Synthesis of dinucleoside acylphosphonites by phosphonodiamidite chemistry and investigation of phosphorus epimerization

Article abstract of DOI:10.3762/bjoc.11.19

The reaction of the diamidite, (iPr₂N)₂PH, with acyl chlorides proceeds with the loss of HCl to give the corresponding acyl diamidites, RC(O)P(N(iPr)₂)₂ (R = Me (7), Ph (9)), without the intervention of sodium to give a phosphorus anion. The structure of 9 was confirmed by single-crystal X-ray diffraction. The coupling of the diamidites 7 and 9 with 5'-O-DMTr-thymidine was carried out with N-methylimidazolium triflate as the activator to give the monoamidites 3'-O-(P(N(iPr)₂)C(O)R)-5'-O-DMTr-thymidine, and further coupling with 3'-O-(tert-butyldimethylsilyl)thymidine was carried out with activation by pyridinium trifluoroacetate/N-methylimidazole. The new dinucleoside acylphosphonites could be further oxidized, hydrolyzed to the H-phosphonates, and sulfurized to give the known mixture of diastereomeric phosphorothioates. The goal of this work was the measurement of the barrier to inversion of the acylphosphonites, which was expected to be low by analogy to the low barrier found in acylphosphines. However, the barrier was found to be high as no epimerization was detected up to 150 °C, and consistent with this, density functional theory calculations give an inversion barrier of over 40 kcal/mol.

Full text of DOI:10.3762/bjoc.11.19

Synthesis of dinucleoside acylphosphonites by
phosphonodiamidite chemistry and investigation of
phosphorus epimerization
William H. Hersh
Full Research Paper
Open Access
Address:
Beilstein J. Org. Chem. 2015, 11, 184–191.
Department of Chemistry and Biochemistry, Queens College and the
Graduate Center of the City University of New York, Queens, NY
11367-1597, USA
doi:10.3762/bjoc.11.19
Received: 02 November 2014
Accepted: 15 January 2015
Published: 30 January 2015
Email:
William H. Hersh - william.hersh@qc.cuny.edu
This article is part of the Thematic Series "Nucleic acid chemistry".
Guest Editor: H.-A. Wagenknecht
Keywords:
acylphosphine; acyl phosphite; chiral; DFT; NMR; nucleic acids;
oligonucleotides
© 2015 Hersh; licensee Beilstein-Institut.
License and terms: see end of document.
Abstract
The reaction of the diamidite, (iPr2N)2PH, with acyl chlorides proceeds with the loss of HCl to give the corresponding acyl
diamidites, RC(O)P(N(iPr)2)2 (R = Me (7), Ph (9)), without the intervention of sodium to give a phosphorus anion. The structure of
9 was confirmed by single-crystal X-ray diffraction. The coupling of the diamidites 7 and 9 with 5′-O-DMTr-thymidine was carried
out with N-methylimidazolium triflate as the activator to give the monoamidites 3′-O-(P(N(iPr)2)C(O)R)-5′-O-DMTr-thymidine,
and further coupling with 3′-O-(tert-butyldimethylsilyl)thymidine was carried out with activation by pyridinium trifluoroacetate/N-
methylimidazole. The new dinucleoside acylphosphonites could be further oxidized, hydrolyzed to the H-phosphonates, and sulfur-
ized to give the known mixture of diastereomeric phosphorothioates. The goal of this work was the measurement of the barrier to
inversion of the acylphosphonites, which was expected to be low by analogy to the low barrier found in acylphosphines. However,
the barrier was found to be high as no epimerization was detected up to 150 °C, and consistent with this, density functional theory
calculations give an inversion barrier of over 40 kcal/mol.
Introduction
Antisense oligonucleotides, which are modified oligonucleo- changes that have occurred since the first antisense drug was
tides that bind to mRNA in order to inhibit translation to approved in 1998 [8]. However, despite ongoing clinical trials,
proteins, have been intensively investigated for decades [1,2]. only recently has a second antisense drug [3] been approved [9].
Modifications of native DNA or RNA are required to permit
pharmaceutical use, for example, for conferring nuclease resis- Most (albeit not all) antisense reagents, including the two
tance, enhanced binding to mRNA, and recruitment of approved drugs, are phosphorothioates, in which one of the
RNase H. A selection of recent publications [3-7] illustrates the terminal oxygen atoms on the phosphodiester internucleotide
184

Beilstein J. Org. Chem. 2015, 11, 184–191.
linker is replaced by sulfur. A potential issue with phosphoro- Although the required acylphosphonites are essentially
thioates is the creation of a new stereocenter at every phos- unknown, acylphosphonates 3 (Figure 1) have been extensively
phorus linker. This results in the formation of an enormous studied over the past few decades [32], both for their synthetic
number of diastereomers that may have different properties, utility [33-37] and as compounds of biological interest [38,39].
such as nuclease stability and RNase H interactions. Some They have also been used for phosphorothioate synthesis, as
[5,10] but not all [11,12] of the non-phosphorothioate antisense will be described shortly. The straightforward route for the syn-
reagents avoid the diastereomer problem, but the extensive thesis of 3 involves the Arbuzov reaction of a trialkyl phosphite
study of phosphorothioates is reason enough to investigate the ester with an acid chloride [40], necessarily giving the dialkyl
synthesis of P-chiral phosphorothioates [13-19]. However, none acylphosphonate with identical alkoxy groups. Thus, no compa-
of the reported methods seem to allow for routine, high-yield rable synthesis of 1 could be carried out since the dinucleoside
synthesis with high stereocontrol [20-23]. The extension to acylphosphonite would require different nucleoside moieties.
phosphorothioate RNA for applications of RNA interference
has also been described [24].
One example can be found in the literature of a non-Arbuzov
route leading to the dinucleoside analog of 3 where as reported
We recently reported the synthesis of a number of chiral disul- by Hata et al., 3a [41] gives 3b via sequential coupling [42,43]
fide sulfurizing reagents [25] and the results of the sulfurization (Scheme 1). While the deoxygenation of phosphine oxides is
of phosphite triesters in order to look for stereoselectivity [26]. known [44,45], deoxygenation of the functionally much more
In order to change such a method into a practical synthesis route complex phosphonate 3b to the dinucleoside analog of 1 was
for P-chiral phosphorothioates, a trivalent dinucleoside phos- not attempted. Only one example of an acylphosphonite analog
phorus moiety is required that undergoes rapid epimerization of 1 was found in the literature, namely, a dinucleoside formate
under the sulfurization conditions – that is, the goal was to carry reported by Caruthers et al. [46]. As shown in Scheme 1, the
out a dynamic kinetic resolution with formation of a single reaction of H-phosphonodiamidite 4 with chloroformate 5 gave
epimeric phosphorothioate.
acyl diamidite 6, which by following standard nucleoside
coupling protocols gave the formate phosphonite 1a. The initial
In order to carry out the required dynamic kinetic resolution, it P–C bond formation yielding 6 was odd, since although the ad-
was hypothesized that acylphosphonites 1 might undergo rapid dition of sodium was intended to give the phosphorus anion
epimerization by comparison to their known acylphosphine (iPr2N)2P− from 4, the authors noted that there was in fact no
analogs 2 (Figure 1). Mislow previously showed that trialkyl evidence of any reaction of 4 with sodium. The phosphonite 1a
and triarylphosphines slowly undergo epimerization only at was reported on the basis of 31P NMR analysis to exist as a
temperatures above approximately 130 °C, with activation mixture of two diastereomers, but the possibility of epimeriza-
barriers near 33 kcal/mol [27], but that acylphosphines 2 tion seems not to have been considered. Instead, 1a was imme-
undergo epimerization much more rapidly [28]. Although diately oxidized to give the phosphonoformate dinucleoside.
acylphosphines are pyramidal, the analogy to amides suggests
that the planar transition state for epimerization was greatly In this paper we report: (1) a more general synthesis of acyl
stabilized. Literature data from several groups suggests that the diamidites, which most likely accounts for Caruthers’ results,
inversion barrier is decreased to below 20 kcal/mol [28-30]. In including the X-ray crystal structure of the benzoyl analog,
order to extend the comparison from phosphines to phosphites, (2) the diamidite conversion to dinucleoside acylphosphonites,
we recently showed that dinucleoside phosphite triesters, like (3) conversion of these compounds to phosphorothioates,
phosphines, epimerize slowly – in this case at 150 °C with an (4) evidence that acylphosphonites, unlike their phosphine
inversion barrier of 33 kcal/mol [31]. Since the replacement of analogs, do not undergo facile inversion, and (5) confirmation
an alkyl or aryl group in phosphines by an acyl group lowered of this observation by density functional theory (DFT) calcula-
the epimerization barrier by >13 kcal/mol, it seemed reasonable tions.
to propose the same could happen by a similar replacement of
an alkoxy group in phosphite triesters.
Results and Discussion
Synthesis of acylphosphonodiamidites. Initially we attempted
to deprotonate H-phosphonodiamidite 4 with butyllithium, but
in all cases, the major reaction product with acetic anhydride
was simply the starting material. However, a minor product
with a peak in the 31P NMR spectrum near 63 ppm was
observed that was comparable to that exhibited by 6 at 52 ppm.
We repeated Caruthers’ method [46], which involved a two
Figure 1: Acyl phosphorus compounds.
185

Beilstein J. Org. Chem. 2015, 11, 184–191.
Scheme 1: Synthesis of a dinucleoside acylphosphonate (3b) and a formate diester (1a).
hour reflux of 4 with sodium, and similarly detected no visible at 227.9 ppm. The signal was observed as a doublet with 1JPC =
disappearance of the sodium. No reaction of this mixture with 22.4 Hz, in poor agreement with the calculated value of −60 Hz,
acetic anhydride was detected, beyond the occasional appear- although both CH3 calculations were in remarkably good agree-
ance of the 63 ppm peak. Removal of the solvent from the 4/Na ment (13C: 31 ppm, 2JPC = 54 Hz). Investigation of the poor
mixture, followed by reaction of the resultant crude material agreement with the one-bond coupling constant has not been
with acetyl chloride, resulted in variable amounts of the 63 ppm attempted, but the cause might be a phosphorus lone pair/car-
peak.
bonyl dihedral angle effect in which the true minimum energy
conformation has not been found [48-50]. Once purified, 7
Since there was no evidence that deprotonation of 4 gave rise to exhibited a strong carbonyl peak in the infrared spectrum at
the product, the direct reaction of 4 with acetyl chloride in the 1654 cm−1, a surprisingly low frequency considering the down-
absence of sodium was examined (Scheme 2). Surprisingly, an
immediate reaction occurred with acetyl chloride, giving a
yellow-orange solution accompanied by copious production of
what was assumed to be gaseous HCl. The hexane-soluble prod-
uct (7, Scheme 2) could be isolated as an orange oil exhibiting a
single peak in the 31P NMR spectrum at 63 ppm, but only in
64% yield. Variable amounts of iPr2NH2+Cl− also formed,
which presumably accounted for the low yield. The characteri-
zation included a doublet for the carbonyl methyl group in both
the 1H and 13C NMR spectra (1H: 2.27 ppm, 3JPH = 8.8 Hz;
13C: 30.7 ppm, 2JPC = 49.7 Hz). However, a peak for the car-
bonyl carbon could not be found nor could any literature prece-
dent be found. A DFT-NMR calculation using Gaussian [47]
was carried out (see Supporting Information File 1 for details),
with the surprising result that the carbonyl carbon peak was
expected at 242 ppm – that is, not at all analogous to an amide
chemical shift. Expansion of the standard sweep width beyond
Scheme 2: Reaction of an H-phosphonodiamidite with acid chlorides.
220 ppm indeed led to observation of the missing carbon peak,
186

Beilstein J. Org. Chem. 2015, 11, 184–191.
field 13C chemical shift. This inconsistency is perhaps due to an
effect caused by the electron-donor diisopropylamide groups.
A rational approach to prevent the formation of the HCl and
iPr2NH2+Cl− that accompany the formation of 7 would involve
the addition of a base. However, the use of Et3N gave a 43:57
mixture of 7 and a new byproduct, 8, with a 31P NMR peak at
49.7 ppm. Although efforts to isolate 8 have been unsuccessful,
the proposed structure (Scheme 2; see Supporting Information
File 1 for details) is consistent both with formation by base-
induced enolization followed by trapping with acetyl chloride,
and with the 1H, 13C, and DEPT NMR spectra. For instance, the
alkene-like hydrogen signals at 5.48 and 5.29 ppm have a
1.1 Hz coupling constant, which is typical of geminal alkyli-
dene hydrogen atoms. The difference between the 1.1 and
7.3 Hz phosphorus–hydrogen coupling constants is plausible for
a vinylphosphine, although the values themselves are low [48].
The NMR spectrum calculated using Gaussian was also in good
agreement with experimental observations (Scheme 2; also,
31P NMR: observed 50 ppm, calculated 52 ppm, and for com-
Figure 2: ORTEP [52] drawing of 9. Selected distances (Å) and angles
(°): P–N1 1.687(1), P–N2 1.679(1), P–C1 1.879(1), C1–O 1.221(1),
C1–C2 1.495(1), C1–P–N1 99.90(5), C1–P–N2 100.59(5), P–C1–O
120.23(8), P–C1–C2 119.78(8), C2–C1–O 119.82(9).
lone pair orbitals. However, as noted above, this does not result
in an amide-like 13C chemical shift.
parison, 7: observed 64 ppm, calculated 70 ppm), although once Nucleoside coupling. The initial attempts at coupling 7 and 9
again, the calculated one-bond coupling constant is in poor with 5′-O-DMTr-thymidine were carried out using the
agreement.
Caruthers procedure with 4,5-dicyanoimidazole (DCI) as the
activator. While formation of the product was observed, it was
The reaction of 4 with benzoyl chloride yielded a similar reac- not a clean reaction. The N-methylimidazolium triflate
tion, namely, copious bubbling of gaseous HCl with formation (NMI·Tf) method reported for the monosubstitution of
of a red-orange solution. In this case, unlike that with acetyl diamidites was found to be effective [53], resulting in relatively
chloride, two peaks were observed in the 31P NMR spectrum, clean formation of amidites 10 and 11 (Scheme 3). The initial
one of which appeared to be due to the desired product, 9, while reaction produced a yellow foam, which was assumed to be the
the other was far downfield at 121 ppm, perhaps due to color of the desired products due to its similarity to 7 and 9.
Ph(CO)P(Cl)N(iPr)2. The use of Et3N resulted in a clean reac- However, the presence of unreacted 5′-O-DMTr-thymidine, the
tion presumably since no enolization is possible, leading to 9 in starting acyls 7 and 9, and downfield 31P NMR peaks that are
89% yield after hexane extraction. A carbonyl peak was characteristic of diesters necessitated chromatographic sep-
observed in the 13C NMR spectrum, similar to that of 7, as a aration. The isolation of 11 as a yellow solid as a 42.5:57.5 mix-
doublet at 220.9 ppm (1JPC = 23.7 Hz), in addition to a strong ture of phosphorus epimers was carried out without difficulty.
carbonyl band in the infrared spectrum at 1631 cm−1.
Little fractionation of diastereomers occurred during multiple
chromatographies. In contrast, the isolation of 10 was some-
X-ray crystal structure of 9. Large crystals of 9 could be what tedious both due to the presence of impurities that were
obtained from hexane, allowing confirmation of the structure by not easily separated and continuous fractionation of the phos-
single-crystal X-ray diffraction [51]. An ORTEP drawing is phorus epimers that occurred during repeated chromatogra-
shown in Figure 2, illustrating the steric crowding around phos- phies. However, 10 was ultimately obtained as a white foam.
phorus due to the diisopropylamino groups. One result is that
the amines are essentially planar, where the sum of the angles Samples of the fast-moving diastereomer and of a mixture of
about each nitrogen atom is 360.0(1)°. However, the phos- the two diastereomers of 10 were fully characterized. They
phorus, as expected, is pyramidal with an N–P–N angle of exhibited the characteristic carbonyl doublets at 226.9 and
112.51(5)° and average N–P–C angles of 100.2°. The dihedral 226.8 ppm in the 13C NMR spectra (both with 1JPC = 25.3 Hz).
angle between the carbonyl and the putative phosphorus lone However, the infrared spectra could not be used to confirm the
pair is 108°, that is, only 18° from the angle required for the presence of the carbonyl since the thymidine ring exhibited a
lone pair and the carbonyl p-orbital to be parallel to each other. strong carbonyl absorption in the same region (i.e., 5′-O-DMTr-
The low CO stretching frequency in the infrared spectrum at thymidine at 1684 cm−1, see Supporting Information File 1).
1631 cm−1 is consistent with the overlap of the carbonyl and The absence of room-temperature epimerization was apparent
187

Beilstein J. Org. Chem. 2015, 11, 184–191.
Scheme 3: Synthesis of dinucleosides.
from both the chromatographic separation as well as the distinct mation of the diester impurities, and the use of excess phos-
peaks in the NMR spectra, and is well-known (as is facile phonamidite since both 10 and 11 could be chromatographi-
chromatographic separation) for the analogous β-cyanoethoxy cally removed from the dinucleosides. However, even a 50%
phosphoramidites [54,55]. A similar characterization of the excess was not sufficient to consume all of the 3′-O-(tert-butyl-
mixture of diastereomers of 11 also exhibited two sets of peaks dimethylsilyl)thymidine, and separation of this alcohol impu-
in the NMR spectra. However, one peculiarity involving only rity from the acyl dinucleosides could not be achieved.
the diisopropylamino moiety was noted, namely, a fluctional
process that resulted in broad peaks for the isopropylmethyl The characterization of the acyl dinucleosides was accom-
signals of both 10 and 11 in the 1H and 13C NMR spectra. Since plished by 1D and 2D 1H and 13C NMR spectra, with confirma-
this was not relevant to phosphorus epimerization, it was not tion of the dinucleoside structures by high-resolution positive
further investigated.
ion ESI mass spectrometry. The characteristic carbonyl doublets
shifted a few ppm upfield with larger P–C coupling constants
Dinucleoside coupling was carried out using the 2:1 pyridinium than those observed for 10 and 11 (223.28 and 223.25 ppm,
trifluoroacetate/N-methylimidazole (PTFA/NMI) activator 1JPC = 38.6 and 41.4 Hz for 12, 211.6 and 211.1 ppm,
developed at Isis to combine phosphonamidites 10 and 11 with 1JPC = 44.3 and 42.4 Hz for 13). Particularly, for 13, the two
3′-tert-butyldimethylsilyl-protected thymidine to give 12 and 13 deoxyribose hydrogen atom networks could be correlated with
[56]. The reactions were monitored by 31P NMR, giving fairly the two H1′ muliplets at 6.2–6.4 ppm in the 2D-COSY NMR
clean conversion from starting materials to the P-epimeric prod- spectrum. The identity of the two networks is based on the
ucts. The reaction of 10 to give 12 was significantly faster than assumption that the downfield H1′ multiplets are due to the
that of 11 to give 13. That is, while the former was nearly DMTr-thymidine, and the upfield H1′ multiplets are due to the
complete in 30 min, the latter was not complete even in 2.5 h. tert-butyldimethylsilyl thymidine, as is seen for the parent alco-
However, prolonged reaction led to formation of diester impuri- hols [31]. For 12, this analysis could not be carried out with as
ties that could not be chromatographically separated. While not much certainty due to the larger amount of the 3′-O-tert-butyl-
identified, the impurities could be due to detritylation although dimethylsilyl thymidine impurity. As previously observed for
the characteristic orange color of the trityl cation was not seen. the β-cyanoethoxy dithymidyl triesters [31], the characteristic
The diesters were less robust than the precursor monoester two pairs of thymidyl methyl doublets (due to allylic coupling
amidites. For instance, rapid chromatographic separation of the to the H6 alkene hydrogen) on the two diastereomers were seen,
products from the ammonium salts formed during the synthesis one pair downfield near 1.8 ppm for the tert-butyldimethylsilyl-
was required to prevent further decomposition. The best proce- thymidine, and one pair upfield near 1.4 ppm for the DMTr-
dure was found to involve short reaction times to minimize for- thymidine.
188

Beilstein J. Org. Chem. 2015, 11, 184–191.
Conversion to phosphorothioates. An additional confirmation tion. Not surprisingly, the phosphine transition state exhibited a
of the structures was possible via oxidation and conversion to trigonal planar geometry, and the gas-phase energy barrier of
the known diastereomeric phosphorothioates. As shown by Hata 32.5 kcal/mol was in excellent agreement with the values
et al., acylphosphonate 3b (Scheme 1) was converted to the reported by Mislow for reactions carried out in non-polar
corresponding phosphorothioate by treatment with a base (to toluene as a solvent. Similarly, the acylphosphine transition
give the H-phosphonate via loss of carboxylate) followed by state exhibited a trigonal planar geometry, and the gas-phase
elemental sulfur [43]. As shown in Scheme 3, oxidation of 12 energy barrier of 21.2 kcal/mol was also in excellent agreement
and 13 with anhydrous tert-butyl hydroperoxide gave the with the typical values noted above. The phosphite, in contrast,
acylphosphonates 14 and 15. These products were immediately did not yield a trigonal planar transition state, but rather gave a
hydrolyzed by addition of approximately 2 equiv of aqueous T-shaped transition state. In fact, such a result has been reported
triethylammonium bicarbonate (TEAB) to give the H-phospho- for PF3 [59,60] and is due to the presence of the highly elec-
nate 16. Treatment of 16 with approximately 4 equiv of phenyl- tronegative atoms bound to phosphorus. While the alkoxy
acetyl disulfide (PADS) [57] gave the known dinucleoside groups of the phosphite triester are considered to be π-electron
phosphorothioate 17 [26,58].
donors, they are strongly electronegative. This results in an
inversion geometry that is rather different from the phosphines,
Phosphorus epimerization. During the many attempts to presumably mediated by oxygen–phosphorus lone pair–lone
purify the acyl dinucleosides by chromatography, it became pair repulsion. The calculated barrier of 41.3 kcal/mol (from the
apparent that some fractionation of the diastereomeric mixtures slightly higher ground state geometry) was not in good agree-
occurred. Due to the sensitivity of the acyls, chromatography on ment with that reported, being 8.3 kcal/mol higher. Unlike the
long columns of silica (with or without 1% NEt3 to deactivate other calculations, this one incorporated the polar acetonitrile
the silica) gave rise to the loss of acyls, without any improve- solvent that was used for the epimerization kinetics [31], but it
ment in diastereomer separation. However, the individual had essentially no effect on the calculated barrier.
samples of fractionated diastereomers of 12 or 13 showed no
evidence of room-temperature epimerization. Heating of
samples at temperatures from 50 to 150 °C over several hours
was carried out in acetonitrile, as was done for the epimeric
phosphite triesters [31]. Although this was not the most rigorous
test (since available samples only ranged in diastereomer ratios
from 1:1 to 1:1.4), no detectable change in the diastereomer
ratios was observed. More importantly, sample decomposition
occurred at all temperatures. Because of this, it would have
been impossible to distinguish between small changes in dia-
stereomer ratio and small differences in rates of diastereomer
decomposition.
Calculation of acylphosphonite inversion barrier. As
described in the Introduction, inversion barriers for phosphines,
acylphosphines, and phosphites are known. These were calcu-
lated first for validation of the method. The structures of chiral
phosphine 18, chiral acylphosphine 19, and chiral phosphite 20
(Scheme 4) were each minimized by DFT using Gaussian (see
Scheme 4: Calculated phosphine, acylphosphine, phosphite, and
acylphosphonite inversion barriers.
Supporting Information File 1 for details), and the vibrational
calculation confirmed a local minimum for each structure.
While the phosphites are complicated by multiple local minima,
which depend on the conformations about the P–O bonds, the The transition state calculations were then carried out on the
energies are comparable (separated by less than 1 kcal/mol) and chiral acylphosphonite 21 with the expectation that a high
do not have a significant impact on the inversion barriers. Tran- phosphite-like barrier might occur. In fact, the barrier of
sition states were located using routines implemented in 41.7 kcal/mol was slightly higher than the phosphite triester
Gaussian (see Supporting Information File 1 for details), and in calculation. Unlike the phosphite T-shaped transition state, this
all cases the vibrational calculation exhibited one negative occurs via a trigonal planar transition state, although like the
frequency that corresponded to the phosphorus inversion vibra- others with one negative frequency corresponding to phos-
189

Beilstein J. Org. Chem. 2015, 11, 184–191.
phorus inversion. The high barrier is consistent with the absence Acknowledgements
of acylphosphonite epimerization of 12 and 13, and this Mass spectra were obtained at the North Carolina State Univer-
suggests that any amide-like stabilization of the phosphorus sity Mass Spectrometry Facility located in the Department of
lone pair in a trigonal transition state is eliminated by Chemistry. Financial support from the City University of New
oxygen–phosphorus lone pair–lone pair repulsion.
York PSC-CUNY Research Award Program is gratefully
acknowledged.
Conclusion
References
High-yield syntheses have been developed for some novel phos-
phorus synthons, including the acyl and benzoyldiamidites via
the surprising condensation of phosphonodiamidite 4 and the
two acid chlorides with elimination of HCl. The nucleoside
acylphosphonamidites were also readily prepared, although the
benzoyl analog 11 was more easily isolated (albeit as the
diastereomeric mixture of P-epimers), while the acetyl analog
10 allowed the diastereomers to be separated following repeated
chromatographies. Final coupling to give the dinucleoside
acylphosphonites proceeded smoothly, but separation of the
pure compounds from the nucleoside precursors could not be
completely achieved. If the pure acylphosphonites are to be
obtained, solid-phase synthesis would probably be required.
The chromatography results suggested that room-temperature
epimerization at phosphorus did not occur, and heating at
50–150 °C also gave no indication of epimerization. Density
functional theory calculations are consistent with this result,
giving a calculated inversion barrier of 42 kcal/mol. While one
might argue that given the results of such a calculation, the lack
of epimerization was a foregone conclusion, it is still important
to test theory with experiment. In conclusion, the analogy
between phosphines and phosphites, in which the low barrier to
acylphosphine inversion suggested that the same might occur
with acylphosphonites, is in this case superficial, and the antici-
pated low-temperature inversion does not occur.
1. Matsukura, M.; Shinozuka, K.; Zon, G.; Mitsuya, H.; Reitz, M.;
Cohen, J. S.; Broder, S. Proc. Natl. Acad. Sci. U. S. A. 1987, 84,
7706–7710. doi:10.1073/pnas.84.21.7706
2. Uhlmann, E.; Peyman, A. Chem. Rev. 1990, 90, 543–584.
doi:10.1021/cr00102a001
3. Crooke, S. T.; Geary, R. S. Br. J. Clin. Pharmacol. 2013, 76, 269–276.
doi:10.1111/j.1365-2125.2012.04469.x
4. Lee, J. J. A.; Yokota, T. J. Pers. Med. 2013, 3, 144–176.
doi:10.3390/jpm3030144
5. Šípová, H.; Špringer, T.; Rejman, D.; Šimák, O.; Petrová, M.;
Novák, P.; Rosenbergová, S.; Páv, O.; Liboska, R.; Barvík, I.;
Štěpánek, J.; Rosenberg, I.; Homola, J. Nucleic Acids Res. 2014, 42,
5378–5389. doi:10.1093/nar/gku125
6. Prakash, T. P.; Graham, M. J.; Yu, J.; Carty, R.; Low, A.; Chappell, A.;
Schmidt, K.; Zhao, C.; Aghajan, M.; Murray, H. F.; Riney, S.;
Booten, S. L.; Murray, S. F.; Gaus, H.; Crosby, J.; Lima, W. F.; Guo, S.;
Monia, B. P.; Swayze, E. E.; Seth, P. P. Nucleic Acids Res. 2014, 42,
8796–8807. doi:10.1093/nar/gku531
7. Magen, I.; Hornstein, E. Brain Res. 2014, 1584, 116–128.
doi:10.1016/j.brainres.2014.04.005
8. Marwick, C. JAMA 1998, 280, 871.
doi:10.1001/jama.280.10.871-JMN0909-6-1
9. Jiang, K. Nat. Med. 2013, 19, 252. doi:10.1038/nm0313-252
10.Marshall, W. S.; Caruthers, M. H. Science 1993, 259, 1564–1570.
doi:10.1126/science.7681216
11.Li, P.; Sergueeva, Z. A.; Dobrikov, M.; Shaw, B. R. Chem. Rev. 2007,
107, 4746–4796. doi:10.1021/cr050009p
12.Uehara, S.; Hiura, S.; Higashida, R.; Oka, N.; Wada, T. J. Org. Chem.
2014, 79, 3465–3472. doi:10.1021/jo500185b
13.Stec, W. J.; Grajkowski, A.; Kobylanska, A.; Karwowski, B.;
Koziolkiewicz, M.; Misiura, K.; Okruszek, A.; Wilk, A.; Guga, P.;
Boczkowska, M. J. Am. Chem. Soc. 1995, 117, 12019–12029.
doi:10.1021/ja00154a001
Supporting Information
Supporting Information contains the experimental
procedures, characterization of new compounds, copies of
1H, 13C and 31P NMR spectra, IR spectra, details of the
X-ray structure determination, and details of the DFT
calculations.
14.Stec, W. J.; Karwowski, B.; Boczkowska, M.; Guga, P.;
Koziołkiewicz, M.; Sochacki, M.; Wieczorek, M. W.; Blaszczyk, J.
J. Am. Chem. Soc. 1998, 120, 7156–7167. doi:10.1021/ja973801j
15.Wang, J.-C.; Just, G. J. Org. Chem. 1999, 64, 8090–8097.
doi:10.1021/jo990242l
16.Wilk, A.; Grajkowski, A.; Phillips, L. R.; Beaucage, S. L.
J. Am. Chem. Soc. 2000, 122, 2149–2156. doi:10.1021/ja991773u
17.Yu, D.; Kandimalla, E. R.; Roskey, A.; Zhao, Q.; Chen, L.; Chen, J.;
Agrawal, S. Bioorg. Med. Chem. 2000, 8, 275–284.
doi:10.1016/S0968-0896(99)00275-8
Supporting Information File 1
Experimental.
[http://www.beilstein-journals.org/bjoc/content/
supplementary/1860-5397-11-19-S1.pdf]
18.Almer, H.; Szabo, T.; Stawinski, J. Chem. Commun. 2004, 290–291.
doi:10.1039/b311912b
Supporting Information File 2
CIF file of compound 9.
19.Oka, N.; Yamamoto, M.; Sato, T.; Wada, T. J. Am. Chem. Soc. 2008,
130, 16031–16037. doi:10.1021/ja805780u
[http://www.beilstein-journals.org/bjoc/content/
supplementary/1860-5397-11-19-S2.cif]
20.Lu, Y. Mini-Rev. Med. Chem. 2006, 6, 319–330.
doi:10.2174/138955706776073439
190

Beilstein J. Org. Chem. 2015, 11, 184–191.
21.Guga, P. Curr. Top. Med. Chem. 2007, 7, 695–713.
doi:10.2174/156802607780487786
50.Hersh, W. H.; Lam, S. T.; Moskovic, D. J.; Panagiotakis, A. J.
J. Org. Chem. 2012, 77, 4968–4979. doi:10.1021/jo3003776
51.Cambridge Crystallographic Data Centre, deposition number CCDC
1030743.
22.Oka, N.; Wada, T. Chem. Soc. Rev. 2011, 40, 5829–5843.
doi:10.1039/c1cs15102a
23.Guga, P.; Koziołkiewicz, M. Chem. Biodiversity 2011, 8, 1642–1681.
doi:10.1002/cbdv.201100130
52.Farrugia, L. J. J. Appl. Crystallogr. 1997, 30, 565.
doi:10.1107/S0021889897003117
24.Nukaga, Y.; Yamada, K.; Ogata, T.; Oka, N.; Wada, T. J. Org. Chem.
2012, 77, 7913–7922. doi:10.1021/jo301052v
53.Xie, C.; Staszak, M. A.; Quatroche, J. T.; Sturgill, C. D.; Khau, V. V.;
Martinelli, M. J. Org. Process Res. Dev. 2005, 9, 730–737.
doi:10.1021/op050077d
25.Mukhlall, J. A.; Noll, B. C.; Hersh, W. H. J. Sulfur Chem. 2011, 32,
199–212. doi:10.1080/17415993.2011.580346
54.Stec, W. J.; Zon, G. Tetrahedron Lett. 1984, 25, 5279–5282.
doi:10.1016/S0040-4039(01)81583-5
26.Mukhlall, J. A.; Hersh, W. H. Nucleosides, Nucleotides Nucleic Acids
2011, 30, 706–725. doi:10.1080/15257770.2011.597366
27.Baechler, R. D.; Mislow, K. J. Am. Chem. Soc. 1970, 92, 3090–3093.
doi:10.1021/ja00713a028
55.Cheruvallath, Z. S.; Sasmor, H.; Cole, D. L.; Ravikumar, V. T.
Nucleosides, Nucleotides Nucleic Acids 2000, 19, 533–543.
doi:10.1080/15257770008035005
28.Egan, W.; Mislow, K. J. Am. Chem. Soc. 1971, 93, 1805–1806.
doi:10.1021/ja00736a053
56.Eleuteri, A.; Capaldi, D. C.; Krotz, A. H.; Cole, D. L.; Ravikumar, V. T.
Org. Process Res. Dev. 2000, 4, 182–189. doi:10.1021/op9900378
57.Cheruvallath, Z. S.; Carty, R. L.; Moore, M. N.; Capaldi, D. C.;
Krotz, A. H.; Wheeler, P. D.; Turney, B. J.; Craig, S. R.; Gaus, H. J.;
Scozzari, A. N.; Cole, D. L.; Ravikumar, V. T. Org. Process Res. Dev.
2000, 4, 199–204. doi:10.1021/op990077b
29.Kostyanovskii, R. G.; Fomichev, A. A.; Zagurskaya, L. M.;
Zakharov, K. S. Izv. Akad. Nauk SSSR, Ser. Khim. 1973, 1915–1916.
30.Vedejs, E.; Donde, Y. J. Am. Chem. Soc. 1997, 119, 9293–9294.
doi:10.1021/ja970369x
31.Mukhlall, J. A.; Hersh, W. H. Inorg. Chim. Acta 2011, 369, 62–70.
doi:10.1016/j.ica.2010.11.022
58.Bartoszewicz, A.; Kalek, M.; Stawinski, J. J. Org. Chem. 2008, 73,
5029–5038. doi:10.1021/jo8006072
32.McKenna, C. E.; Kashemirov, B. A. Top. Curr. Chem. 2002, 220,
201–238. doi:10.1007/3-540-45731-3_8
59.Dixon, D. A.; Arduengo, A. J.; Fukunaga, T. J. Am. Chem. Soc. 1986,
108, 2461–2462. doi:10.1021/ja00269a063
33.Evans, D. A.; Johnson, J. S. J. Am. Chem. Soc. 1998, 120,
4895–4896. doi:10.1021/ja980423p
60.Gilheany, D. G. Chem. Rev. 1994, 94, 1339–1374.
doi:10.1021/cr00029a008
34.Jang, K. P.; Hutson, G. E.; Johnston, R. C.; McCusker, E. O.;
Cheong, P. H.-Y.; Scheidt, K. A. J. Am. Chem. Soc. 2014, 136, 76–79.
doi:10.1021/ja410932t
License and Terms
35.Wen, Y.-Q.; Hertzberg, R.; Moberg, C. J. Org. Chem. 2014, 79,
6172–6178. doi:10.1021/jo500895u
This is an Open Access article under the terms of the
Creative Commons Attribution License
36.Demir, A. S.; Esiringü, l.; Göllü, M.; Reis, Ö. J. Org. Chem. 2009, 74,
2197–2199. doi:10.1021/jo8026627
(http://creativecommons.org/licenses/by/2.0), which
permits unrestricted use, distribution, and reproduction in
any medium, provided the original work is properly cited.
37.Jiang, H.; Paixão, M. W.; Monge, D.; Jørgensen, K. A.
J. Am. Chem. Soc. 2010, 132, 2775–2783. doi:10.1021/ja9097803
38.Kluger, R.; Pike, D. C. J. Am. Chem. Soc. 1977, 99, 4504–4506.
doi:10.1021/ja00455a052
39.Breuer, E.; Safadi, M.; Chorev, M.; Gibson, D. J. Org. Chem. 1990, 55,
6147–6153. doi:10.1021/jo00312a021
The license is subject to the Beilstein Journal of Organic
Chemistry terms and conditions:
40.Berlin, K. D.; Taylor, H. A. J. Am. Chem. Soc. 1964, 86, 3862–3866.
doi:10.1021/ja01072a053
(http://www.beilstein-journals.org/bjoc)
41.Sekine, M.; Satoh, M.; Yamagata, H.; Hata, T. J. Org. Chem. 1980, 45,
4162–4167. doi:10.1021/jo01309a019
The definitive version of this article is the electronic one
which can be found at:
42.Kume, A.; Fujii, M.; Sekine, M.; Hata, T. J. Org. Chem. 1984, 49,
2139–2143. doi:10.1021/jo00186a012
doi:10.3762/bjoc.11.19
43.Fujii, M.; Ozaki, K.; Sekine, M.; Hata, T. Tetrahedron 1987, 43,
3395–3407. doi:10.1016/S0040-4020(01)81630-6
44.Krenske, E. H. J. Org. Chem. 2012, 77, 1–4. doi:10.1021/jo202370x
45.Krenske, E. H. J. Org. Chem. 2012, 77, 3969–3977.
doi:10.1021/jo300346g
46.Yamada, C. M.; Dellinger, D. J.; Caruthers, M. H. J. Am. Chem. Soc.
2006, 128, 5251–5261. doi:10.1021/ja060112b
47.Gaussian 09, Revision D.01; Gaussian, Inc.: Wallingford, CT, 2009.
48.Fedorov, S. V.; Krivdin, L. B.; Rusakov, Y. Yu.; Chernysheva, N. A.;
Mikhailenko, V. L. Magn. Reson. Chem. 2010, 48, S48–S55.
doi:10.1002/mrc.2614
49.Fedorov, S. V.; Krivdin, L. B.; Rusakov, Y. Yu.; Ushakov, I. A.;
Istomina, N. V.; Belogorlova, N. A.; Malysheva, S. F.; Gusarova, N. K.;
Trofimov, B. A. Magn. Reson. Chem. 2009, 47, 288–299.
doi:10.1002/mrc.2386
191