (see table

ABSTRACT

1.2, pp. 6–7 in Ref. 1); a similar level of methylation of Cyt is

The photochemistry of the dinucleoside monophosphate thymi-

dylyl-(39-59)-5-methyl-29-deoxycytidine (Tpm⁵dC)hasbeenstud-

ied in aqueous solution using both 254 nm and UV-B radiation.

A variety of dinucleotide photoproducts containing 5-methyl-

cytosine (m⁵C) have been isolated and characterized. These

include two cyclobutane dimers (CBD) (the cis-syn [c,s]and trans-

syn forms), a (6-4) adduct and its related Dewar isomer, and

two isomers of a product in which the m⁵C moiety was converted

into an acrylamidine. Small amounts of thymidylyl-(39-59)-

thymidine (TpT) were also formed, presumably as a secondary

photoreaction product. In addition, a photoproduct was charac-

terized in which the m⁵C moiety was lost, thus generating 39-

thymidylic acid esterified with 29-deoxyribose at the 5-hydroxyl

on the sugar moiety. The c,s CBD of Tpm⁵dC readily undergoes

deamination to form the corresponding CBD of TpT. The

kinetics of this deamination process has been studied; the cor-

responding enthalpy and entropy of activation for the reaction

have been evaluated at pH 7.4 as being, respectively, 73.4 kJ/mol

and 2103.5 J/K mol. Deamination was not observed for the other

characterized photoproducts of Tpm⁵dC.

observed in human DNA as well (2). In plant DNA, the level of

methylation of Cyt is considerably higher. For example, about 31%

of the Cyt are in the form of m⁵C in wheat DNA (1).

The presence of m⁵C in DNA has potential for increasing the

amount of light absorbed by that DNA in the UV-B region of the

spectrum. This nucleobase has an absorption spectral proﬁle that is

shifted toward longer wavelengths, as compared with Cyt itself. For

example, the absorption spectrum of 5-methyl-29-deoxycytidine

(m⁵dC) displays an absorption maximum at 278 nm, as compared

with 271 nm for 29-deoxycytidine (3). This implies that m⁵C

residues are more likely to capture light in the UV-B region of the

spectrum than Cyt residues. This, in turn, indicates that substituting

Cyt in DNA with m⁵C will increase the UV-B absorbance of the

DNA, as compared with DNA not containing m⁵C.

In the past few years, it has become clear that m⁵C residues

within DNA play an important biological role (for an overview, see

Ref. 4). Methylated Cyt appear to be involved in the regulation of

chromatin structure and gene expression (5), whereas changes in

normal patterns of Cyt methylation seem to be involved in

tumorigenesis (Ref. 6 and references therein), ageing (Refs. 7,8 and

references therein) and a variety of human syndromes involving

abnormal gene expression (9). On the photobiological level, it

appears that m⁵C photoproducts, in particular cis-syn (c,s)

cyclobutane dimers (CBD), are likely to be involved in mutagenic

events. A variety of studies have demonstrated that mutagenic hot

spots in UV-B–irradiated DNA are associated with 59-pyrimidine-

m⁵C-guanine sequences (see for example Refs. 10–13 and

references therein). Such sequences in native DNA should be

susceptible to formation of c,s CBD. Deamination of, say, the c,s

thymine (Thy)–m⁵C CBD would yield the corresponding c,s Thy-

Thy CBD, which could then be ﬁxed as a mutated site.

The photochemistry of m⁵C itself has become increasingly well

explored during the past 15 years. This compound and its 29-

deoxyribonucleoside undergo light-induced rearrangement reac-

tions (through putative Dewar intermediates) in acetonitrile to form

ureidoacrylonitriles (14). The m⁵C moiety reacts with alkyl amines

to form ring-opened products, namely akylcarbamoyl-3-amino-

acrylamidines (15) and with alcohols to form carboalkoxy-3-

aminoacrylamidines (16). Both m⁵C and its deoxyribonucleoside

react with water to form acrylamidines (17). Other studies of the

reactions of m⁵C in water have suggested that photodeamination,

photooxidation and photo-induced demethylation occur to form

Thy, 5-hydroxymethylcytosine and Cyt, respectively (18).

INTRODUCTION

The photochemistry of 5-methylcytosine (m⁵C) has been a topic of

a signiﬁcant amount of interest during the last 15 years. This

particular nucleobase is a signiﬁcant constituent of mammalian and

plant DNA. For example, methylated cytosines (Cyt) account for

{Posted on the website on 19 October 2004

*To whom correspondence should be addressed: Department of Pharma-

ceutical Chemistry, School of Pharmacy, S-926, University of California,

San Francisco, CA 94143-0446, USA. Fax: 415-476-0668; e-mail:

shetlar@cgl.ucsf.edu

Abbreviations: CBD, cyclobutane dimer; c,s, cis-syn; Cyt, cytosine; DAD,

diode array detector; dCpdC, 29-deoxycytidylyl-(39-59)-29-deoxycyti-

dine; dCpT, 29-deoxycytidylyl-(39-59)-thymidine; ESI, electrospray

ionization; HPLC, high-performance liquid chromatography; MALDI,

matrix-assisted laser desorption ionization; m⁵C, 5-methylcytosine;

m⁵dC, 5-methyl-29-deoxycytidine; m⁵dCpdC, 5-methyl-29-deoxycytidyl-

yl-(39-59)-29-deoxycytidine; m⁵dCpT, 5-methyl-29-deoxycytidylyl-(39-

59)-thymidine; MeOH, methanol; NMR, nuclear magnetic resonance;

Thd, thymidine; Thy, thymine; TpdA, thymidylyl-(39-59)-29-deoxyade-

nosine; TpdC, thymidylyl-(39-59)-cytosine; Tpm⁵dC, thymidylyl-(39-59)-

5-methyl-29-deoxycytidine; TpT, thymidylyl-(39-59)-thymidine; t,s,

trans-syn; TSP, 3-(trimethylsilyl)propionic-2,2,3,3-d₄acid, sodium salt.

Ó 2005 American Society for Photobiology 0031-8655/05

404

Photochemistry and Photobiology, 2005, 81 405

The photochemical reactions of m⁵C to form CBD, (6-4) adducts

and Dewar adducts have also received some attention. Early

studies suggested that there was no evidence for formation of CBD

when m⁵dC was irradiated at 254 nm in the frozen state (19). A

second study from the same laboratory investigated the photo-

chemistry of m⁵C at 254 nm when incorporated into a viral DNA in

which all Cyt residues were replaced with m⁵C (20). Once again,

no evidence was found for formation of CBD. However, a later

study suggested that a CBD containing m⁵C, as well as

a monomeric m⁵C-containing product, were formed when de

novo methylated DNA was irradiated at either 254 or 302 nm (21);

however, these products were not structurally characterized. The

photoreactions of m⁵C in the context of the duplex poly-29-

deoxyguanosine:poly-29-deoxy-5-methylcytidine have also been

studied. Applying radioimmunoassay methods, it was found that

CBD and (6-4) type adducts of m⁵C were both present in this

duplex of homopolymers after it was irradiated with either UV-B

or UV-C light (22). Recently, it has been shown that m⁵C, when

irradiated in the frozen state with light predominately in the UV-B

region, formed three m⁵C cyclobutane homodimers (23) that could

be structurally characterized. Similar irradiation of equimolar

mixtures of m⁵C and Thy produced three m⁵C-Thy cyclobutane

heterodimers (23), including the biologically relevant c,s and trans-

syn (t,s) dimers.

The photoreactions of dinucleoside monophosphates have

received considerable attention during the past 40 years. One

reason is that the reactions involving the nucleobase components

contained in these dinucleotides (or in oligonucleotides as well) are

considered to be more reﬂective of those occurring in DNA or

RNA than those reactions occurring when the same moieties are

irradiated as free nucleobases, nucleosides or nucleotides in ﬂuid

solution. For example, the Dewar adduct of Thy was discovered

only when the structure of the last uncharacterized photoproduct

formed in the thymidylyl-(39-59)-thymidine (TpT) system was

ﬁnally determined (24,25). The Dewar adduct of Thy in the TpT

system is not a primary photoproduct, but is a secondary product

produced from the corresponding (6-4) adduct by absorption of

UV-B light (24). Isolation of the Dewar product of either the

nucleobase Thy or the nucleoside thymidine (Thd), after the

irradiation of the (6-4) adducts of these compounds, has not been

reported. The presence of structural constraints, found in the

context of the dinucleotide environment, may be conducive to

light-induced formation of the Dewar adduct of Thy from its

precursor (6-4) adduct in TpT.

Cyt are among the products arising from these reactions (Refs. 29–

32 and references therein). Mass spectral methods have been

developed recently for detection and quantitation of the four c,s

CBD, the four (6-4) adducts and the four Dewar adducts of TpT,

dCpdT, TpdC and dCpdC in photoreacted DNA (Ref. 33 and

references therein). This method involves digestion of the

irradiated DNA with an enzymatic cocktail that reduces it to

dinucleoside monophosphates containing these photoproducts and

nucleosides of unreacted bases. (The literature concerning di-

nucleotide photochemistry up to 1989 is surveyed in Ref. 34,

whereas a more recent review is given in Ref. 35; leading

references to more recent work are given in Refs. 33,36.)

Of particular interest here is the photochemistry of dinucleoside

monophosphates containing m⁵C. Only one study of the photo-

chemistry of such compounds has been published, namely that of

Douki and Cadet (29); these workers described their work on three

compounds, namely 5-methyl-29-deoxycytidylyl-(39-59)-thymi-

dine (m⁵dCpT), 5-methyl-29-deoxycytidylyl-(39-59)-29-deoxycyti-

dine(m⁵dCpdC), andthymidylyl-(39-59)-5-methyl-29-deoxycytidine

(Tpm⁵dC). Of these three compounds, the photochemistry of

m⁵dCpT was the most amenable to study. Two compounds that

contained m⁵C were isolated and characterized by proton nuclear

magnetic resonance (NMR) and UV spectroscopy and by mass

spectrometry; one was identiﬁed as the t,s m⁵C-Thy CBD and the

other as the m⁵C-Thy (6-4) adduct. Mass spectral and photochemical

evidence indicated that the c,s m⁵C-Thy CBD and the m⁵C-Thy

Dewar adduct were present in a fraction that could not be separated

into pure components. Additional evidence, supporting the assign-

ment of the (6-4) adduct of m⁵dCpT as a compound containing an

undeaminated m⁵C residue, was provided by a study that showed that

the ﬂuorescence properties of this compound were different from

those of the (6-4) adduct of TpT (37). In addition to the products

containing m⁵C, two products were formed that were also produced

when TpT was irradiated; these were the c,s Thy-Thy CBD and the

(6-4) adduct of TpT. Douki and Cadet (29) suggested that these

particular products may have arisen from photoreactions of excited

m⁵CpdT, presumably via deamination of the m⁵C component of the

dinucleotide in its excited state.

In the case of the Tpm⁵dC and m⁵dCpdC systems, m⁵C-con-

taining photoproducts were not isolated (29), although the authors

indicated that they could not rule out that certain products eluting

with very small high-performance liquid chromatography (HPLC)

peak areas could be such compounds. However, it was found that the

predominant photoreactions of Tpm⁵dC led to the same c,s CBD and

(6-4) adduct that were formed in the UV-irradiated TpT system.

Irradiation of the m⁵dCpdC system produced two products that were

characterized; these were the (6-4) product and the corresponding

Dewar adduct of TpdC (29).

A signiﬁcant body of work exists on the photochemistry of di-

nucleoside monophosphates. Most experimental attention has been

directed toward those dinucleotides containing two pyrimidines,

although studies on the photochemistry of thymidylyl-(39-59)-

29-deoxyadenosine (TpdA) and 2-deoxyadenosylyl-(39-59)-29-

deoxyadenosine have been fruitful. The remaining dinucleoside

monophosphates containing purine nucleobases appear to be un-

reactive toward formation of speciﬁc photoproducts (26) (for

a review of the photochemistry of the purine containing dinucleo-

side phosphates, see Ref. 27; more recent information on the

structure of the TpdA photoproduct is given in Ref. 28). Along

with studies of the photochemistry of TpT, the photochemical

reactions of three dinucleoside monophosphates of particular

interest in DNA have been investigated, namely 29-deoxycyti-

dylyl-(39-59)-29-deoxycytidine (dCpdC), thymidylyl-(39-59)-29-de-

oxycytidine (TpdC), and 29-deoxycytidylyl-(39-59)-thymidine

(dCpT). CBD, (6-4) products, Dewar adducts and hydrates of

In this article, we present the results of a study of the photo-

chemistry of Tpm⁵dC in aqueous solution. This dinucleoside mono-

phosphate is one of the compounds whose photoreactions were

studied previously by Douki and Cadet (29), but for which these

workers were unsuccessful in isolating m⁵C-containing products

for characterization. The study of the photoreactivity of Tpm⁵dC

has proven more tractable under the conditions used in our ex-

periments, and therefore, we have been able to isolate and charac-

terize

a

number of photoproducts containing m⁵C. This

dinucleotide sequence is of particular photobiological interest

because it is one that occurs naturally in mammalian DNA in the

trinucleotide sequence thymidylyl-(59-39)-5-methyl-29-deoxycyti-

dylyl-(59-39)-29-deoxyguanosine.

406 Lech Celewicz et al.

ꢀ208C until required for use. The resultant material was characterized by

mass spectrometry and by proton and ¹³C NMR spectroscopy. Greater

experimental detail concerning this synthesis of Tpm⁵dC and its spectro-

scopic characterization, as well as similar information for m⁵dCpT, will be

reported elsewhere (L. Celewicz, in preparation).

MATERIALS AND METHODS

General aspects. HPLC solvents were from Fisher (Fair Lawn, NJ); NMR

solvents were from Aldrich (Milwaukee, WI). The dinucleotides TpT and

TpdA were purchased from Sigma (St. Louis, MO). A Rainin binary

gradient pumping system (Emeryville, CA) coupled to a Hewlett-Packard

1040A diode array detector (DAD) (Palo Alto, CA) was used for HPLC.

Analytical and preparative separations were carried out on a Shiseido

Capcell UG120 RP18 4.5 3 150 mm reverse phase column (5 lm particle

size; Yokohama, Japan); hereafter, this column is termed the Capcell

column. Before injection of HPLC samples, they were subjected to spin

ﬁltration, using Costar Spin-X micro-centrifuge ﬁlter tubes containing a 0.2

lm nylon ﬁlter (Corning Incorporated, Corning, NY).

RESULTS AND DISCUSSION

Results of exploratory studies on the effects of irradiation

of the Tpm⁵dC system

UV spectra were run on a Hewlett-Packard 8452A diode array spectro-

meter. One-dimensional NMR spectra were run either at 600 MHz on a

Varian INOVA NMR spectrometer (Palo Alto, CA) or at 500 MHz on a

Bruker DRX 500 MHz NMR spectrometer (Rheinstetten, Germany). Most

assignments of resonances corresponding to particular protons were made

by comparison with those seen in the NMR spectra of similar well-studied

compounds.

Preliminary studies were done to optimize irradiation times and

separation conditions and to explore the general nature of the

photoproducts produced when Tpm⁵dC (I) is irradiated. In a typical

run, the content of one tube of lyophilized Tpm⁵dC (as the

ammonium salt; see above) was dissolved in distilled water to

make a stock solution. A portion of this stock solution was then

diluted 300-fold and the UV spectrum measured. Using an

estimated e_maxvalue of 18 150 M^ꢀ1cm^ꢀ1at 271 nm (obtained by

adding the e_maxvalues for Thd and m⁵dC [9650 M^ꢀ1cm^ꢀ1Ref. 41

and 8500 M^ꢀ1cm^ꢀ1Ref. 42, respectively]), the concentration of the

diluted solution was obtained. From this value and the dilution

factor, the concentration of the stock solution was calculated. A

portion of the stock solution was then diluted with doubly distilled

water to 0.5 mM and the remainder of the stock was frozen until

needed for further use.

A sample (0.7 mL) of the 0.5 mM solution of Tpm⁵dC was

placed in a Type 30 cuvette and irradiated at about 48C with 254

nm light for varying lengths of time (see above); the pH of the

solution at this temperature was about 4.6. After each irradiation

time, 20 lL of the contents of the cuvette was removed and

injected on the Capcell column (see above). The following gradient

(Gradient A) was used: t 5 0 min: % MeOH 5 0; 4 min, 0; 6 min,

25; 7 min, 25; 7.25, 0; 10 min, 0. The ﬂow rate was 3 mL/min and

the aqueous eluent was 25 mM ammonium formate, pH 7. After

128 min (a dose of 123 kJ m^ꢀ2) about 26% of the parent compound

had reacted, whereas after 300 min (288 kJ m^ꢀ2), about 50% had

disappeared. After 16 h irradiation under the same conditions, most

of the parent compound had disappeared; however, the HPLC

chromatogram indicated that the set of reaction products was very

complex after this extended period of irradiation. A second set of

exploratory experiments was done in which a similar Tpm⁵dC

sample was irradiated at pH 4.6 with light centered at 312 nm from

unﬁltered BLE 1T158 lamps. After 128 min, about 20% of the

parent had disappeared. To get an idea of how far the reaction

could be pushed with this longer wavelength radiation, we

irradiated for 15 h and found that around 48% of the Tpm⁵dC

had reacted.

Matrix-assisted laser desorption ionization (MALDI) mass spectrometry

was carried out on an Applied Biosystems 4700 mass spectrometer

(Applied Biosystems, Foster City, CA) operating in the reﬂector mode and

using external calibration. The matrix used was a-cyano-4-hydroxycin-

namic acid. Electrospray ionization (ESI) mass spectrometry was done on

either a Sciex API300 triple quadrupole electrospray instrument (Toronto,

Canada) or on a Waters Micromass ZQ 4000 instrument (Beverly, MA).

Irradiation methods. Exploratory runs were done in stoppered black self-

masking quartz UV cuvettes holding a volume of 0.7 mL (Type 30,

Precision Cells, Farmingdale, NY); the sample compartment in these cells is

2 mm wide and 1.0 cm deep. Preparative irradiations were done on 3.5 mL

volumes contained in stoppered quartz spectrophotometer cuvettes (Type

110, Hellma Cells, Plainview, NY), which are constructed to hold this total

volume; the sample compartment in these cells is 1.0 cm deep. Most

irradiations for preparative purposes were done with 254 nm light provided

by unﬁltered Spectronics BLE 1T155 lamps housed in a Spectroline

XX-15A dual lamp holder (Spectronics, Westbury, NY). Irradiations for

preparative purposes were done in a cold room maintained at 48C. A cuvette

containing the sample to be irradiated was placed on its side on an

aluminum block (9 cm long, 7.5 cm wide and 5 cm high) such that it was

situated directly underneath and parallel to one of the lamp tubes; the lamp

housing itself was supported on blocks 7.5 cm high. The distance between

the bottom of the cylindrical lamp tube and the top surface of the cuvette

was 2.3 cm. The irradiance at the surface of the cuvette was about 16 J m^ꢀ2

s

^ꢀ1, as measured using a Spectronics DM-254N Ultraviolet Meter. Some

irradiations were done with light centered at 312 nm; in these situations, the

light was provided by Spectronics BLE 1T158 lamps housed in the same

holder and the sample was situated as for the 254 nm irradiations.

Preparation of Tpm⁵dC. The synthesis of Tpm⁵dC was performed using

the phosphotriester approach (38,39). 59-O-Dimethoxytritylthymidine was

phosphorylated with 2-chlorophenylphosphorodi-(1,2,4-triazolide) to give,

after hydrolysis, the triethylammonium salt of the 39-(2-chlorophenyl)-

phosphate of 59-O-dimethoxytritylthymidine. The latter compound was

allowed to react with 4-N-benzoyl-5-methyl-29-deoxycytidine in the

presence of 1-(mesitylenesulphonyl)-3-nitro-1,2,4-triazole in pyridine as

a condensing agent. The product of the reaction, protected Tpm⁵dC, was

puriﬁed by silica gel column chromatography (silica gel 70–230 mesh 60 A,

Aldrich) using chloroform–methanol (MeOH) (50:1 vol/vol) as eluent

(yield 75%, 230 mg). Deblocking of the protected Tpm⁵dC was

accomplished via a two-step procedure (40). In the ﬁrst step, the acid

labile 59-O-dimethoxytrityl group was removed by treatment of the

A typical HPLC chromatogram, resulting from injection of 200

lL of freshly irradiated Tpm⁵dC irradiated for 128 min at 254 nm

at pH 4.6 and 48C with an unﬁltered 1T155 lamp is shown in Panel

b of Fig. 1. We found nine major product peaks to be of interest for

this study; these are designated 1–9 in Fig. 1b. Four peaks,

corresponding to minor products, were present. Three of these,

labeled with small letters (v, w and x) were not studied in detail;

however, one minor product peak was identiﬁed, namely that

labeled TpT (6-4), which elutes at about 4.6 min and corresponds

to the (6-4) photoadduct of TPT. This identiﬁcation was based on

the identity of its elution time with that of an authentic sample of

this compound and the close correspondence of its absorption

spectrum proﬁle, as obtained with the spectral capture capability of

protected Tpm⁵dC with

a 2% solution of benzenesulfonic acid in

a chloroform–MeOH solution (7:3 vol/vol) at 08C for 10 min. The partially

protected Tpm⁵dC was then puriﬁed by silica gel column chromatography

(silica gel 70–230 mesh 60 A, Aldrich) using chloroform–MeOH (25:1 vol/

vol) as eluent. The base labile groups 4-N-benzoyl and 2-chlorophenyl were

removed by treatment with concentrated ammonia at 508C for 5 h. Finally,

the unprotected Tpm⁵dC was puriﬁed by HPLC on a Shiseido Capcell Pak

C₁₈UG 120 reverse phase column (103250 mm, 5 lm particle size), using

25 mM ammonium formate–MeOH (84/16) ﬂowing at a rate of 5 mL/min.

The appropriate fraction containing pure Tpm⁵dC from each HPLC run was

collected. The collected eluent was then distributed into a number of 50 mL

plastic tubes and lyophilized to remove excess ammonium formate. The

tubes containing the resultant freeze-dried material were kept in a freezer at

Photochemistry and Photobiology, 2005, 81 407

Figure 1. Panel a displays the HPLC

chromatogram of a freshly irradiated

reaction mixture of Tpm⁵dC photoprod-

ucts, obtained via irradiation of 0.5 mM

Tpm⁵dC for 128 min with an unﬁltered

Spectroline BLE-1T158 lamp (UV-B out-

put centered at 312 nm). Panel b shows the

HPLC chromatogram of a second freshly

photolyzed Tpm⁵dC sample (0.5 mM) that

had been irradiated for 128 min with

a Spectroline BLE-1T155 lamp with out-

put mainly at 254 nm. Both irradiated

solutions were at pH 4.6 and at 48C. The

injection volume in each case was 200 lL

and the wavelength of detection was 230

nm. The irradiation conditions, the col-

umn used and the HPLC gradient used are

as described in Materials and Methods.

the DAD, with that of the authentic TpT (6-4) adduct. The peaks

labeled y and z are minor impurities in the parent compound. Panel

a in Fig. 1 shows the HPLC chromatogram obtained after Tpm⁵dC

at pH 4.6 and 48C was similarly irradiated with light centered at

312 nm. The biggest difference between the chromatograms

displayed in Fig. 1a,b is that Peak 5 eluting at 5.2 min in Panel 1b

is barely detectable in Panel 1a. We did not ﬁnd any evidence for

formation of the TpT (6-4) product in this irradiated solution.

The UV spectra for the compounds corresponding to a number

of these peaks, obtained using the ‘‘on the ﬂy’’ spectral capture

capabilities of the DAD, are shown in Fig. 2. These spectra will be

discussed in the course of the discussion below.

via analysis of material freshly irradiated at pH 7 or above, do not

show a signiﬁcant presence of Peak 2.

Preparative isolation of the Tpm⁵dC photoproducts

In a typical preparative run, 3.5 mL portions of Tpm⁵dC (0.5 mM, pH

7.5) were placed in Type 21 quartz cuvettes (see above) and

irradiated for 5 h at 254 nm in the cold as described above. (A total of

about 45 mL of 0.5 mM Tpm⁵dC was irradiated to ensure isolation of

a sufﬁcient amount of each photoproduct for NMR and mass spectral

study.) The resulting irradiated solution was partitioned directly on

the Capcell column, using an isocratic ﬂow of 3 mL/min. The eluent

used was 84% (25 mM ammonium formate)–16% MeOH. Each run

required 7 min under these conditions. Seven fractions were col-

lected: F1 (0.25–0.5 min), F2 (0.5–0.8 min), F3 (1–1.2 min), F4 (1.6–

2.1 min), F5 (2.6–3.1 min), F6 (3.1–4.1 min) and F7 (5.6–6.3 min).

Rechromatography of F1 using Gradient A on the Capcell column

showed that F1 contained the material corresponding to Peaks 1, 2, 3

and 4 in Fig. 1, along with some material corresponding to Peak 5.

The UV spectrum of the peak corresponding to F2 showed that this

fraction mainly corresponded to Peak 5, whereas the spectra of the

peaks collected in F3, F4 and F5 showed that these fractions

contained material corresponding to Peaks 6, 7 and 8, respectively.

Fraction F6 contained unreacted Tpm⁵dC, whereas the spectrum of

the peak corresponding to F7 showed that it corresponds to Peak 9

in Fig. 1.

We noted variation in the pH of solutions prepared from

different lyophilized tubes of Tpm⁵dC (see the last section of

Materials and Methods) after dissolution of the contents of such

tubes in doubly distilled water, followed by dilution to 0.5 mM in

Tpm⁵dC. The measured pH ranged from 4.6 to 7.5, with the

contents of most tubes yielding solutions with pH values above 7.

The reason for this variability is probably because of the tubes

containing differing amounts of residual ammonium formate after

lyophilization of the parent Tpm⁵dC, which was isolated using

ammonium formate eluent at pH 7; larger amounts of residual

ammonium formate would give solutions with pH values near 7.

The qualitative and quantitative aspects of the observed photo-

chemistry are similar at both pH 4.6 and pH 7–7.5. However, the

pH does affect the stability of the c,s Tpm⁵dC CBD corresponding

to Peak 1 (see below for details relevant to the identiﬁcation of this

compound); this compound readily decomposes, particularly at

acid pH, to form the corresponding TpT CBD, which elutes as

Peak 2 in Fig. 1. Although most of our studies were done with

samples at pH 7 or above, the run made to prepare Fig. 1 was done

with a sample irradiated at pH 4.6, so as to show the relationship of

elution time of the TpT c,s CBD to that of other photoproducts.

HPLC chromatograms similar to those shown in Fig. 1, obtained

Fraction 1 was chromatographed preparatively using Gradient A

and the individual compounds corresponding to Peaks 1, 2, 3 and 4

were collected (F1-1, F1-2, F1-3 and F1-4); each of these fractions

contained one major product. Each major product was then

repuriﬁed for NMR and mass spectrometric study, using the same

chromatographic conditions used for their isolation. Between 0.3

and 1.0 mg of each product was obtained using this procedure. In

the subsequent text, we will use the following designations for the

various products, which correspond to individual fractions in

408 Lech Celewicz et al.

Figure 2. Normalized overlaid UV spectra of various photoproducts produced in the Tpm⁵dC system, along with those of related compounds. Panel

a displays the normalized spectral absorbance of the c,s and t,s CBD, namely II (C1) (2a.2) and V (C4) (2a.1) as a function of wavelength in nanometers.

Panel b presents the spectra corresponding to the (6-4) products arising from irradiation of Tpm⁵dC (C5, VI) (2b.1) and m⁵dCpT (2b.2). Panel c shows the

spectra of the Dewar adducts produced upon irradiation of Tpm⁵dC (C3, VII) (2c.1) and m⁵dCpT (2c.2). Finally, Panel d displays the spectra of acrylamidine

product C7 (VIII) (2d.2) and the depyrimidation product C6 (IX) (2d.1) produced upon irradiating Tpm⁵dC. The spectra were obtained by using the ‘‘on the

ﬂy’’ spectrometric capabilities of a Hewlett-Packard 1040A DAD. The spectrum of each puriﬁed compound of interest was captured as its corresponding

peak eluted from the Capcell column; the eluent used was 25 mM ammonium formate ﬂowing at a rate of 2 mL/min.

which they are found: C(ompound)1: F1-1; C2 (F1-2); C3 (F1-3);

C4 (F1-4); C5 (F2); C6 (F3); C7 (F4); C8 (F5); C9 (F7). (Table 1

summarizes the relationship between the product[s] in each fraction

[e.g. F2 contains C5], as well as to the peak[s] in Fig. 1 that

correlate to each fraction [e.g. Peak 5] and the structure of the

corresponding product[s] in that fraction [e.g. VI], as given in

Scheme 1.)

observations that this compound decomposes to form the

corresponding III and that it reverts upon irradiation with 254

nm light to form parent Tpm⁵dC, provides conclusive evidence

that C1 is indeed described by the structure given by II.

The mass spectral and proton NMR data are consistent with the

assignment of C4 as the t,s CBD of Tpm⁵dC (V). Electrospray mass

spectrometry in the negative ion mode indicates that C4 has the

expected value for [MꢀH] of 544. The assignable NMR data are as

follows: (D₂O, 600 MHz, 3-(trimethylsilyl)propionic-2,2,3,3-d₄

acid, sodium salt [TSP]) Tp: 5.26 (1H, H19(dd)), 3.70 (1H, H29(m)),

2.59 (1H, H20(dd)), 4.78 (1H, H39(m)), 4.22 (1H, H49(dd)), 3.69

(2H, (H59 and H50)(m)), pm⁵dC: 5.75 (1H, H19(dd)), 2.14 (1H,

H29(m)), 1.93 (1H, H20(dd)), 4.52 (1H, H39(overlapping dd)), 4.00

(1H, H49 (broadened doublet)), 4.11 (2H, (H59 and H50)(m)). Two

singlet peaks, each corresponding to three protons, occur at 1.54 and

1.45 ppm and can be assigned to the two methyl groups. Two broad

singlets occur at 4.17 and 4.12 ppm and probably correspond to the

protons at Position 6 on the two pyrimidine rings. These assign-

ments are consistent with the extensive sets of NMR data provided

by Liu and Yang (31) and by Kan et al. (43) for the TpT t,s CBD, as

well as the data given by Douki and Cadet (29) for the m⁵dCpT t,s

Characterization of the photoproducts

Products C1 and C4 have UV spectra similar to those of m⁵C-Thy

mixed dimers (23). (The UV spectra of C1 and C4 are shown in

Panel a of Fig. 2). This suggests that one of these products is the c,s

CBD and the other the t,s CBD of Tpm⁵dC. (Steric constraints

make it unlikely that cis,anti and trans,anti CBD can be formed in

dinucleotides, such as Tpm⁵dC.) Irradiation of either of these two

compounds with 254 nm light results in its complete conversion to

Tpm⁵dC, as evidenced by production of a single compound with an

absorption spectrum and HPLC behavior identical to that of

Tpm⁵dC.

When C1 is allowed to stand, it converts into C2. Irradiation of

C2 with 254 nm light produces TpT (IV). Irradiation of TpT itself

at 254 nm, using the conditions described above, produces as

a major product a compound with a retention time and UV

spectrum identical to that of C2. On the basis of the published

chromatographic behavior of the c,s TpT dimer (see, for example,

ﬁg. 5 in Ref. 29), C2 can be identiﬁed as this compound (III in

Scheme 1). Therefore, C1 can be identiﬁed as the c,s CBD of

Tpm⁵dC (II), which undergoes thermal deamination to form III as

shown in Scheme 1. This implies that C4 is the Tpm⁵dC t,s CBD

(V), shown in Scheme 1. This latter compound is thermally stable

upon standing at 48C and at room temperature.

Consistent with the identiﬁcation of C1 as II are mass spectral

data. In particular, the ESI mass spectrum in the negative ion mode

gives the expected value for [M ꢀ H] of 544. Under conditions

required to prepare a pure NMR sample, II decomposed to

a mixture of II and III, the c,s TpT CBD. However, the mass

spectral data and the UV absorption spectral proﬁle, along with the

Table 1. Correlation between fraction number (F), product designation

(C), peak number (as labeled in Fig. 1) and the corresponding structure in

Scheme 1

Peak label

in Fig. 1

Fraction

C(ompound)

Structure or identity

F1

F1-1

F1-2

F1-3

F1-4

F2

F3

F4

F5

F6

1, 2, 3, 4

1

C1, C2, C3, C4

II, III, V, VII

C1

C2

C3

C4

C5

C6

C7

C8

II

III

VII

V

VI

IX

VIII

I

2

3

4

5

6

7

8

Tpm⁵dC

9

Tpm⁵dC

C9

F7

IV

Photochemistry and Photobiology, 2005, 81 409

Table 2. Proton NMR chemical shifts (in ppm with respect to TSP as an

internal standard) of selected photoproducts of Tpm⁵dC. All spectra were

run in D₂O. For comparative purposes, the chemical shift data for the t,s

CBD and (6-4) product of TpT, given by Douki and Cadet (29), as well as

the corresponding data for the TpT Dewar adduct of Taylor et al. (25) are

also tabulated. The parentheses in the second column of the table denote the

nucleoside component of the dinucleoside monophosphate the NMR data in

that row refers to. Information about the multiplet structure of individual

resonances is given in the discussion of the structural assignment for each

species

Compound

H19 H29 H20 H39 H49 Me

H6

TpT t,s CBD (T)pT

Tp(T)

5.36 3.39 2.68 4.84 4.21 1.58 4.22

5.80 2.27 2.07 4.63 3.98 1.52 4.33

V (C4)

(T)pm⁵dC 5.26 3.70 2.59 4.78 4.22 1.54 4.12

Tp(m⁵dC) 5.75 2.14 1.93 4.52 4.00 1.45 4.17

TpT (6-4)

adduct

(T)pT

Tp(T)

6.27 1.55 2.25 4.08 3.76 1.83 5.17

6.60 3.15 2.67 4.82 4.23 2.40 8.09

VI (C5)

(T)pm⁵dC 6.18 1.31 2.13 3.86 3.70 1.68 5.01

Tp(m⁵dC) 6.54 3.02 2.58 4.83 4.15 2.35 8.03

Scheme 1. The photoreactions of Tpm⁵dC in aqueous solution. Also

shown are the thermal deamination reaction of the c,s CBD of Tpm⁵dC (II)

to form III, the corresponding CBD of TpT, and the photoinduced reaction

of III to form TpT (IV). The notation hm_Bindicates that this reaction is

most signiﬁcant when UV-B light is used to induce photoreaction.

TpT Dewar

adduct

(T)pT

Tp(T)

6.30 2.52 2.36 4.56 3.89 1.57 4.73

5.64 2.28 2.27 4.49 3.89 2.10 5.33

VII (C3)

(T)pm⁵dC 6.31 2.52 2.38 4.54

*

1.55 4.76

2.15 5.36

Tp(m⁵dC) 5.65 2.28 2.28 4.51

* These chemical shifts cannot be determined with certainty because of

extensive overlap in the regions where the corresponding resonances

occur.

CBD. (For comparative purposes, the chemical shift data for C4 (V),

as well as that for C3 and C5, are tabulated in Table 2, along with the

corresponding data given by Douki and Cadet (29) for the t,s CBD

and the (6-4) adduct of TpT and by Taylor et al. (25) for the TpT

Dewar adduct.)

H59 and H50 protons for the Tp and pm⁵dC ribose rings appear in

a region containing overlapping multiplets lying between 3.92 and

4.10 ppm. The preceding assignments are consistent with the

previous assignments for the (6-4) and Dewar adducts of TpT

given by Rycyna and Alderfer (44), Kan et al. (43), Taylor et al.

(25) and Douki and Cadet (29), as well as the data given for the

(6-4) product of m⁵dCpdT given in Ref. 29.

Product C5 has an absorption spectrum characteristic of a (6-4)

product, with a k_maxof 327 nm and k_minat 271 nm (Fig. 2b.1). The

molecular mass [M ꢀ H] of C5, obtained using ESI mass

spectrometry in the negative ion mode, is 544; this corresponds

to the expected molecular mass of the structure given by VI. When

the ESI spectrum was run in the positive ion mode, the

predominant peak was at 568, which corresponds to the sodium

adduct of VI in which the phosphate is protonated. Irradiation of

C5 with 312 nm light for 16 min converts it completely into

a compound with the same elution properties and absorption

spectrum as C3. The absorption spectrum of C3 (Fig. 2c.1) has

The UV spectra of C7 and C8, obtained using the spectral capture

capabilities of the DAD, suggest that they contain moieties that are

unsaturated; in addition, the two spectra are almost superimposable,

differing only slightly in the values of k_max; C7 has an absorption

maximum at 283 nm, whereas the k_maxfor C8 is at 281 nm. The k_min

for both compounds is at 239 nm. (The UV spectrum of C7 is shown

in Fig. 2d.2.) The compounds C7 and C8 rapidly interconvert;

HPLC analysis of either freshly ‘‘puriﬁed’’ product indicates that

both compounds are present. Because these compounds cannot be

individually maintained in a pure state, the mass spectrum of

a mixture containing both C7 and C8 was determined using ESI in

the positive ion mode. The (M þ H^þ) peak for these compounds

occurs at 520.2, whereas the (Mþ Na^þ) peak occurs at 542.2; an M

value of 519.2 corresponds to the mass that would be expected for

the acrylamidine product given by structure VIII. In this compound,

a CO fragment has been lost, but two hydrogens have been gained

for a net loss of 26 from the mass (545) of the parent Tpm⁵dC. This

type of compound has been previously characterized by Celewicz

and Shetlar (17) as the nature of photoproduct arising when either

m⁵C or m⁵dC are irradiated in water. It was shown in this study that

the acrylamidine type products exists as a mixture of two isomeric

compounds that differ only slightly in their UV spectra. It was

suggested that the difference between these two products is that one

corresponds to a compound in which the acrylamidine moiety takes

the E conformation and the other the Z conformation; the

acrylamidine moiety shown in VIII is in Z conformation. Further

evidence that C7 and C8 are acrylamidines is provided by the results

k

_max5 223 nm and a shallow minimum at 221 nm. Conversely,

C3, when irradiated with 254 nm light, is converted quantitatively

into a compound with elution properties and a UV spectrum

identical to those of C5. The molecular mass [M ꢀ H] of C3 is also

544, as expected for an adduct containing both the Thy and m⁵C

moiety. These observations indicate that C5 should be identiﬁed as

the (6-4) adduct of Tpm⁵dC (VI), whereas C3 is the corresponding

Dewar adduct of Tpm⁵dC (VII).

Analysis of the proton NMR spectra of these two compounds

yields data that are consistent with these identiﬁcations. In the case

of C5, the following proton NMR assignments can be made: (D₂O,

600 MHz, TSP) Tp: 6.18 (1H, H19(dd)), 1.31 (1H, H29(ddd)), 2.13

(1H, H20(ddd)), 3.86 (1H, H39(m)); 3.70 (1H, H49(m)), 3.96

(1H, H59(dd), 3.86 (1H, H50(m)), 1.68 (3H, Me(s)), 5.01 (1H,

H6(s)), pm⁵dC: 6.54 (1H, H19(dd)), 3.02 (1H, H29(ddd)), 2.58

(1H, H20(m)), 4.83 (1H, H39(m)), 4.15 (1H, H49(m)), 3.80 (1H,

H59(dd)), 3.70 (1H, H50(m)), 2.35 (3H, Me (s)), 8.03 (1H, H6,(s)).

For C3, similar assignments can be made as follows: (D₂O, 500

MHz, TSP) Tp: 6.31 (1H, H19(d)), 2.52 (1H, H29(m)), 2.38 (1H,

H20(m)), 4.54 (1H, H39(m)); 1.55 (3H, Me, (s)), 4.76 (1H, H6(s)),

pm⁵dC: 5.65 (1H, H19(dd)), 2.28 (1H, H29(m)), 2.28(1H, H20(m)),

4.51 (1H, H39(m)), 2.15 (3H, Me(s)), 5.36 (1H, H6(s)). The H49,

410 Lech Celewicz et al.

of a chemical study. Treatment of a mixture of these two compounds

with 1,19-carbonyldiimidazole, as described in Ref. 17, produced

Tpm⁵dC in good yield; this was proven by the correspondence of the

HPLC behavior and UV spectrum of the product with that of

Tpm⁵dC.

depicted by IX should be considered as a symbol representing

these multiple forms, rather than a single structure. Although it is

difﬁcult to make a deﬁnitive assignment of the resonances in this

spectrum, we were able to make several assignments, as follows:

(D₂O, 600 MHz, TSP, AP 5 apyrimidinic) Tp: 6.33 (1H,

H19(broad triplet)), 1.91 (3H, Me(s)), 7.67 (1H, H6(s)); pAP:

The proton NMR spectrum of freshly isolated C7 is quite

complex because of a number of overlapping multiplets that arise

because of the presence of both C7 and C8 (see above); such

overlap is particularly evident in the resonances corresponding to

the acrylamidine moiety and its attached sugar. However, the

following NMR features can be assigned with some degree of

certainty. The relative spectral intensities in several cases show that

one isomer of VIII predominates in amount; the chemical shifts

corresponding to the dominant form (when evident) are italicized.

(D₂O, 600 MHz, TSP, AA 5 aminoacrylamidine) Tp: 6.33 (1H,

H19(t)) and 6.29 (1H, H19(dd)), 2.26 (2H, H29, H20(m)), 3.98 (1H,

H39(m)), 3.80 (1H, (H59or H50)(dd)), 3.84 (1H, (H59 or H50)(dd)),

1.898 and 1.905 (3H, Me(s)), 7.65 (1H, H6(s)) and 7.66 (1H,

H6(s)); pAA: 5.44 (1H, H19(t)) and 5.39 (1H, H19(t)), 2.43 and

2.63 (1H, H29, H20(both complex multiplets)), 4.78 (1H,

H39(overlapped by the H₂O peak), 1.70 and 1.72 (3H, Me(s)),

7.35 and 7.36 (1H, 3H(s)). Complex peaks centered at 3.98 and

4.20 probably contain resonances corresponding to the Tp 49

resonance and the pAA 49, 59 and 50 resonances. On the basis of

the ratio of the areas of the AA methyl peaks, it can be calculated

that the two isomers are present in about a 2:1 ratio at the ambient

temperature of the NMR experiment.

5.62 (;0.5 H, H19(broadened

t or overlapping dd)) and

5.59(;0.5H,H19(broadened d)). (The presence of two peaks for

H19 suggests that both the a and b anomeric forms of the 29-

deoxyribose moiety are indeed contributing to the NMR spectrum.)

There are two complex peaks centered at 2.44 and 2.55; these

probably correspond to the Tp and pAP 29 and 20 protons. Two

very complex regions of overlapping resonances lie between 3.77

and 3.89 ppm and between 3.89 and 4.07 ppm. These regions

probably contain the Tp and the pAP 49, 59 and 50 resonances. In

addition, there are four broadened resonances between 4.20 and

4.50 in which splitting and shoulders can be detected. A couple of

these have integrations corresponding to half a proton, suggesting

that they are representative of different anomeric forms of the AP

deoxyribose moiety. These peaks, in total, probably correspond to

resonances due to H39 for the Tp and pAP moieties. Between 2.10

and 2.60 ppm are three complex resonances corresponding to

a total of four protons; these likely correspond to the H29 and H20

protons of the Tp and pAP groupings.

To conﬁrm that our assignment of C6 as IX is correct, we

prepared an authentic sample of IX. We depurinated the

dinucleoside monophosphate TpdA by letting it stand in 0.1 M

HCl at 378C for either 5 h or 15 h; after the former time, small

amounts of TpdA remained, whereas incubation for 15 h produced

quantitative depurination. The protocol used is similar to that

described by Weinfeld et al. (45) for production of IX from TpdA.

Analysis of the reaction mixture by HPLC using Gradient A on the

Capcell column, after neutralization with 1 M ammonium formate,

showed that only two products were present in signiﬁcant amounts,

namely adenine and the desired depurinated product. Compound

IX was puriﬁed using the Capcell column with 90% ammonium

formate–10% MeOH ﬂowing at 3 mL/min as eluent; under these

conditions, the adenine peak eluted as a tailing peak at 0.7 min and

the desired IX at 1.99 min. The UV spectral properties of the

puriﬁed IX, as measured by the DAD (which measures spectral

intensities at 2 nm intervals) were very similar to those in the

literature (46) (k_max5 267, k_min5 235, A₂₆₀/A₂₈₀5 1.6).

The HPLC chromatogram of puriﬁed IX using Gradient A on

the Capcell column yielded a single peak that had the same

retention time as C6 in a freshly irradiated solution of Tpm⁵dC (0.5

mM, irradiated in the cold at 254 nm for 2 h). Overlay of the UV

absorption spectra of C6 and IX indicated that the spectrum of C6

was identical to that of IX. Coinjection of IX with the irradiated

Tpm⁵dC solution resulted in signiﬁcant enhancement of the height

of C6; there was no indication of peak broadening. Because IX and

C6 coelute and have identical UV absorption spectra, we conclude

that C6 is indeed IX.

Product C9 has an absorption spectrum and HPLC retention time

identical to that of authentic TpT (IV). Additional support for this

assignment is provided by identity of the chemical shift values of

the various resonances in the proton NMR spectrum, run at 600

MHz in D₂O with TSP as standard, with those published by

Rycyna and Alderfer (44).

When the absorption spectra of C6 and C9 are compared in

normalized form, they are almost superimposable; C6 displays

a k_maxat 267 nm and k_minat 235 nm; see Fig. 2.d.1. However, the

molecular mass of C6 corresponds to that expected for TpT that has

lost one Thy (or Tpm⁵dC that has lost a m⁵C). After replacement of

the m⁵C moiety with OH, the net loss in molecular mass for

Tpm⁵dC is 107; this corresponds to an expected molecular mass of

438 for a species, which is protonated on the phosphate moiety.

The molecular mass [M ꢀ H] determined in the negative ion mode

by ESI is 436.8, whereas the molecular mass of the [MþNa^þ] peak

determined by MALDI mass spectrometry is 460.98. These two

masses indeed correspond to the molecular mass expected for

a species of structure IX, assuming that the phosphate is protonated

in both the ESI and the MALDI mass spectra. The structure

displayed as IX can be regarded as the 5-ester of 39-thymidylic acid

with 2-deoxy-b-erythro-pentofuranose with the free 29-deoxyri-

bose taking the b anomeric form. Aspects of the proton NMR data

are consistent with the assignment of C6 as having this general

type of structure; in particular, there is only one methyl resonance

and one resonance corresponding to H6. However, the sugar

regions of the proton NMR spectrum are very complex with many

resonances being signiﬁcantly broadened, as compared with the

corresponding resonances in the compounds studied above. This is

probably because of the fact that the free deoxyribose moiety can

exist in two ring forms that are anomeric at C1 in that sugar moiety,

as well as in an open chain conformation; these multiple forms

probably equilibrate with one another reasonably rapidly, account-

ing for the fact that we see only one HPLC peak. Thus the structure

Compounds II, V, VI and VIII are primary products

in the photoreaction of Tpm⁵dC

In one set of experiments, we followed the course of the

photochemical reaction of Tpm⁵dC as a function of irradiation

time. For these studies, samples of Tpm⁵dC (0.5 mM, pH 7.4) were

irradiated at 254 nm and 48C in self-masking 0.7 mL cuvettes as

described in Materials and Methods. Irradiation times of 8, 16, 32,

Photochemistry and Photobiology, 2005, 81 411

64, 96 and 128 min were used and the irradiance was 16 J m^ꢀ2s^ꢀ1

at the upper surface (area 5 0.7 cm²) of the irradiation cell

(see Materials and Methods). Calculation shows that after 8 min the

dose, on the basis of surface area was 7.68 kJ m^ꢀ2, which

corresponds to an amount of energy of 0.54 J being received at the

surface of the cell. Thus, a corresponding average dose of about

0.77 J cm^ꢀ3, on the basis of volume, was received by the contents

of the cell after 8 min irradiation. Doses at other times can be

readily obtained by multiplication by an appropriate time scaling

factor. The irradiated solutions (20 lL injections) were analyzed by

HPLC using Gradient A on the Capcell column. Each of the

products could be detected using 230 nm radiation. However, peak

areas corresponding to VI and the two forms of VIII were

considerably larger when detected at 327 and 280 nm, respectively.

The presence of ﬁve compounds (II, V, VI and the two forms of

VIII) was evident after 8 min, as evidenced by the appropriate

retention time and absorption spectrum for the each peak. The

peaks areas for VI and the two forms of VIII increased steadily

throughout the irradiation period. The peak areas of II and V both

increased in area between 8 and 64 min and then remained

essentially constant over the next two irradiation intervals. The

photochemical behavior observed for the peaks corresponding to

II, V, VI and VIII is that expected for primary photoproducts. The

behavior exhibited by II and V is that expected of a compound that

reaches a photoequilibrium in which photoreversal of product to

parent compound occurs at the same rate as it is formed.

retention time as IX. This impurity has a very similar absorption

spectrum to that of IX and, indeed, may be IX. However, using

a second impurity with a signiﬁcantly larger peak area (eluting at

1.05 min and with an area that remained essentially constant with

increasing irradiation time) as an internal standard, it was found

that the difference in the normalized area of the peak eluting at 1.2

min after a given irradiation time and the normalized area

corresponding to the impurity peak in the unirradiated sample

steadily increased. Between irradiation times of 4 and 24 min, this

area increased about ﬁve-fold. At the same time, the area

corresponding to Tpm⁵dC decreased to about 90% of its original

area. Although these results tend to support the idea that IX is

a primary photoproduct, rather than a secondary photoproduct,

they must be at this point be regarded as only suggestive. More

extensive work will be required to prove this point beyond doubt.

Effect of sample deoxygenation on the photoreactions

of Tpm⁵dC

We did a study to determine whether the qualitative photochemisty

of Tpm⁵dC was altered by removal of oxygen from the reaction

mixture, as compared with the photochemistry occurring when the

irradiated samples were equilibrated with air. These comparative

studies were done on two samples (0.45 mM, 10 mM in sodium

phosphate buffer, pH 7.0) contained in quartz NMR tubes from

Wilmad (Buena, NJ) (sample size, 500 lL). One sample was ﬁtted

with a rubber septum designed for NMR tubes (Wilmad) and the

other with a normal plastic top. The tube with the septum was

deoxygenated for 20 min by bubbling with 99.997% nitrogen

ﬂowing through a long syringe needle (Wilmad) that reached to the

bottom of the NMR tube; a second small narrow-gauge needle

piercing the septum allowed egress of the nitrogen. The two tubes

were taped in equivalent positions to the block described in Materials

and Methods and irradiated for 1 h at 228C. The resultant reaction

mixtures were analyzed via HPLC using the Capcell column.

For studies designed to examine the effect of deoxygenation on

production of the two isomers of VIII and IX, the eluent was 86%

ammonium formate–18% MeOH ﬂowing at a rate of 3 mL/min.

These conditions nicely resolve IX from any possible Thd

contamination (which elutes at 1.47 min). The resultant chromato-

grams for the two systems indicated no qualitative differences

between the aerated and deoxygenated system in the region where

IX and the two isomers of VIII elute, although there were slight

quantitative differences in peak areas for the peaks corresponding

to these three compounds in the chromatograms for the two

irradiated systems.

The c,s TpT CBD (III) and the Tpm⁵dC Dewar adduct (VII)

could be ﬁrst detected after 64 min of irradiation and then increased

in amount over the next two irradiation intervals. This is the

behavior expected of secondary products; in these two cases, III is

putatively produced by thermal deamination of II and VII by

photoisomerization of the initially produced VI. In these time-

course studies, we did not detect a peak eluting at a retention time

corresponding to that of the TpT (6-4) adduct (see Fig. 1, Panel b).

The question of whether IX is a primary photoproduct

In the studies described above, we were unable to determine if IX

was a primary product. Under the conditions of the gradient used for

HPLC analysis in those studies, the peak for this particular compound

was poorly resolved from a peak corresponding to an impurity in the

parent Tpm⁵dC. Only after 64 min irradiation could we determine that

IX was deﬁnitely present in the reaction mixture. We therefore

developed HPLC conditions that achieved better separation of IX

from this impurity peak. For our studies of IX, 0.7 mL samples of

Tpm5dC (0.45 mM, 10 mM in sodium phosphate, pH 7) were

irradiated at 254 nm and 48C as described in Materials and Methods.

Irradiation times of 4, 8, 16 and 24 min were used. The samples were

analyzed on the Capcell column using 86% ammonium formate–14%

MeOH, ﬂowing at a rate of 3 mL/min. We focused on the appearance

of a peak at 1.2 min, which corresponded in elution time to IX; under

these conditions, the two isomers of VIII elute at 1.8 and 2.6 min,

whereas parent Tpm⁵dC elutes at 3.5 min. The peaks corresponding to

II, V, VII and VII elute as a set of poorly resolved peaks nearer to the

beginning of the chromatogram.

For the sake of completeness, a second set of HPLC runs was

made on the same two samples to study the effects of sample

deoxygenation on formation of II, V, VI and VII. Therefore,

Gradient A (see Materials and Methods) was used on the Capcell

column. Again, there were not large differences in the peak areas

for those peaks corresponding to these compounds in the

chromatogram corresponding to the deoxygenated solution, as

compared with the aerated solution.

The results outlined above, taken as a whole, suggest that oxygen

is not an effective quencher for those excited states participating in

formation of the various photoproducts of Tpm⁵dC discussed in this

paper and also indicate that oxygen is not required as a reactant on

the pathway leading to loss of m⁵C by Tpm⁵dC to form IX.

Even using these improved conditions, we found examination of

the formation of IX to be problematic. It is produced in

a considerably smaller amount than the photoproducts whose

study was described in the previous section. An impurity, present

in a small amount in unirradiated samples, elutes at the same

412 Lech Celewicz et al.

Effects of irradiation wavelength on the distribution of

products in the Tpm⁵dC system

acetophenone to sensitize the triplet state reactions of m⁵dCpT.

They found evidence for four products; the two main products

were the c,s TpT CBD and a second product, eluting with the same

retention time as the c,s m⁵dCpT CBD that slowly decomposed to

form the c,s dimer of TpT. Minor products with the same retention

time as the t,s CBD of TpT and m⁵dCpT were also observed.

We carried out a similar experiment with Tpm⁵dC. A solution

(0.7 mL, pH 7.4) that was 0.5 mM in Tpm⁵dC and 14 mM in

acetophenone was irradiated in the cold as described in Materials

and Methods, using the long-wavelength lamp emitting mainly at

312 nm. The reaction was followed as a function of time, using

Gradient A and the Capcell column. After 128 min, about 45% of

the parent had reacted and three products were present, namely II,

V and III, the c,s TpT CBD. The ratio of the integrated peak areas

for the two Tpm⁵dC dimers (II/V), measured at 230 nm, was 2.2,

whereas the ratio of the areas corresponding to II and III was about

7.9. There were no peaks corresponding to the Tpm⁵dC (6-4)

product or the corresponding Dewar adduct. Similarly, peaks

suggesting the formation of VI, VIII and IX were absent. These

results indicate that the triplet state of Tpm⁵dC can react to form

the two CBD of this dinucleotide monophosphate and, at the same

time, supports (but does not prove) the hypothesis that the reactions

to form VI, VIII and IX occur from the singlet state of Tpm⁵dC.

Figure 1a,b indicate that there are signiﬁcant differences in the

distribution of products formed when Tpm⁵dC is irradiated for

128 min with 254 nm light and with UV-B light. As pointed out

previously, the (6-4) product VI (corresponding to Peak 5 in Fig. 1)

is present in only small amounts in this system when irradiated

with radiation that is predominately in the UV-B, whereas it is the

dominant product when irradiation occurs at 254 nm. Another

signiﬁcant difference between the product distributions at the two

wavelengths is the considerably greater amounts of the dimer

products II and V that are present in the Tpm⁵dC system when

irradiation is done with long wavelength light. An enhanced feeling

for the differences in the two proﬁles in Fig. 1 can be obtained by

comparison of the areas corresponding to the various classes of

products at each irradiation wavelength. When the irradiation was

done using 254 nm light (Fig. 1b), the integrated areas

corresponding to Peaks 1 and 4 were 194 and 232, respectively,

whereas the areas of Peaks 3 and 5 were 120 and 1242,

respectively. However, when the long wavelength light was used

(Fig. 1a), the corresponding areas for Peaks 1 and 4 were 660 and

873 and the areas of Peaks 3 and 5 were 97 and 35. The relatively

small areas corresponding to II and V in the system irradiated at

254 nm, as compared with the area corresponding to VI, is to be

expected. These two compounds have very signiﬁcant absorbance

values at 254 nm (see Fig. 2a) and, thus, continuous destruction of

these dimers by absorption of light of this wavelength would occur.

As described above, a study of the peak area for each of the various

products formed as a function of irradiation time, using 254 nm

radiation with time points of 8, 16, 32, 64, 96 and 128 min and

using the irradiation conditions described in Materials and

Methods, showed that the areas corresponding to II and V had

reached a steady state at 64 min, whereas the peak area

corresponding to VI and the sum of the peak areas of the two

isomers corresponding to VIII continued to increase with time of

irradiation. These results indicate that if one desires to isolate CBD

II and V, then irradiations should be done with lamps emitting

primarily long wavelength radiation. On the other hand, irradiation

at 254 nm is the best route to take if a relatively high yield of the

(6-4) product VI is needed. (It should be noted that direct

comparisons of peak areas for various products in Fig. 1a,b are not

valid because the amount of light absorbed by the two irradiated

parent systems is different. Thus, it is not feasible to compare peak

areas to evaluate the relative reactivity of Tpm⁵dC to form a given

individual product under the two sets of irradiation conditions.)

A study was made of the effect of ﬁltering the 254 nm lamps

with Vycor shields on the amounts of IX and the two isomers of

VIII formed during 2 h of irradiation under the conditions

described in Materials and Methods. It was found that the amount

of IX decreased by a factor of about 3.3, as compared with the

situation where the lamp was unﬁltered, whereas the summed areas

corresponding to C7 and C8 decreased by a factor of about 1.1.

Thus, it appears that short-wavelength components of the unﬁltered

germicidal lamp enhance the rate of the depyrimidation process

while having little effect on acrylamidine formation.

Thermal stability of the c,s CBD of Tpm⁵dC near neutral pH

As mentioned in the Introduction, deamination of c,s Tpm⁵dC

CBD in the context of DNA is thought to play a role in ﬁxing

mutations at sites in which a Thy residue has replaced a m⁵C

moiety. To gain quantitative information about the kinetics of the

thermal deamination of the c,s CBD of Tpm⁵dC (II) to form III,

we followed the kinetics of decomposition of II as a function of

temperature using a sample obtained by irradiating under the

following conditions: 0.5 mM Tpm⁵dC, pH 7.4, irradiation time of

128 min with long-wavelength radiation. (Although external

buffers [e.g. phosphate] were not added to this solution, it is likely

the pH of 7.4 is because of the presence of residual ammonium

formate in the solution that remained after lyophilization; for

example, 25 mM ammonium formate that has stood in air for

several days has a pH of around 7.2.) We analyzed the solution

immediately after irradiation, using Gradient A on the Capcell

column; there was a barely detectable peak corresponding to III in

the freshly irradiated solution. For each run, 70 lL portions of

a freshly irradiated solution of Tpm⁵dC were placed in ﬁve 1.5 mL

screw-top plastic vials (Sarstedt, Newton, NC). These vials were

placed in covered water-ﬁlled beakers situated in water baths

maintained at a constant temperature. The remaining irradiated

solution was stored on ice. After 6 min, to allow the vials to come

to equilibrium with the surrounding water, the clock was started

and, after set time intervals, individual vials were removed and

plunged into an ice bath. After agitation of each vial on a vortex

mixer, followed by centrifugation, 20 lL was injected on the

Capcell column and chromatographed using Gradient A as

described in Materials and Methods; samples for each time point

were run in duplicate. The integrated area, measured at 230 nm,

corresponding to Peak 1 (see Fig. 1) for each incubation time was

averaged. The averaged integrated area corresponding to the

Tpm⁵dC t,s CBD (Peak 4), which is stable over the heating periods

used and which can thus serve as an internal standard, was also

evaluated for each time point. A normalization factor for each time

point, to be applied to corresponding averaged peak areas for Peak

Acetophenone photosensitizes formation of both

the c,s and t,s Tpm⁵dC CBD

In one of their studies on the photochemistry of the m⁵dCpT

system, Douki and Cadet examined the results of using

Photochemistry and Photobiology, 2005, 81 413

1, was calculated by dividing the averaged area for Peak 4

associated with the ﬁrst time point by the averaged area for Peak 4

corresponding to the time point to be corrected. The experimental

peak area for Peak 1, at each time point, was then multiplied by its

corresponding normalization factor. This normalization eliminates

the effects of injection volume variation.

they calculated the half-life to be 6.8 h. In another study, Lemaire

and Ruzsicska (48) measured the rate constant for deamination of

the thymidylyl-(39-59)-29-deoxycytidine c,s CBD as 1.5 3 10^ꢀ3

min^ꢀ1in 0.05 M phosphate buffer at 258C and pH 7; from this value,

the half-life can be evaluated as 7.7 h. In yet another study,

Hariharan and Johns (49) examined the stability of the cytidine

cyclobutane homodimers formed when cytidylyl-(39-59)-cytidine

was irradiated and found that the originally formed dimer had a half

life of 7.2 h at 98C and pH 6.8. The product from this reaction,

a heterodimer containing both uracil and Cyt residues, underwent

deamination with a half-life of about 5 h.

The natural log of the normalized peak area for Peak 1 at each

time, measured at 230 nm, was plotted as a function of time of

incubation using Graph III for the Macintosh (Computer

Associates, Islandia, NY); the slope of each line was then

evaluated using a least squares ﬁtting procedure included in this

program. Runs were made at four temperatures; the values for k,

the pseudo ﬁrst-order rate constants (in h^ꢀ1), obtained from these

four runs are as follows: 548C, 0.164; 598C, 0.297; 658C, 0.431;

718C, 0.684. All correlation coefﬁcients were 0.995 or higher.

The above kinetic data were plotted appropriately to estimate the

enthalpy and entropy of activation for the deamination reaction of

Tpm⁵dC c,s dimer. According to the Eyring transition state theory

of chemical kinetics, the temperature dependence of the rate

constant for a reaction in solution can be mathematically described

as shown in the equation given immediately below (see, for

example, Ref. 47, pp. 476–480), in which the transmission

coefﬁcient has been set equal to one:

Comparison of the appropriate half-lives, given above for several

Cyt-containing c,s CBD, with those of II indicates that methylation

of Cyt at the 5-position results in signiﬁcant stabilization of this

moiety toward thermal deamination. At 258C, for example, the

calculated half-life of II at pH 7.4 was about 58 h, as compared with

7.7 h for the analogous dimer of TpdC (see above).

The c,s dimer of Tpm⁵dC is quite stable to deamination under

the conditions of our experiments, particularly at low temperatures.

However, it is not as stable as the corresponding c,s CBD of

m⁵dCpT, for which the half life was measured at pH 7 at 258C in

0.01 M phosphate buffer. For this deamination reaction, the rate

constant was measured as 10^ꢀ5min^ꢀ1(56 3 10^ꢀ4

corresponding half-life is 1155 h 5 48.1 days.

h

^ꢀ1) (29). The

à

k ¼ ðRT=NhÞe^{DS =R}e^{ꢀDH =RT}

The effect of pH on deamination of the Tpm⁵dC c,s CBD

This can be rewritten in the form of ln(k/T) 5 ln(R/Nh) þ DS^à/R ꢀ

DH^à/RT. According to the latter equation, the value of DH^àcan be

evaluated from the slope of a graph of ln(k/T) versus 1/T, and the y

intercept of such a graph yields a quantity from which DS^àcan be

calculated. We analyzed the above data using this equation; the re-

sulting plot yields values of DH^à573.4 kJ/mol and DS^à5ꢀ103.5 J/K

mol for the deamination reaction of the c,s CBD of Tpm⁵dC. The

correlation coefﬁcient describing the linear ﬁt of the rate constant

data for the Eyring plot was 0.992. Note that, because the rate

constants listed in the previous paragraph are in h^ꢀ1, the corre-

sponding value of R/Nh must also be in h^ꢀ1K^ꢀ1; therefore we used

Figure 1 shows HPLC chromatograms obtained after HPLC

analysis of a system in which the pH of the irradiated Tpm⁵dC

solution was 4.6. In these chromatograms, the presence of the c,s

TpT CBD (Peak 2) is evident. Indeed, this particular chromatogram

was chosen to display the elution behavior of this compound.

When the irradiation was done with the long-wavelength lamp

(Fig. 1a), the area of Peak 2 was comparable with that

corresponding to the Dewar adduct VII. The HPLC analysis was

done immediately after irradiation, which implies that III was

formed during the course of the irradiation. However, after carrying

out analogous irradiations at pH 7.5 accompanied by immediate

HPLC analysis of the photolyzed solution, we found that TpT c,s

CBD was present in, at most, trace amounts in the reaction

mixtures. To establish that this was an effect of pH on deamination

of the Tpm⁵dC c,s dimer, rather than some other factor present in

the system at pH 4.6, we adjusted the pH of 0.7 mL of the solution,

originally at a pH of 4.6, to 7.4 by making it 10 mM in phosphate.

Irradiation of this sample for 90 min in the cold at long

wavelengths, followed immediately by HPLC analysis, produced

a chromatogram in which there was no evidence of formation of III

during the course of the irradiation. (We found no evidence for

deamination of V, VI and VII in any of our experiments.)

In a qualitative experiment designed to test the effect of standing

at pH 4.6 on II, as compared with standing at near neutral pH, we

irradiated a 0.7 mL sample at pH 4.6 for 90 min in the cold with

long wavelength radiation and then placed 70 lL portions of the

irradiated sample into two plastic 1.5 mL Sarstedt tubes with screw

caps. To the contents of one tube, we added 7 lL of 100 mM

sodium phosphate (pH 7.4). We then allowed the two tubes to

stand for 64 h in a refrigerator and analyzed the two solutions via

HPLC as described previously. The HPLC chromatogram of the

sample at pH 4.6 displayed only a trace of Tpm⁵dC c,s CBD,

whereas the other sample, adjusted in pH with phosphate buffer,

retained around 85% of the original II.

a value for R/Nh of 7.483 10¹³^ꢀ1K^ꢀ1in the evaluation of DS^à.

h

Using the values of DH^àand DS^àevaluated above, estimates can

be made of the half-lives of the c,s Tpm⁵dC CBD at various

temperatures; we evaluated these at four different temperatures of

interest: 378C, 17.8 h; 258C, 58.3 h; 48C, 592.3 h 5 24.7 days and

08C, 959.8 h 5 39.9 days. The value for the calculated half-life at

378C is comparable with that experimentally measured at 378C for

the c,s m⁵C-Thy mixed CBD, namely 21.6 h (23).

A second, less extensive run was made in which the pH was

buffered at pH 7.55 by making the solution 5 mM in sodium

phosphate. Three values of k (in h^ꢀ1) were determined: 548C,

0.188; 598C, 0.271 and 658C, 0.477; all correlation coefﬁcients

were 0.997 or better. Plotting and analysis of these data, as des-

cribed above, led to values of DH^à5 75.3 kJ and DS^à5ꢀ97.3 J/K,

which is in reasonable agreement with the studies made directly on

the unbuffered irradiated solution at pH 7.4. The correlation

coefﬁcient for the linear ﬁt of the rate constant data for the Eyring

plot was 0.997.

It is interesting to compare the calculated half-lives of the c,s

Tpm⁵dC CBD with the half-lives that have been measured for

corresponding CBD formed when dinucleotides containing Cyt and

Thy or two Cyt are irradiated. Douki and Cadet (30) measured the

rate constant for deamination of the c,s CBD of dCpT at pH 7 in 10

mM phosphate buffer at room temperature; from this rate constant,

414 Lech Celewicz et al.

To obtain an idea of how much faster II underwent deamination

at pH 4.6 than at pH 7.4, we studied the kinetics of deamination of

II at 548C, as described in the previous section. The pseudo ﬁrst-

order rate constant determined for the deamination process was k 5

3.4 h^ꢀ1, which compares with the corresponding value of k50.169

Evidence that acrylamidine products analogous to C7 and

C8 are formed in the m⁵dCpT system

In view of our observation that products C7 and C8, the two

isomeric forms of acrylamidine product VIII, are formed in the

Tpm⁵dC system, we did a study seeking evidence for the formation

of analogous compounds in the m⁵dCpT system. The parent

compound (0.5 mM), which was prepared using essentially the

method described in Ref. 29, was irradiated with 254 nm radiation

under the same conditions described in Materials and Methods.

The HPLC of 20 lL of this reaction mixture, run using Gradient A

on the Capcell column, showed two closely eluting peaks at 6.75

and 6.82 min with absorption spectra very similar to those of C7

and C8; parent m⁵dCpT eluted at 7.62 min. Using the spectral

capture capability of the DAD, it was found that the peak eluting

ﬁrst had a k_maxat 277 nm, whereas the peak eluting at 6.82 min

had k_max5 283 nm; both had k_minat 240 nm. Using the same

gradient on Column A, but with an injection size of 500 lL of

irradiated material, the material eluting between 6 and 6.4 min,

corresponding to the unresolved peaks of the putative acrylami-

dines, was collected and submitted for ESI mass spectral analysis

in the positive ion mode. The [M þ H^þ] peak had a mass of 520.2,

whereas there was a [M þ Na^þ] peak at 542.1. These are the same

masses as were measured for the mixture of C7 and C8 in the

Tpm⁵dC system. These results strongly suggest that acrylamidine

products, analogous to VIII, are produced because of photolysis of

the m⁵CpdT system. It is likely that the reason these compounds

were not detected by Douki and Cadet is that they adhere strongly

to the column and are not eluted until the solvent contains an

appreciable concentration of MeOH. The HPLC proﬁle for the

irradiated m⁵dCpT system, shown in ﬁg. 5 in Ref. 29, was run in an

isocratic mode using 25 mM ammonium formate as eluent.

In agreement with the results of Douki and Cadet (29), we found

that compounds were present in the reaction mixture with the

expected HPLC and UV spectral properties of the c,s and t,s CBD,

the (6-4) product and Dewar adduct of m⁵dCpT. Using the Capcell

column with Gradient A, we isolated the (6-4) adduct; this

compound eluted between 1.2 and 1.7 min. We then irradiated the

(6-4) product at 312 nm with the BLE-1T158 lamps for 12 min.

The HPLC, run on Column A using 25 mM ammonium formate at

3 mL/min, displayed a dominant peak eluting at 0.75 min

corresponding to the Dewar adduct; a relatively small amount of

(6-4) remained after 12 min irradiation. We did not isolate these

two compounds from this reaction mixture; however, using the

spectral capture capability of the Hewlett-Packard DAD, we

obtained the UV spectrum of the (6-4) product and of the Dewar

adduct. These spectra are displayed in Panels b and c in Fig. 2,

along with the corresponding spectra of the Tpm⁵dC products.

Both the (6-4) product and the Dewar adduct of m⁵dCpT show

absorption spectra that are shifted signiﬁcantly to the red in the

region below 260 nm, as compared with the corresponding spectra

of the Tpm⁵dC products. This would be expected because the

m⁵dCpT products contain a saturated m⁵C ring, whereas the

Tpm⁵dC products contain a saturated Thy ring.

h

^ꢀ1obtained at pH 7.4. Thus, the deamination process at pH 4.6 is

about 20-fold faster than the same process at pH 7.4, at least at

548C. The rate of deamination has also been found to be dependent

on pH for c,s CBD of dinucleoside monophosphates containing

Cyt, rather than m⁵C. For example, detailed studies by Lemaire and

Ruzsicska indicated that the deamination rate of the c,s CBD of

TpdC and dCpT are dependent on both pH and buffer

concentration; both have maximal rates of deamination at a pH

of around 3.8.

The above experiments suggest that careful control of pH is

required to maximize the stability of the Tpm⁵dC c,s CBD.

Although it does slowly deaminate at physiological pH when

stored at temperatures near 08C, the rate of its deamination is

accelerated markedly in acid solution (e.g. pH 4.6). Further studies

will be required to elucidate the detailed dependence of the rate of

deamination upon pH and its mechanistic implications.

Effects of temperature of irradiation on deamination

Almost all of our irradiations were carried out at 48C. However, we

did an experiment to qualitatively compare the photochemistry

occurring at ambient temperature with that taking place at 48C. We

irradiated two 0.7 mL samples of 0.5 mM Tpm⁵dC (pH 4.6) at 254

nm for 120 min under almost identical conditions, the only

difference being that one sample was maintained at 48C and the

other was maintained at room temperature (about 228C). In both

cases, the irradiation setup was that described in Materials and

Methods; in each case, the sample being irradiated rested on an

aluminum block (9 cm long, 7.5 cm wide and 5 cm high), so as to

maintain the temperature constant during the period of photoreac-

tion. Both samples were placed on ice immediately after photolysis

and successively analyzed by HPLC. The signiﬁcant differences

between the two systems, in terms of the HPLC proﬁle, were that the

peak area for C6 (IX) is increased about three-fold, as compared

with the irradiation done at 48C, whereas the peak area of TpT (IV)

was enhanced about seven-fold at 228C. The amounts of the TpT c,s

CBD (III) found in both systems were comparable in amount. The

integrated peak area corresponding to II, measured at 230 nm, was

about 15% larger in the system irradiated at 48C, as compared with

that irradiated at 228C. One additional interesting observation was

made. The leading edge of the peak corresponding to II for the

sample irradiated at room temperature contains a substance that

displays an absorption maximum at about 285 nm. The elution time

and the spectrum of this new compound are very similar to that of

the Z isomer of the acrylamidine formed by reaction of m⁵C with

water (17). A trace of this material was also present in the leading

edge of the peak corresponding to II in the system irradiated at 48C.

(Indeed, the UV spectrum of the material eluting in Peak 1 in Panel

b in Fig. 1 indicates some contamination of this peak with putative

acrylamidine; this was not true of Peak 1 in Panel a.) The greater

amount of this acrylamidine product in the system irradiated at 228C

may be because of the larger amounts of m⁵C released in this system

by formation of IX. This enhanced amount of m⁵C could then react

in a secondary photoprocess to form the larger amount of the

acrylamidine observed in the system irradiated at 228C.

Comparison of the results of the present study with those of

previous work

There are substantial differences in the results of our study of the

photochemistry of Tpm⁵dC and those obtained by Douki and Cadet

(29). (In both studies, experiments for preparative purposes were

Photochemistry and Photobiology, 2005, 81 415

done using light from germicidal lamps emitting predominately at

254 nm.) In particular, we have isolated and characterized six

photoproducts containing m⁵C, as well as one product in which the

m⁵C has been lost from the dinucleoside phosphate. On the other

hand, Douki and Cadet found that the c,s CBD and (6-4) adduct of

TpT were the major products formed, although two other

compounds were present in their reaction mixtures in amounts

too small to characterize. In our studies, the TpT c,s CBD appears

as a secondary product resulting from thermal deamination of the

corresponding c,s CBD of Tpm⁵dC; at most, a trace of the (6-4)

adduct of TpT was observed in our photoreacted solutions of

Tpm⁵dC and then only after high doses had been delivered to the

solutions being irradiated.

Therefore, our samples received 100.8 J over the 5 h period used

for preparative irradiations; this corresponds to a dose of 28.8 3

10⁶J m^ꢀ35 28.8 J cm^ꢀ3(or alternatively, 2.88 3 10⁵J m^ꢀ2). Thus,

our samples, irradiated for preparative purposes, received a dose

about 27-fold larger on a volume basis than the solutions of Douki

and Cadet (or 192-fold larger on a surface area basis). Therefore, it

cannot be maintained that initially formed m⁵C-containing

products were initially formed and destroyed because these

products are the predominant products in our systems after much

higher doses of radiation. A second possibility might be that, for

some reason, the two TpT containing products are rapidly

produced at small doses, but then the rate of production of m⁵C-

containing products predominates as the dose increases. However,

the results described in a previous section, in which we showed that

II, V, VI and the two isomers of VIII were primary products,

suggest that, even at doses comparable with those of Douki and

Cadet, the (6-4) adduct and the c,s CBD of TpT are not formed as

primary photoproducts. In that section, it was shown that after an

average dose of 0.77 J cm^ꢀ3(7.7 kJ m^ꢀ2based on average surface

dose) as compared with the 1.06 J cm^ꢀ3received by the sample of

Douki and Cadet (surface dose, 1.5 kJ m^ꢀ2), the only detectable

HPLC peaks present were those due to II, V, VI and the two

isomers of VIII. Thus, the doses received by our sample irradiated

for 8 min (either based on average dose per unit volume or per unit

area) and the sample irradiated for 1 h by Douki and Cadet are

within the same order of magnitude. Hence, it is questionable

whether such an explanation can account for the disparateness of

the result of Douki and Cadet and those reported in this study.

Still another possibility to be considered is that the distribution

of absorption events within the solutions being irradiated might be

quite different, in some way affecting the products being formed.

For example, absorption of a signiﬁcantly higher number of

photons near the solution surface in the more concentrated solution

of Douki and Cadet could lead to a higher percentage of excited

states being produced near the surface that could, in some manner,

react differently (e.g. deaminate more rapidly) than molecules in

bulk solution, thus promoting formation of III and the (6-4) adduct

of TpT. Indeed, the premise for such an argument can be shown to

hold. The 254 nm radiation incident on the samples in both our

experiments and in those of Douki and Cadet was essentially

completely absorbed. The ratio of the e value at 254 nm for

Tpm⁵dC to that of e_max(at 271 nm) can be measured to be 1.38.

Thus, e₂₅₄can be evaluated, using the estimated value of e_maxof

18 150 given previously, to be about 13 150. The absorbance of

a 0.5 mM solution at 254 nm along a 1 cm path length of our

irradiation cell can be calculated to be about 6.6, whereas a similar

calculation for the 3.6 mM solution of Douki and Cadet along

a path length of 0.14 cm produces an absorbance of 6.2. However,

the fraction of incident radiation that was absorbed along a given

distance of passage through the two irradiated solutions was quite

different, being about 7.2-fold higher for the solution of Douki and

Cadet. This implies that the number of photons absorbed and

number of excited states produced in this latter solution, at any

given depth traversed by the light beam, should be about 7.2-fold

higher, suggesting that the number of excited state molecules

produced near the surface in the solution of Douki and Cadet were

7.2-fold higher than the number produced near the surface in our

solutions. Although the considerations outlined in this paragraph

do establish one signiﬁcant distinction between conditions holding

in our irradiated solutions and those of Douki and Cadet, the

outlines of a mechanism describing how this difference could

In the paragraphs below, we will discuss possible answers to the

following question: ‘‘Why is there such a disparity between the

results of the two sets of experiments?.’’ This discussion makes use

of some of the results outlined in preceding sections.

One obvious possibility is that the Tpm⁵dC used in our

experiments differs in some important manner from that of Douki

and Cadet. This particular factor is impossible to assess

experimentally because samples of the material used by Douki

and Cadet in their experiments no longer exist (T. Douki, personal

communication). However, control experiments run by Douki and

Cadet, using NMR and mass spectrometry, indicated that the

Tpm⁵dC used for their experiments did not contain signiﬁcant

amounts of TpT, excluding this factor as an explanation for the

formation of c,s CBD and (6-4) adducts of TpT as the predominant

products in their studies of this system (J. Cadet and T. Douki,

personal communication). It should be noted that the photochem-

istry of m⁵dCpT reported by Douki and Cadet and that observed in

the limited set of experiments we carried out on this compound is

quite similar. The main differences in the results of these two

studies are that the amount of deaminated CBD product present

immediately after irradiation was signiﬁcantly smaller in our

studies than was found in the studies of Douki and Cadet and that

we did not observe formation of the deaminated (6-4) adduct of

m⁵dCpT in our experiments. (Control experiments by Douki and

Cadet also showed that their m⁵dCpT was did not contain

detectable TpT [J. Cadet, personal communication].)

Another alternative to be considered is that the doses, received

by the sample in the experiment of Douki and Cadet and by the

samples used in our experiments, were dramatically different. For

example, it might be postulated that m⁵C-containing photoproducts

were formed and then destroyed by the absorption of much higher

doses of radiation in the system studied by Douki and Cadet.

However, the following reasoning rules out that possibility. Douki

and Cadet indicate that their sample, irradiated for 1 h with

a germicidal lamp, received a total UV dose of 1.5 kJ m^ꢀ2when

expressed in terms of surface area. Thus, the lamp in their

experimental setup provided an average irradiance of 0.42 J m^ꢀ2

s^ꢀ1to the surface of their Tpm⁵dC solution. The 25 mL sample

used (at a concentration of about 3.6 mM) was contained in a 15

cm diameter petri dish; thus their irradiated solution had a surface

area of about 1.77 3 10^ꢀ2m²and a depth of about 1.4 3 10^ꢀ3m.

Therefore, their sample received an amount of energy correspond-

ing to 26.5 J during its irradiation; on a volume basis, this

corresponds to a dose of 1.06 3 10⁶J m^ꢀ35 1.06 J cm^ꢀ3. The

irradiance that our solutions (0.5 mM) were exposed to was

measured as 1.6 mW cm^ꢀ25 16 J m^{ꢀ2 ꢀ1}; our preparative

s

irradiations were done for 5 h and the surface area of the 3.5 mL

cuvettes in which samples were irradiated was 3.5 3 10^ꢀ4m².

416 Lech Celewicz et al.

account for the discrepancy between the two sets of results are

murky.

the reaction conditions used by Douki and Cadet differ in some

unidentiﬁed manner from those we used and that this divergence is

responsible for the differing results.

Another possibility that can be considered is that the Tpm⁵dC

starting material of Douki and Cadet (or the aqueous medium in

which it was dissolved) contained a small amount of impurity

(other than TpT) that, in some manner, promoted direct formation

of the c,s CBD and (6-4) products of TpT or, alternatively,

dramatically increased the rate of deamination of the initially

formed Tpm⁵dC c,s CBD and (6-4) adduct. For example, such an

agent might promote highly efﬁcient deamination of the m⁵C

residue in the excited state of Tpm⁵dC, resulting in formation of the

c,s CBD and (6-4) adducts of TpT as the dominant products.

However, this must be considered a highly speculative idea that

because of the non-accessibility of the parent starting material(s)

cannot be tested.

One general conclusion that can be drawn, on the basis of the

studies described in this article, is that the type of photoinduced

deamination process suggested byDouki and Cadet (i.e. deamination

of Tpm⁵dC in its excited state, leading to formation of the c,s CBD

and (6-4) adduct of TpT as predominant primary photoproducts,

rather than the corresponding Tpm⁵dC compounds) does not act as

an important mechanism under the reaction conditions we used.

However, the operation of such a mechanism cannot be ruled out

under conditions that were not encompassed by our studies.

SUMMARY AND CONCLUSIONS

There are several other potential causes for the differences in our

results and those of Douki and Cadet. These workers did not buffer

their solution during irradiation (T. Douki, personal communica-

tion) and did not indicate the pH of the unbuffered solution in

which reaction occurred. Thus, it is possible that pH differences

might play a role in explaining the disparity between our results. In

our studies, we have examined in detail at the photochemistry of

Tpm⁵dC at pH 4.6 and 7.4 in the absence of added buffer and at pH

7.55 in 5 mM phosphate buffer (see above). In each of these cases,

we found that the Tpm⁵dC products described above were formed

as primary photoproducts; however, the stability of the c,s CBD of

Tpm⁵dC, once formed, is dependent on pH, being considerably

more susceptible to deamination at the lower pH. We did not ﬁnd

evidence for formation of either the c,s CBD or the (6-4) adduct of

TpT as primary photoproducts at pH 7.4 in unbuffered solution or

in the phosphate-buffered solution. At pH 4.6, there was always

some c,s TpT CBD present in freshly irradiated solutions; however,

this likely resulted from deamination of parent Tpm⁵dC c,s CBD.

The (6-4) adduct of TpT was observed only as a trace photoproduct

at pH 4.6 and then only after high doses of 254 nm radiation.

The limited scope of our studies, on the effect of pH on the

photochemistry of Tpm⁵dC, do not allow us to rule out the

possibility that this factor plays a role in explaining the disparity

between our results and those of Douki and Cadet. A signiﬁcantly

different photochemistry, as compared with that occurring in our

studies, might result in solutions irradiated at pH values

signiﬁcantly higher or lower than the ones we used in our studies.

Another possibility to consider is that the temperature of

irradiation could affect the photochemistry. Douki and Cadet did

not control the temperature during their irradiations (T. Douki,

personal communication), whereas our preparative irradiations were

done at about 48C. We made a comparative study of the

photochemistry of Tpm⁵dC at room temperature (about 228C)

with that occurring at 48C when the two samples are irradiated under

otherwise identical conditions. There was not a signiﬁcant differ-

ence in the amount of c,s TpT CBD formed in the two experiments.

There was no evidence of formation of signiﬁcant amounts of the

TpT (6-4) adduct at either temperature. These results suggest that

temperature of irradiation is probably not important in explaining

the dissimilarities of our results and those of Douki and Cadet.

In summary, although a number of factors can be suggested (and

a number discounted) as potential reasons for the disparity in the

photochemistry of Tpm⁵dC described by the experiments reported

in this study and by those of Douki and Cadet, it is not possible to

make any deﬁnitive statements concerning the reason(s) for the

deviations in the two sets of results. The possibility remains that

We have shown that UV irradiation of the biologically relevant

dinucleoside monophosphate Tpm⁵dC leads to formation of eight

products. Two of these are CBD of Tpm⁵dC, namely the c,s and t,s

isomers (II and V). The (6-4) product (VI) and Dewar adduct (VII)

of Tpm⁵dC have also been isolated and characterized. Two

acrylamidine products (VIII) are formed, presumably via reaction

of the m⁵C moiety with water. We have also isolated a product

(IX) in which the dinucleotide has lost its m⁵C moiety via

a depyrimidination reaction. The c,s CBD of TpT (III) is formed as

a secondary thermal reaction product, resulting from deamination

of II. Finally, small amounts of TpT (IV) were isolated.

The distribution of photoproducts is dependent upon the

wavelength of light used for promoting photoreaction. Irradiation

with either unﬁltered or Vycor ﬁltered lamps emitting predomi-

nately at 254 nm leads to preferential formation of the (6-4) product,

the Dewar adduct, the acrylamidine products and the depyrimidi-

nated dinucleotide. Irradiation with lamps that emit primarily at 312

nm yields the two CBD of Tpm⁵dC as predominant products,

although the Dewar adduct and the two acrylamidine products are

also present in signiﬁcant amounts in the resultant photoreaction

mixture. The acrylamidine products, analogous to VIII, and the

depyrimidinated photoproduct IX have not been previously

observed in the context of dinucleotide photochemistry.

The c,s CBD of Tpm⁵dC undergoes thermal deamination to form

the corresponding dimer of TpT. Rate constants for this process

have been measured at four different temperatures and the enthalpy

and entropy of activation for the deamination reaction have been

evaluated at pH 7.4. These results are likely to be relevant to

understanding the mutational hot spots that are observed at

Tpm⁵dC sites in UV-irradiated eukaryotic DNA (Ref. 13 and

references therein). The rate of the deamination reaction of II

appears to be considerably enhanced at the acidic pH of 4.6, as

compared with pH 7.4, which has implications for the design of

studies to detect II itself in UV-irradiated DNA and chromatin.

As mentioned above, the c,s Tpm⁵dC dimer is thought to play

a role in UV-induced mutagenesis. However, the other products

characterized here may also have signiﬁcance for understanding

the deleterious effects of UV radiation on eukaryotic cellular

systems, particularly that in the UV-B region. If such products are

indeed produced photochemically at key Tpm⁵dCpdG sites (e.g. in

regions of chromosomal DNA associated with transcriptional

regulation), then incomplete repair (e.g. replacement of the

damaged m⁵C with Cyt without subsequent methylation to restore

the original base sequence) could induce alterations in gene

expression with attendant biological consequences.

Photochemistry and Photobiology, 2005, 81 417

Acknowledgements—Research support from the National Science Founda-

tion (CHE-0131203) is gratefully acknowledged. Also acknowledged is

the Bio-organic, Biomedical Mass Spectrometry Resource (A. L. Burlin-

game, Director), supported by NIH Division of Research Resources Grant

RR 01614 and, especially, David Maltby who took special pains to

obtain needed mass spectra in a timely manner. We also thank Peter

Madrid, who ran a number of the early mass spectra on the Waters

Micromass ZQ 4000 mass spectrometer, and to the laboratory of Profes-

sor Kip Guy for allowing us to have access to this instrument. Drs. Jean

Cadet and Thierry Douki provided helpful information concerning the

conditions under which they did their experiments on Tpm⁵dC; these two

researchers, as well as the referees, made useful comments that were very

useful in revision of the manuscript for publication. Finally, we acknowl-

edge the help of Professor Arnold Falick, who ran MALDI mass spectra

as required.

20. Ehrlich, M., M.-F. Dove and L.-H. Huang (1986) Photolysis of

methylated DNA. Photobiochem. Photobiophys. 11, 73–79.

21. Barna, T., J. Malinowski, P. Holton, M. Ruchirawat, F. F. Becker and

J.-N. Lapeyre (1988) UV-induced photoproducts of 5-methylcytosine in

a DNA context. Nucleic Acids Res. 16, 3327–3340.

22. Mitchell, D. L. (2000) Effects of cytosine methylation on pyrimidine

dimer formation in DNA. Photochem. Photobiol. 71, 162–165.

23. Shetlar, M. D., V. J. Basus and A. M. Falick (2004) The cyclobutane

dimers of 5-methylcytosine and their deamination products. Photochem.

Photobiol. Sci. 4, 968–979.

24. Johns, H. E., M. L. Pearson, J. C. LeBlanc and C. W. Helleiner (1964)

The ultraviolet photochemistry of thymidylyl-(39-59)-thymidine. J. Mol.

Biol. 9, 503–524.

25. Taylor, J.-S., D. S. Garrett and M. P. Cohrs (1988) Solution-state

structure of the Dewar pyrimidinone photoproduct of thymidylyl-(39-

59)-thymidine. Biochemistry 27, 7206–7215.

26. Bose, S. N. and R. J. H. Davies (1982) The photoreactivity of purine-

containing deoxydinucleoside monophosphates in ice. Biochem. Soc.

Trans. 10, 350–351.

REFERENCES

27. Davies, R. J. H. (1995) Ultraviolet radiation damage in DNA. Biochem.

1. Adams, R. L. P. and R. H. Burdon (1984) Molecular Biology of DNA

Methylation. Springer-Verlag, New York.

Soc. Trans. 23, 407–418.

28. Zhao, X. D., S. Nadji, J. L. F. Kao and J. S. Taylor (1996) The structure

of d(TpA)*, the major photoproduct of thymidylyl-(39-59)-deoxyade-

nosine. Nucleic Acids Res. 24, 1554–1560.

29. Douki, T. and J. Cadet (1994) Formation of cyclobutane dimers and (6-

4) photoproducts upon far-UV photolysis of 5-methylcytosine-contain-

ing dinucleoside monophosphates. Biochemistry 33, 11942–11950.

30. Douki, T. and J. Cadet (1992) Far-UV photochemistry and photosen-

sitization of 2-deoxycytidylyl-(39-59)-thymidine: isolation and charac-

terization of the main photoproducts. J. Photochem. Photobiol. B: Biol.

15, 199–213.

31. Liu, F.-T. and N. C. Yang (1978) Photochemistry of cytosine

derivatives. 1. Photochemistry of thymidylyl-(39-59)-deoxycytidine.

Biochemistry 17, 4865–4876.

32. Taylor, J.-S., H.-F. Lu and J. J. Kotyk (1990) Quantitative conversion

of the (6-4) photoproduct of TpdC to its Dewar valence isomer upon

exposure to simulated sunlight. Photochem. Photobiol. 51, 161–167.

33. Douki, T. and J. Cadet (2001) Individual determination of the yield of

the main UV-induced dimeric pyrimidine photoproducts in DNA

suggests a high mutagenicity of CC photolesions. Biochemistry 40,

2495–2501.

34. Cadet, J. and P. Vigny (1990) The photochemistry of nucleic acids. In

Bioorganic Photochemistry, Vol. 1 (Edited by H. Morrison), pp. 1–272.

John Wiley and Sons, New York.

35. Ruzsicska, B. P. and D. G. E. Lamaire (1995) DNA photochemistry. In

CRC Handbook of Organic Photochemistry and Photobiology (Edited

by W. M. Horspool and P.-S. Song), pp. 1289–1315. CRC Press, Boca

Raton, FL.

36. Douki, T., T. Zalizniak and J. Cadet (1997) Far-UV-induced dimeric

photoproducts in short oligonucleotides: sequence effects. Photochem.

Photobiol. 66, 171–179.

37. Blais, J., T. Douki, P. Vigny and J. Cadet (1994) Fluorescence quantum

yield determination of pyrimidine(6-4)pyrimidone photoadducts. Pho-

tochem. Photobiol. 59, 402–404.

2. Friso, S., S.-W. Choi, G. G. Dolnikowski and J. Selhub (2002) A

method to assess genomic DNA methylation using high-performance

liquid chromatography/electrospray ionization mass spectrometry. Anal.

Chem. 74, 4526–4531.

3. Sharonov, A., T. Gustavsson, S. Marguet and D. Markovitsi (2003)

Photophysical properties of 5-methylcytidine. Photochem. Photobiol.

Sci. 2, 362–364.

4. Millar, D. S., R. Holliday and G. W. Grigg (2003) Five not four: history

and signiﬁcance of the ﬁfth base. In The Epigenome: Molecular Hide

and Seek (Edited by S. Beck and A. Olek), pp. 3–20. Wiley-VCH,

Weinheim, Germany.

5. Robertson, K. D. (2003) DNA methylation and chromatin-unraveling

the tangled web. Oncogene 21, 5361–5369.

6. Szyf, M. (2003) Targeting DNA methylation in cancer. Ageing Res.

Rev. 2, 299–328.

7. Richardson, B. (2003) Impact of aging on DNA methylation. Ageing

Res. Rev. 2, 245–261.

8. Issa, J.-P. (2003) Living longer: the aging epigenome. In The

Epigenome: Molecular Hide and Seek (Edited by S. Beck and A.

Olek), pp. 141–149. Wiley-VCH, Weinheim, Germany.

9. Walter, J. and M. Paulsen (2003) Epigenetic trouble: human diseases

caused by epimutations. In The Epigenome: Molecular Hide and Seek

(Edited by S. Beck and A. Olek), pp. 103–116. Wiley-VCH, Weinheim,

Germany.

10. You, Y. H., C. Li and G. P. Pfeifer (1999) Involvement of 5-methyl-

cytosine in sunlight-induced mutagenesis. J. Mol. Biol. 293, 493–503.

11. You, Y.-H. and G. P. Pfeifer (2001) Similarities in sunlight-induced

mutational spectra of CpG-methylated transgenes and the p53 gene in

skin cancer point to an important role of 5-methylcytosine residues in

solar UV mutagenesis. J. Mol. Biol. 305, 389–399.

12. Ikehata, H., T. Masuda, H. Sakata and T. Ono (2003) Analysis of

mutation spectra in UVB-exposed mouse skin epidermis and dermis:

frequent occurrence of C-T transition at methylated CpG-associated

dipyrimidine sites. Environ. Mol. Mutagen. 41, 280–292.

13. Lee, D.-H. and G. P. Pfeifer (2003) Deamination of 5-methylcytosines

within cyclobutane pyrimidine dimers is an important component of

UVB mutagenesis. J. Biol. Chem. 278, 10314–10321.

14. Shaw, A. A. and M. D. Shetlar (1990) 3-ureidoacrylonitriles: novel

products from the photoisomerization of cytosine, 5-methylcytosine and

related compounds. J. Am. Chem. Soc. 112, 7737–7742.

15. Hom, K., G. Strahan and M. D. Shetlar (2000) Ring opening

photoreactions of cytosine and uracil with ethylamine. Photochem.

Photobiol. 71, 243–253.

38. Chattopadhyaya, J. B. and C. B. Reese (1979) Some observations

relating to phosphorylation methods in oligonucleotide synthesis.

Tetrahedron Lett. 5059–5062.

39. Reese, C. B. (1978) Chemical synthesis of oligo-nucleotides and poly-

nucleotides by phosphotriester approach. Tetrahedron 34, 3143–3179.

40. Stawinski, J., T. Hozumi, S. A. Narang, C. P. Bahl and R. Wu (1977)

Arylsulfonyltetrazoles, new coupling reagents and further improve-

ments in triester method for synthesis of deoxyribooligonucleotides.

Nucleic Acids Res. 4, 353–371.

41. Fox, J. J. andD. Shugar (1952) Spectrophotometric studiesof nucleic acid

derivatives and related compounds as a function of pH. II. Natural and

synthetic pyrimidine nucleosides. Biochim. Biophys. Acta 9, 369–384.

42. Fox, J. J., D. van Praag, I. Wempen, I. L. Doerr, L. Cheong, J. E. Knoll,

M. L. Eidenoff, A. Bendich and G. B. Brown (1959) Thiation of

nucleosides. II. Synthesis of 5-methyl-29-deoxycytidine and related

pyrimidine nucleosides. J. Am. Chem. Soc. 81, 178–187.

43. Kan, L.-S., L. Voituriez and J. Cadet (1988) Nuclear magnetic resonance

studies of cis-syn, trans-syn, and 6-4 photodimers of thymidylyl(39-

59)thymidine monophosphate and the cis-syn photodimers of

16. Shaw, A. A. and M. D. Shetlar (1989) Ring-opening photoreactions of

cytosine and 5-methylcytosine with aliphatic alcohols. Photochem.

Photobiol. 49, 267–271.

17. Celewicz, L. and M. D. Shetlar (1992) The photochemistry of 5-

methylcytosine and 5-methyl-29-deoxycytidine in aqueous solution.

Photochem. Photobiol. 55, 823–830.

18. Privat, E. and L. C. Sowers (1996) Photochemical deamination and

demethylation of 5-methylcytosine. Chem. Res. Toxicol. 9, 745–750.

19. Ehrlich, M. and M.-F. Dove (1983) Photolysis at 254 nm of 5-

methyldeoxycytidine. Photobiochem. Photobiophys. 6, 121–126.

418 Lech Celewicz et al.

thymidylyl(39-59)thymidine cyanoethyl phosphotriester. Biochemistry

27, 5796–5803.

44. Rycyna, R. E. and J. L. Alderfer (1985) UV irradiation of nucleic acids:

formation, puriﬁcation and solution conformational analysis of the ‘6-4’

lesion of dTpdT. Nucleic Acids Res. 13, 5949–5963.

45. Weinfeld, M., M. Luzzi and M. C. Paterson (1989) Selective hydrolysis

by exo- and endonucleases of phosphodiester bonds adjacent to an

apurinic site. Nucleic Acids Res. 17, 3735–3745.

sites: design of a ³²P-postlabeling assay for apurinic sites in DNA.

Biochemistry 29, 1737–1743.

47. Chang, R. (2000) Physical Chemistry for the Chemical and Biological

Sciences. University Science Books, Sausalito, CA.

48. Lemaire, D. G. E. and B. P. Ruzsicska (1993) Kinetic analysis of the

deamination reactions of cyclobutane dimers of thymidyly-39,59-29-

deoxycytidine and 29-deoxycytidylyl-39,59-thymidine. Biochemistry 32,

2525–2533.

46. Weinfeld, M., M. Luzzi and M. C. Paterson (1990) Response of phage

49. Hariharan, P. V. and H. E. Johns (1968) Dimer photoproducts in

cytidylyl-(39-59)-cytidine. Photochem. Photobiol. 8, 11–22.

T4 polynucleotide kinase toward dinucleotides containing apurinic

Article Doi

The photochemistry of thymidylyl-(3′-5′)-5-methyl-2′- deoxycytidine in aqueous solution

DOI: 10.1562/2004-06-15-RA-201.1

Source and publish data:

Authors:

Article abstract of DOI:10.1562/2004-06-15-RA-201.1

Full text of DOI:10.1562/2004-06-15-RA-201.1

Products guided by the article

R&D Labs maybe for 76958-58-2

Relevant to this article

Hot Product