Lowering Solid Support Costs for Oligonucleotide Synthesis

1
Lowering Solid Support Costs for Oligonucleotide
Synthesis
Marc L. Rothstein and Dianne M. Rothstein
Prime Synthesis, Inc. Aston, PA 19014

Results
Objective
Once the best low-cost CNA linker was identified
by the screening test, it was compared to the
more expensive benchmark process in a synthesis
test. For each support, a number of 21-mer
deoxy-oligo nucleotides were synthesized at
Integrated DNA technologies, using their
proprietary synthesis instrumentation, developed
to operate at maximum reagent efficiency. The
oligos were analyzed by CE and MALDI for full
length product yield, oligo purity, n-1
purity, and average coupling efficiency
Table 1 Comparison of Benchmark and
Low Cost CPG Performance
Background
Product ID Coupling Oligo Purity n-1 yield (full length)
N16 (Benchmark) 99.19 84.71 1.12 58.38
N12 (Low Cost) 99.21 85.66 1.44 55.05
N/A (Competitor) 99.46 86.73 2.78 44.50
The attributes of CPG as a solid support include
high specific surface area, dimensional stability
in organic solvents, large and stable pore spaces
and high rigidity. Most reaction schemes used to
link nucleoside to the CPG surface, begin by
coupling a bi-functional organosilane to the
native silanol groups of the glass surface. (Fig.
1) With proper immobilization techniques,
nucleoside loadings of over 100 (µm/g) can be
obtained on a 500 Å pore size support. Under
proper deposition conditions, a uniform
distribution of reactive groups can be achieved,
which will result in higher levels of ligand
accessibility. Together with the preferred pore
structure, this accounts for the continued
popularity of CPG in such difficult syntheses as
RNAis. Since the silanol groups of the CPG are
capable of reacting with nucleoside
phosphoramidites, it is important to eliminate
unreacted surface silanols on the finished
nucleoside support, thus minimizing the
occurrence of an oligo impurity lacking only the
intended 3 base, (n-1). In one study, NMR
Spectroscopy was used to detect and characterize
residual silanols on the surface of chemically
modified silica (2) . Silanol groups were
classified as free or associated. The
associated forms (Fig. 2) consisted of two
adjacent silanols with a hydrogen bound to each
other (vicinal) or a silanol with two hydroxyls
sitting on the same atom (geminal). Another
study (3) concluded that associated silanols were
reactive to trimethylchlorosilane (TCMS), while
free silanols were more reactive towards
hexamethyldisilazane (HMDS). Preliminary work
with aminopropyl linkers on CPG showed two
silanization approaches could provide complete
coverage of all silanol groups 1) repetitious
silanization (monolayer conditions) and 2) single
silanization, followed by a dual silanol capping
treatment with TMCS and HMDS. In a similar study
CPG was functionalized with a proprietary long
chain alkykamine linker (CNA) developed at
Ribozyme Pharmaceuticals. CNA silane was found
to be highly reactive towards silanol groups.
From this, a high performance LCAA-linked
nucleoside CPG process was developed, that
utilized both the novel linker AND capping
silanes. This product gave the lowest n-1
impurities of any CPG tested to date, but was
costly to produce. With this performance as a
benchmark, a more cost-effective process was
developed, utilizing the CNA precursor silane to
its BEST advantage.
Methods
In order to quickly evaluate the various silanol
reduction schemes, a rapid screening test was
developed. Once screened, the best performing
supports were evaluated in actual oligonucloetide
synthesis tests. The screening test (Residual
Silanol Test) was based on the fact that
nucleoside phosphoramidites can couple to the
hydroxyl group of a silanol with low efficiency,
but a nucleoside succinate cannot.
The test scheme is shown below
Figure 4 Representative CE Test Results
Amino CPG is saturated with the nucleoside
succinate base 1 in an exhaustive reaction
Mono-nucleoside loaded CPG is saturated with the
nucleoside phosphoramidite base 2 in an
exhaustive reaction
(Loading 2 Loading 1) Incremental Loading


i.e Amount of nucleoside phosphoramidite base 2
that reacted with RESIDUAL SILANOL groups on CPG
INCREASE in Incremental Loading to
actual n-1 impurity (Figure 3)

• Residual Silanol Test Confirmation
• Used two CPG samples of different silanol
  reduction schemes
• Synthesized 18-mer deoxy oligos with AKTA (GE
  Healthcare) synthesizer
• 1mM scale
• Analysis of n-1 related impurities with CE and
  MALDI
• Two extraneous impurities explained
• n-3 phosphorothiolated impurity seen by CE was
  actually n-1 with extra charged species
• Impurity with MW corresponding to
  (n-1)94 phosphorothiolated impurity of n-1
  oligo

Figure 3 Correlation between Residual Silanol
Test and Actual n-1 Impurities
Summary
The objective of lowering the cost of a linker
CPG, without sacrificing its synthesis
performance was achieved.
The "Low Cost" production protocol cost 30 less
than the corresponding "Benchmark" protocol.
In addition to providing a rapid screening tool
in this study, a useful QC test for the
evaluation of residual silanol levels resulted
from this project.

• An independent check on residual amine groups
  showed complete amine capping

Acknowledgements
References
We would like to thank Dr. Nanda Sinha of Avecia
Biotechnology for his helpful


suggestions and evaluations in this study. We
would also like to thank Trey Martin, Mike Marvin
and Dr. Yakov Leutchy for their evaluations in
this work.
1. Wincott, F.E., Further Improvement to
the Large Scale Production of Ribozymes , Tides
2000 Presentation 2. Holik, M. and
Matejkova, B. (1981) J. Chromatography, Vol. 213
p. 33. 3. Snyder, I.R., Principles of
Adsorption Chromatography, M. Decker, New York,
1968.