Research Article Single-Round Patterned DNA Library...

Research ArticleSingle-Round Patterned DNA Library MicroarrayAptamer Lead Identification

Jennifer A Martin12 Peter A Mirau3 Yaroslav Chushak12 Jorge L Chaacutevez14

Rajesh R Naik3 Joshua A Hagen1 and Nancy Kelley-Loughnane1

1Human Effectiveness Directorate 711 Human Performance Wing Air Force Research Laboratory Wright-Patterson Air Force BaseDayton OH 45433 USA2The Henry M Jackson Foundation for the Advancement of Military Medicine 6720A Rockledge Drive Bethesda MD 20817 USA3Materials and Manufacturing Directorate Air Force Research Laboratory Wright-Patterson Air Force Base OH 45433 USA4UES Inc 4401 Dayton-Xenia Road Dayton Dayton OH 45433 USA

Correspondence should be addressed to Nancy Kelley-Loughnane nancykelley-loughnane1usafmil

Received 12 November 2014 Revised 22 April 2015 Accepted 27 April 2015

Academic Editor Chih-Ching Huang

Copyright copy 2015 Jennifer A Martin et al This is an open access article distributed under the Creative Commons AttributionLicense which permits unrestricted use distribution and reproduction in any medium provided the original work is properlycited

A method for identifying an aptamer in a single round was developed using custom DNAmicroarrays containing computationallyderived patterned libraries incorporating no information on the sequences of previously reported thrombin binding aptamersTheDNA library was specifically designed to increase the probability of binding by enhancing structural complexity in a sequence-space confined environment much like generating lead compounds in a combinatorial drug screening library The sequencedemonstrating the highest fluorescence intensity upon target addition was confirmed to bind the target molecule thrombin withspecificity by surface plasmon resonance and a novel imino protonNMR2DNOESY combination was used to screen the structurefor G-quartet formationWe propose that the lack of G-quartet structure inmicroarray-derived aptamers may highlight differencesin binding mechanisms between surface-immobilized and solution based strategies This proof-of-principle study highlights theuse of a computational driven methodology to create a DNA library rather than a SELEX based approachThis work is beneficial tothe biosensor field where aptamers selected by solution based evolution have proven challenging to retain binding function whenimmobilized on a surface

1 Introduction

Aptamers are oligonucleotide molecular recognition ele-ments selected through a synthetic iterative evolutionaryprocess termed SELEX (Systematic Evolution of Ligandsby EXponential enrichment) [1 2] Since their discoveryaptamers have been selected for a variety of targets fromions to whole cells and implemented in applications such astherapeutics purification or as biosensor detection ligandsThe use of aptamers for these applications over other typesof recognition elements is warranted by their well-reportedadvantages including the ease and reproducibility of chem-ical synthesis simplicity of modifying aptamers with fluo-rescent tags or surface immobilization chemistries (aminebiotin etc) and relatively high stability to degradation [3]

Despite the advantages of aptamers several significantdrawbacks are inherent to the standard SELEX processOne example is the timescale of aptamer selection whichtypically requires an average of 12 cycles and a minimumof 2ndash6 months not including initial optimization processesvalidation of aptamer candidates or structural analysis [4 5]This is typically a result of the low partitioning efficiency (theability to separate binding sequences from nonbinders in aselection round) of conventional partitioning methods usedin SELEX [6] Furthermore SELEX suffers from polymerasechain reaction (PCR) bias where the PCR has been reportedto amplify oligonucleotides unequally resulting in an inac-curacy of comparative representation within a pool as theselection progresses [7] A second form of bias is introducedby the cloning and Sanger sequencing method used for

Hindawi Publishing CorporationJournal of Analytical Methods in ChemistryVolume 2015 Article ID 137489 8 pageshttpdxdoiorg1011552015137489

2 Journal of Analytical Methods in Chemistry

aptamer identification The sequences reported likely reflectthe most abundant sequences present but may not reportthose that have artificially lower numbers due to factors suchas PCR bias or cloning efficiency or due to low sampling ofthe entire sequence space (diversity) of the final pool [8]

Several groups have proposed using DNAmicroarrays toaddress the possible SELEX biases and expedite the aptameridentification process [9ndash11] DNA microarrays function byidentifying locations of fluorescently labeled targets andcorrelating themwith the position of known sequences cova-lently synthesized on the arrayThe arrays are fully customiz-able so the user can define the exact sequences of interest pro-duced on the arrays A single microarray experiment can becompleted in less than one day with an additional 1-2 days fordata analysisThis characteristic is a function of the improvedpartitioning efficiency available by covalently linking thesequences to the surface Higher stringency conditions canbe applied to identify sequences with more ideal bindingproperties PCR cloning and sequencing are not requiredsince known sequences are in predefined locations Alsomicroarrays are particularly useful for identifying aptamersfor biosensor applications since the response of an aptamerselected by solution-based SELEX may be significantlydiminished when it is tethered to a sensor surface [12ndash14]

A major drawback of microarray use is that the highestdensity arrays have a maximum of sim106 sequences incontrast to SELEX methods which evolve from an initiallibrary of sim1015 sequences However combinatorial drug-screening libraries successfully identify binders with only103ndash106 different compounds in the starting library due tothe diversity of the functional groups of the compounds[15] Extending this premise to oligonucleotides it has beendetermined that the probability for a sequence to bind atarget improves with increasing structural complexity [16]This means unstructured sequences or oligonucleotidesthat form simple structures such as those in a randomoligonucleotide library have reduced potential to show anytype of function Constituents of the random pool consist ofmostly unpaired regions combined with short (low stability)stem-loop structures and the probability of containing anabundance of more complex high-affinity aptamers in thestarting random library is low [17] Several groups haveexplored this issue experimentally by optimizing the SELEXstarting library to contain more complex partially structuredsequences [18ndash20] These studies showed that the partiallystructured libraries provided more sequences that bindthe target andor sequences with higher binding affinitiescompared to a completely random starting library

The same principle can be broadened to biasing a mi-croarray starting library to contain sequences with increasedstructural complexity and thus enhanced potential for targetbinding In previous works [9ndash11] starting libraries consistedof 102ndash104 initial sequences and evolved aptamers throughin silico genetic algorithms in multiple chip generations(rounds) These studies used either naturally fluorescenttargets utilized a target with well-studied aptamer bindingmotifs or took into account the characteristics of knownaptamers for library design

In this work we applied DNA library patterning in orderto circumvent the microarray density problem and rapidlyidentified an aptamer to the target molecule thrombin ina single round employing a pattern which considered nostructural information from previously reported thrombinaptamers We show that this aptamer does not form thewell-known G-quartet structure reported in early iterationsof thrombin aptamers by utilizing a combination of iminoproton NMR and 2D NMR for structural characterizationthat is simpler than legacy methods of establishing sequence-specific assignments The NMR results also raise questionson whether the aptamer identification platform (surface-immobilized or solution-based) may significantly influencethe binding mechanism of the final aptamer These resultsgenerally demonstrate a promisingmethod for rapidly identi-fying and characterizing aptamers whichmay directly benefitthe field of aptamer biosensors where immobilization ofsolution-selected aptamers on a sensing platform has provento be challenging

2 Materials and Methods

21 Chemicals and Equipment The chemicals used were IgE(Fitzgerald) Thrombin (Haematologic Technologies Inc)BSA and HSA (Sigma) neuropeptide Y (Pheonix Pharma-ceuticals) IllustraNAP-25 desalting columns andCy3Mono-Reactive Dye Pack (GEHealthcare) NanoDrop (Thermo Sci-entific) nuclease free water (Gibco) Microarray equipmentconsisted of the following custom 8times15 k DNAmicroarrays8 times 15 k gasket slides ozone barrier slides hybridizationchambers scanner cassettes hybridization oven and High-resolution Microarray Scanner (all Agilent) and slide rackand wash dishes (Shandon) and Kimtech polypropylenewipes (Kimberly-Clark) All DNA was purchased throughIDT

4A018 GGTTGGTTTTTCAATCAGCGATCGCG-GAATCCAGGGTTAGGCGGCCAACC (with andwithout 31015840-T

10-Biotin moiety)

TFBS GGTTGGTGTGGTTGG

Buffers Binding [PBSMTB]-1x PBS (81mMNa2HPO4

11 mM KH2PO4 27mM KCl 137mM NaCl pH

74) + 1mM MgCl2+ 01 Tween-20 and 1 BSA

Washing [PBSM]-1x PBS (81mM Na2HPO4 11mM

KH2PO4 27mMKCl 137mMNaCl pH 74) + 1mM

MgCl2 Rinse [R]-14 dilution of PBSM and nuclease

free water

22 Protein Labeling Thrombin was labeled with Cy3 using aCy3 Mono-Reactive Dye Pack (GE Healthcare) Protein wasdiluted to 1mgmL in 1mL 01M Na

2CO3buffer (pH = 93)

and incubated for 30 minutes The product was purified withTexas Red protein labeling size exclusion column (MolecularProbes) and the dye-to-protein ratio (DP) was calculated as08DP using the manufacturerrsquos instructions with UVVisdetection

Journal of Analytical Methods in Chemistry 3

23 Microarray Starting Library Design UNAFold softwarewas used to screen DNA sequences from the patternedlibrary generated using Perl scripts and to determine whichsequences were folding according to a predefined set ofcriteria Constraints were set to maximize the number ofpotential binders (1) 1st base should be paired with the50th base (2) the total number of unpaired bases 10 ltunpairedlt 30 (3) there should be at least two 4-unpaired basestretches The secondary structure of generated sequenceswas evaluated using the UNAFold package with the followingset of parameters 119879 = 25∘C [Na+] = 100mM and [Mg2+] =5mMThese settings represent generalized conditions relatedto aptamer studies encompassing a variety of buffers andapplications Only sequences with a secondary structurethat passed these selection criteria were candidates for themicroarray selection experiments Sequences were analyzedat random until 50000 sequences were reported to fit thecriteria Five thousand out of 50000 total sequences that fitthe constraints were randomly incorporated onto the 8times15 kmicroarray chipwith a 31015840-T

10spacer in duplicate or triplicate

Controls were synthesized with a minimum of 10 replicatesalso containing a 31015840-T

10spacer

24 Microarray Procedure Blocking with PBSMTB was per-formed on the DNA microarray loaded into the gasket for1 h at room temperatureThe slides were quickly edge-tappedto remove excess buffer Seventy 120583L Cy3-thrombin (100 nM)in PBSMTB was loaded onto a gasket slide then incubatedwith the array for 2 hrs at 20∘C in a hybridization cham-berhybridization oven Slides were quickly disassembled inwater and washed for 3min in a PBSM buffer with the sliderack and stir plate then transferred to a 50mL conical tubewith 14 PBSM bufferwater for 1min using a shaker plateSlides in the slide rack were then dipped in a 50mL conicaltube of nuclease-free water to remove any remaining salt andwashed for 1min on a shaker plate The microarray slidewas slowly withdrawn from the water to promote a driersurface the back of the slide was wiped with ethanol andthen placed in a 50mL conical tube with a polypropylenewipe at the bottom and centrifuged at 4150 rpm for 3minThemicroarray was loaded into a scanner cassette and coveredwith an ozone barrier slide before scanning The arrays werescanned using Agilent Scan Control software Images (TIFF)were generated using 20-bit imaging at 5120583m (8times15 k arrays)Data was extracted using Agilent Feature Extraction softwareversion 10731 Mean fluorescence intensity and standarddeviation of replicates were determined using code written inPerl Statistical significance was calculated using a two-tailed119905-test at 95 confidence interval for potential binders identi-fied through the microarray compared to control sequencesStatistical significance was used as a metric for differentiatingreal binding events from experimental artifacts

25 Surface Plasmon Resonance (SPR) SPR Biacore stud-ies were carried out on a CM7 (GE Healthcare) chipwith neutravidin custom immobilized on the surfaceThe surface was activated with a mixture of 1-ethyl-3-[3-dimethylaminopropyl]carbodiimide (02M) andN-hydroxy-

succinimide (005M) for 420 sec at 10 120583Lmin Neutravidin(10mgmL Thermo Scientific) was dissolved in HyClonewater then diluted 110 in 10mM sodium acetate buffer (pH =45) and added at 30 120583Lmin for 210 sec The surface wasblocked with 1M ethanolamine for 600 sec at 5 120583Lmin 31015840-biotin-T

104A018 aptamer was heated to 95∘C then intro-

duced at 5 120583M in 5mM MgSO4for 180 sec at 30 120583Lmin

Samples were diluted to 273120583M in HEPES buffer fromBiacore (10mM HEPES 150mM NaCl 005 SurfactantP20) and dialyzed into HEPES buffer overnight The sampleswere serially diluted to appropriate concentrations thenintroduced onto the chip at 30120583Lmin for 30 sec Dataanalysis was performedwith BIAevaluation software for threeexperiments of three replicates for each compound withthe background subtracted data (reference channel 1) andplotted in GraphPad Prism 5 using a one site total bindingmodel Each experiment included freshly prepared reagentsintroduced onto the same chip in triplicate

26 NMR Structural Analysis Proton NMR of aptamer solu-tions was performed on a 400MHz Bruker Avance NMRspectrometer DNA was dissolved in PBSTMB buffer withH2O D2O ratios of 90 10 and 0 100 at aptamer concen-

trations of 1mM for 4A018 and 13mM for the TBA 15merThe imino proton spectra were acquired with the W5 watersuppression pulse sequence and 2D NOESY spectra wereacquired with 1024 points in the direct dimension and 256points in the indirect dimensionwith a sweepwidth of 10 ppmand a mixing time of 02 s

3 Results and Discussion

A full description of pattern design and analysis is includ-ed in Supporting Information text and Figure S1 in Supple-mentaryMaterial available online at httpdxdoiorg1011552015137489 The pattern PT2 was applied to the microarrayfor assessing thrombin binding Several potential thrombinaptamer candidates were identified from the microarraywork The top 15 candidates were ranked by fluorescenceintensity (Figure 1) in comparison to the positive controlsof the reported thrombin fibrinogen binding site aptamer(TFBS) and thrombin heparin binding site aptamer (THBS)[21 22] A mutation of the streptavidin aptamer (SA) withan extra guanine base inserted at position 19 was used asa negative control due to its low binding observed with avariety of different targets in preliminary work [23] All ofthe top 15 reported sequences demonstrated over 12x highermean fluorescence intensity values than SA with the highestintensity sequence 4A018 reporting mean values over 120xhigher than SA All fluorescence intensity values for the top15 sequences were statistically significant compared to the SAcontrol (119901 lt 00001 95 CI) A close-up of the fluorescenceintensity values of the top 5 sequences compared to SA is alsoshown (Figure 1 inset)

The sequences for each of the top 15 candidates arereported in Table S1 with the predicted secondary structuresof the top 5 depicted in Figure S2 The first 6ndash8 bases of all15 sequences are similar and also resemble the first 6ndash8 bases


0100002000030000400005000060000700008000090000

100000

TFBS

THBS

4A01

81A

175

1A52

32A

353

4A20

20A

044

2A13

33A

645

3A32

51A

529

3A58

91A

926

0A41

80A

189

4A20

8SA

Sequence ID

Fluo

resc

ence

inte

nsity

02000400060008000

10000120001400016000

4A018 1A175 1A523 2A353 4A202 SASequence ID

Fluo

resc

ence

inte

nsity

Figure 1 Microarray performance of the top 15 ranked potential thrombin binders of PT2 compared to controls Inset close-up view of theintensities of the top 5 ranked sequences compared to the negative control Error bars represent standard deviation of replicates of fluorescencevalues obtained from a 2 h incubation of 100 nM Cy3-thrombin with the microarray at 20∘C

of the TFBS sequence Previous microarray studies showeda similar trend with the final thrombin binding sequencesafter four rounds of mutation also demonstrating a 51015840-GGTTGG consensus sequence [9] While the pattern (PT2)does constrain certain bases in specified positions it doesnot account for the abundance of sequence similarity in thefirst several bases of microarray candidates Further analysisof the sequences included in the initial library indicated thatthe number of sequences containing 51015840-GGTTGG (34 of5000 total sequences) was similar to the 35 of sequencescontaining a 51015840-GGCCGG signifying that the library wasnot biased with an abundance of the motif (see SupportingInformation text and Table S2) Random sequences (TableS3) without the 51015840-GGTTGGmotif did not exhibit significantbinding (Figure S3)

To confirm the microarray results due to the low numberof technical replicates on the array the top ranked sequence4A018 was assessed for target binding by SPR Two separateflow channels were immobilized with biotinylated 4A018with a 31015840-T

10linker included to permit some flexibility of

the sequence and increase the distance of the potentialbinder from the surface of the SPR chip Proteins similarin molecular weight andor also found in the body bovineserum albumin (BSA) and human serum albumin (HSA) aswell as a compoundwith an isoelectric point (pI) close to thatof thrombin neuropeptideY (NPY) were assayed for bindingin addition to thrombin

Thrombin was the only molecule that significantly inter-acted with 4A018 in either of the flow channels (Figure 2)Flow channel 4 (Figure 2(b)) had an increased relativeresponse compared to channel 3 (Figure 2(a)) likely dueto higher immobilization of 4A018 in channel 4 (24820RU) versus channel 3 (17546 RU) The curve fit produceddissociation constants (119870

119889) of 404 plusmn 031 120583M in channel 3

and 396 plusmn 029 120583M from channel 4 (Figure S4) while BSAHSA and NPY demonstrated negligible binding and didnot fit the curve (Figures 2 and S4) Flow channels 3 and

4 served as independent replicates confirming the responsewith each measurement plusmn1 of the mean 119870

119889(400 120583M)

These SPR results show that 4A018 has affinity for thrombinprotein rather than the fluorescent tag used in themicroarrayexperiments and that the interaction is specific to thrombindespite addition of proteins of similar molecular weight andpI Furthermore it demonstrates that the use of proper con-trols and replicates during amicroarray experiment enhancesthe ability to differentiate a real binding event from spuriousinteractions in a massively parallel format

One reason the affinity of themicroarray selected aptamermay be lower than that of the solution-based aptamers(nanomolar dissociation constants) is because solution-basedSELEX is an evolutionary process designed to select the ldquobestrdquobinders by employing multiple rounds of increasingly highstringency conditions Furthermore a much smaller startingpopulation was utilized in the microarray experiment andconstraints were imposed that may preclude identification ofa sequence with a higher affinity In addition the SELEX gen-erated aptamers were fully optimized by truncating primersand nonessential bases and identifying a consensus sequenceFurther optimization of patterns and stringency conditionsas well as mutationaltruncation analysis could lead toaptamerswith enhanced target affinityThe reported119870

119889of the

microarray-identified sequence may also be improved underfully optimized binding conditions or by applying geneticalgorithms to the lead compound over multiple microarrayrounds [9ndash11] Refinement of these conditionswill be aided byconsideration of the sequences of more confirmed thrombinbinders from themicroarray and by testingmore sequences ofthe patterned libraries following this proof-of-concept work

In contrast with promoting the sequence as an idealaptamer this SPR binding study instead validated the overallhypothesis that a patterned library aids in identifying anaptamer despite the relatively low sequence space covered ina microarray compared to SELEX The enhanced stringencyof the microarray conditions generated a sequence with high


0 2 4 6 8 10 12

800

600

400

200

0

Rela

tive r

espo

nse (

RU)

ThrombinBSA

HSANPY

[Sample] (120583M)

(a)

0 2 4 6 8 10 12

800

1000

1200

1400

600

400

200

0

Rela

tive r

espo

nse (

RU)

ThrombinBSA

HSANPY

[Sample] (120583M)

(b)

Figure 2 SPR response of 4A018 with analytes in flow channel 3 (a) and flow channel 4 (b) Analytes assayed were thrombin (circle) BSA(square) HSA (triangle) and NPY (downward triangle) at 0ndash109120583M Error bars represent standard deviation of replicates for the mean ofthree separate experiments each performed in triplicate

target specificity and a binding sequence was identified byapplying the very first pattern designedwith no considerationto reported thrombin binders The binding demonstrated inthe SPR experiments shows that an aptamer identified whileimmobilized on a microarray surface will retain bindingwhen transitioned to a new biosensing platform requiringaptamer surface linkage Designing a microarray with lowreplicates of different patterns may serve as a screeningmechanism to determine optimal patterns or homologousstretches necessary for target binding Rather than providingthe highest affinity sequence this proof-of-principle studyhighlights the use of a computational driven methodology tocreate a DNA library rather than a SELEX based approach

Previous reports have shown that the well-known TFBSand THBS thrombin aptamers fold into G-quartet structures[21 22] Due to the 51015840- and internal (including fiveGG repeatsand one GGG) structural similarity of 4A018 and TFBSwe investigated the possibility that 4A018 also formed a G-quartet structure

G-quartets have a number of unique features includingin-plane pairing of four guanine bases slow imino protonexchange high thermal stability and syn conformationsabout some guanine glycosidic linkages that can be identifiedby NMR [24] While previous studies have used site-specificNMR assignments andmultidimensional NMR to determinethe three dimensional structure of aptamers [25] we showhere that simple screening NMR experiments can be used torule out G-quartet formation in the 4A018 aptamer

The imino proton NMR spectra is very sensitive toaptamer folding and the hydrogen bonding patterns resultingfrom the formation of G-quartets AT and GC base pairsloops and mismatched base pairs [24] The imino protonsexchange rapidly with water and cannot be directly observedunless they are protected by hydrogen bond formation orfolded into loops The lowest energy structure of 4A018

(a)

(b)

16 15 14 13 12 11

Chemical shift (ppm)

Figure 3 The 400MHz proton imino spectra of 4A018 at (a) 278and (b) 298K

(Figure S2) shows a 7-base pair stem (G1-T7G44-C50) anda stem starting at C23-G43 with three or more base pairsand three additional base pairs following two non-pairednucleotides This conformation incorporates two G doubletsinto stem 1 and two G doublets and a triplet into stem2 and one of the G doublets is split between stem 1 andstem 2 In addition each of stems 1 and 2 has two AT basepairs which would be observed in the chemical shift rangeof 135ndash145 ppm rather than the higher field range of 10-11 ppmobserved for the T imino protons in short loops Giventhese differences it should be possible to distinguish betweenG-quartet formation and the calculated structure shown inFigure S2

A comparison of the imino proton spectra of 4A018 at278 (Figure 3(a)) and 298K (Figure 3(b)) shows two wellresolved peaks that can be assigned to the AT (139 ppm)and GC (128 ppm) base pairs [26 27] The relative ATGCratio of the peak areas is sim1 15 in NMR (Figure 3(a)) whichis comparable to the 1 175 ratio expected from the number


0

1

2

3

4

5

6

778 75


Chem

ical

shift

(ppm

)

GH8

GH1998400

H8H6H2998400 H2

998400998400

(a)

78 75


0

1

2

3

4

5

6

7

Chem

ical

shift

(ppm

)

H8H6H2998400 H2

998400998400

(b)

Figure 4 The 2D NOESY NMR spectra for TFBS (a) and 4A018 (b) in D2O acquired with a 02 s mixing time

of AT and GC pairs in the lowest energy structure (FigureS2) The spectra show significant changes as the temperatureincreases from 278 to 298K which is consistent with theformation of short helical regions as in stem 2 rather thana more stable G-quartet No peaks are observed in the regionbetween 10 and 12 ppmwhich contains the signals from basesin protected folds and mismatched base pairs The two GTmismatches predicted for 4A018 in Figure S2 are located at theend of stem regions andwould be difficult to observe byNMRdue to solvent exposure This result suggests that the actualstructure is consistent with the predicted structure shown inFigure S2

Another feature of G-quartets identifiable in the NMRspectra is the syn conformation of the glycosidic angle insome of the guanine bases [24] The syn conformation canbe detected by 2D Nuclear Overhauser Effect spectroscopy(NOESY) [27] since the cross peak intensities depend onthe inverse sixth power of the internuclear distance and thechange from the anti to the syn conformation shortens thedistance between theGH8andGH11015840 protons from37 to 21 AAdditionally software designed to predict oligonucleotidesecondary structure is typically unreliable in ability to reportthe formation of higher order structures (including G-quartets) which can be confirmed by 2D NOESY

The 2DNOESY spectra for TFBS (Figure 4(a)) and 4A018aptamer (Figure 4(b)) show the chemical shift correlationbetween the DNA H8 and H6 base protons (65ndash85 ppm)with the H11015840 sugar protons (5ndash65 ppm) the H21015840 and H210158401015840sugar protons (18ndash3 ppm) and the thymine methyl protons(1 ppm) [27] The G-quartet conformation of the TFBSaptamer has four guanines in the syn conformation thatgive rise to four strong GH8-GH11015840 cross peaks enclosedin the circle in Figure 4 The four GH8-GH11015840 cross peaks

have similar chemical shifts in both dimensions and can bevisualized in an expanded viewof the spectra (not shown)NostrongGH8-GH11015840 cross peaks are observed in the 2DNOESYspectra for 4A018 showing that the solution conformation of4A018 does not contain guanines in the syn conformationThis supports the hypothesis that 4A018 does not adopt aG-quartet structure and it backs the imino proton NMRresults suggesting the accuracy of the predicted structure(Figure S2)

The microarray evolution leading to a thrombin aptamerby Platt similarly did not adopt a G-quartet structure [9]While the microarray work described here and by Plattmay essentially preclude this structural feature due to spa-tial andor researcher-imposed initial library constraints itis possible that solution-based methods may promote theselection of G-quartets to thrombin in ways which remainunclear [21 22 28]This aspect of different structures selecteddepending on the selection mechanism may be of particularinterest to researchers intending to apply solution-selectedaptamers immobilized on a platform for biosensor design

4 Conclusions

This work illustrates the potential of DNA microarray tech-nology for aptamer identification and highlights patternedlibraries designed without prior binding sequence consider-ation as a viable solution to the limitations on microarrayoligonucleotide surface density This method emphasizes arational computational driven methodology to DNA librarycreation rather than a SELEX approach While the resultsof these initial proof-of-concept studies may not currentlyimprove upon SELEX in terms of aptamer affinity themicroarrays rapidly provide a starting point to performadditional experiments to generate higher affinity aptamers


based on the identified sequences Binding candidates canbe identified and ranked in less than one week utilizingmicroarray experiments presenting methodology that ismore amenable to potential high throughput applicationsthan traditional SELEX One area of immediate impactfor this methodology is directed at the biosensor field byrendering it feasible to identify a functional aptamer directlyimmobilized on a solid support to mitigate the variabilityor elimination of affinity often observed in aptamers thatare selected in solution for applications that involve a sur-face linkage Furthermore aptamer based electrochemicalor gold nanoparticle biosensing technologies amplify signaland detect analytes at levels orders of magnitude lowerthan the 119870

119889 reducing the reliance of sensor performance

on affinity [29 30] These platforms also rely on a confor-mational change of the aptamer to indicate target bindingso knowledge of the structural properties of an aptamer isessential in effective sensor platform designThis work showsthat a novel combination of the imino proton NMR and2D NOESY simplifies screening for aptamer conformationcompared to establishing sequence-specific assignmentsTheNMR studies also demonstrated that the microarray selectedaptamer did not form the G-quartet structure commonto solution-based SELEX thrombin aptamers This findingraises consideration that different binding modes may domi-nate in surface-immobilized aptamer identification strategiesin comparison to traditional solution-based SELEX There-fore microarray aptamer identification may be complimen-tary to SELEX in the sense that different types of binderscould be produced depending on the desired applicationFuture focus areas include investigating the performance ofdifferent patterns including in-depth motif analysis of bothbinders and nonbinders as well as the effects of a combinedmicroarraySELEX scheme

Conflict of Interests

The authors declare that there is no conflict of interestsregarding the publication of this article

Acknowledgments

The authors thank Dr Richard Chapleau and Dr DavidRiddle for their assistance with SPR measurements anddata interpretation and Agilent Technologies for the loanof the Agilent High-resolution Microarray Scanner Thisresearchwas performedwhile the author (Jennifer AMartin)held a National Research Council Research AssociateshipAward at Wright-Patterson Air Force Base This work wassupported by the Air Force Office of Scientific Research AirForce Human Signatures Branch (711 HPWRHXB) CoreAir Force Research Laboratory Bio-X Strategic TechnologyThrust Biotronics and Defense Forensic Science Center

References

[1] A D Ellington and J W Szostak ldquoIn vitro selection of RNAmolecules that bind specific ligandsrdquo Nature vol 346 no 6287pp 818ndash822 1990

[2] C Tuerk and L Gold ldquoSystematic evolution of ligands byexponential enrichment RNA ligands to bacteriophage T4DNApolymeraserdquo Science vol 249 no 4968 pp 505ndash510 1990

[3] R Stoltenburg C Reinemann and B Strehlitz ldquoSELEX-a(r)evolutionary method to generate high-affinity nucleic acidligandsrdquo Biomolecular Engineering vol 24 no 4 pp 381ndash4032007

[4] S D Jayasena ldquoAptamers an emerging class of molecules thatrival antibodies in diagnosticsrdquo Clinical Chemistry vol 45 no9 pp 1628ndash1650 1999

[5] S C B Gopinath ldquoMethods developed for SELEXrdquo Analyticaland Bioanalytical Chemistry vol 387 no 1 pp 171ndash182 2007

[6] M V Berezovski M U Musheev A P Drabovich J V Jitkovaand S N Krylov ldquoNon-SELEX selection of aptamers with-out intermediate amplification of candidate oligonucleotidesrdquoNature Protocols vol 1 no 3 pp 1359ndash1369 2006

[7] N Savory K Abe K Sode and K Ikebukuro ldquoSelectionof DNA aptamer against prostate specific antigen using agenetic algorithm and application to sensingrdquo Biosensors andBioelectronics vol 26 no 4 pp 1386ndash1391 2010

[8] M Cho Y Xiao J Nie et al ldquoQuantitative selection of DNAaptamers through microfluidic selection and high-throughputsequencingrdquo Proceedings of the National Academy of Sciences ofthe United States of America vol 107 no 35 pp 15373ndash153782010

[9] M Platt W Rowe D C Wedge D B Kell J Knowles andP J R Day ldquoAptamer evolution for array-based diagnosticsrdquoAnalytical Biochemistry vol 390 no 2 pp 203ndash205 2009

[10] C G Knight M PlattW Rowe et al ldquoArray-based evolution ofDNA aptamers allows modelling of an explicit sequence-fitnesslandscaperdquoNucleic Acids Research vol 37 no 1 article e6 2009

[11] R Asai S I Nishimura T Aita and K Takahashi ldquoIn Vitroselection of DNA aptamers on chips using a method forgenerating pointmutationsrdquoAnalytical Letters vol 37 no 4 pp645ndash656 2004

[12] J R Collett J C Eun and A D Ellington ldquoProduction andprocessing of aptamer microarraysrdquo Methods vol 37 no 1 pp4ndash15 2005

[13] E J Cho J R Collett A E Szafranska and A D EllingtonldquoOptimization of aptamer microarray technology for multipleprotein targetsrdquo Analytica Chimica Acta vol 564 no 1 pp 82ndash90 2006

[14] E Katilius C Flores and N W Woodbury ldquoExploring thesequence space of a DNA aptamer using microarraysrdquo NucleicAcids Research vol 35 no 22 pp 7626ndash7635 2007

[15] S E Osborne and A D Ellington ldquoNucleic acid selection andthe challenge of combinatorial chemistryrdquo Chemical Reviewsvol 97 no 2 pp 349ndash370 1997

[16] J M Carothers S C Oestreich J H Davis and J W SzostakldquoInformational complexity and functional activity of RNAstructuresrdquo Journal of the American Chemical Society vol 126no 16 pp 5130ndash5137 2004

[17] Y Chushak and M O Stone ldquoIn silico selection of RNAaptamersrdquo Nucleic Acids Research vol 37 no 12 article e872009

[18] J H Davis and J W Szostak ldquoIsolation of high-affinity GTPaptamers from partially structured RNA librariesrdquo Proceedingsof the National Academy of Sciences of the United States ofAmerica vol 99 no 18 pp 11616ndash11621 2002

[19] X LuoMMckeague S Pitre et al ldquoComputational approachestoward the design of pools for the in vitro selection of complexaptamersrdquo RNA vol 16 no 11 pp 2252ndash2262 2010


[20] K M Ruff T M Snyder and D R Liu ldquoEnhanced functionalpotential of nucleic acid aptamer libraries patterned to increasesecondary structurerdquo Journal of the American Chemical Societyvol 132 no 27 pp 9453ndash9464 2010

[21] L C Bock L C Griffin J A Latham E H Vermaas and J JToole ldquoSelection of single-stranded DNA molecules that bindand inhibit human thrombinrdquo Nature vol 355 no 6360 pp564ndash566 1992

[22] D M Tasset M F Kubik and W Steiner ldquoOligonucleotideinhibitors of human thrombin that bind distinct epitopesrdquoJournal of Molecular Biology vol 272 no 5 pp 688ndash698 1997

[23] J A Bittker B V Le and D R Liu ldquoNucleic acid evolutionandminimization by nonhomologous random recombinationrdquoNature Biotechnology vol 20 no 10 pp 1024ndash1029 2002

[24] R F Macaya P Schultze F W Smith J A Roe and JFeigon ldquoThrombin-binding DNA aptamer forms a unimolecu-lar quadruplex structure in solutionrdquoProceedings of theNationalAcademy of Sciences of the United States of America vol 90 no8 pp 3745ndash3749 1993

[25] D J Patel A K Suri F Jiang et al ldquoStructure recognitionand adaptive binding in RNA aptamer complexesrdquo Journal ofMolecular Biology vol 272 no 5 pp 645ndash664 1997

[26] G Zheng A M Torres and W S Price ldquoSolvent suppressionusing phase-modulated binomial-like sequences and applica-tions to diffusion measurementsrdquo Journal of Magnetic Reso-nance vol 194 no 1 pp 108ndash114 2008

[27] K Wuthrich NMR of Proteins and Nucleic Acids John Wiley ampSons New York NY USA 1986

[28] G V Kupakuwana J E Crill III M P McPike and P N BorerldquoAcyclic identification of aptamers for human alpha-thrombinusing over-represented libraries and deep sequencingrdquo PLoSONE vol 6 no 5 Article ID e19395 2011

[29] J A Hagen S N Kim B Bayraktaroglu et al ldquoBiofunction-alized zinc oxide field effect transistors for selective sensing ofriboflavin with current modulationrdquo Sensors vol 11 no 7 pp6645ndash6655 2011

[30] J A Martin J L Chavez Y Chushak R R Chapleau JHagen andN Kelley-Loughnane ldquoTunable stringency aptamerselection and gold nanoparticle assay for detection of cortisolrdquoAnalytical and Bioanalytical Chemistry vol 406 no 19 pp4637ndash4647 2014

Submit your manuscripts athttpwwwhindawicom

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Inorganic ChemistryInternational Journal of

Hindawi Publishing Corporation httpwwwhindawicom Volume 2014

International Journal ofPhotoenergy


Carbohydrate Chemistry

International Journal of


Journal of

Chemistry


Advances in

Physical Chemistry

Hindawi Publishing Corporationhttpwwwhindawicom

Analytical Methods in Chemistry

Journal of

Volume 2014

Bioinorganic Chemistry and ApplicationsHindawi Publishing Corporationhttpwwwhindawicom Volume 2014

SpectroscopyInternational Journal of


The Scientific World JournalHindawi Publishing Corporation httpwwwhindawicom Volume 2014

Medicinal ChemistryInternational Journal of


Chromatography Research International


Applied ChemistryJournal of



Theoretical ChemistryJournal of


Journal of

Spectroscopy

Analytical ChemistryInternational Journal of


Journal of


Quantum Chemistry


Organic Chemistry International

ElectrochemistryInternational Journal of



CatalystsJournal of


aptamer identification The sequences reported likely reflectthe most abundant sequences present but may not reportthose that have artificially lower numbers due to factors suchas PCR bias or cloning efficiency or due to low sampling ofthe entire sequence space (diversity) of the final pool [8]

Several groups have proposed using DNAmicroarrays toaddress the possible SELEX biases and expedite the aptameridentification process [9ndash11] DNA microarrays function byidentifying locations of fluorescently labeled targets andcorrelating themwith the position of known sequences cova-lently synthesized on the arrayThe arrays are fully customiz-able so the user can define the exact sequences of interest pro-duced on the arrays A single microarray experiment can becompleted in less than one day with an additional 1-2 days fordata analysisThis characteristic is a function of the improvedpartitioning efficiency available by covalently linking thesequences to the surface Higher stringency conditions canbe applied to identify sequences with more ideal bindingproperties PCR cloning and sequencing are not requiredsince known sequences are in predefined locations Alsomicroarrays are particularly useful for identifying aptamersfor biosensor applications since the response of an aptamerselected by solution-based SELEX may be significantlydiminished when it is tethered to a sensor surface [12ndash14]

A major drawback of microarray use is that the highestdensity arrays have a maximum of sim106 sequences incontrast to SELEX methods which evolve from an initiallibrary of sim1015 sequences However combinatorial drug-screening libraries successfully identify binders with only103ndash106 different compounds in the starting library due tothe diversity of the functional groups of the compounds[15] Extending this premise to oligonucleotides it has beendetermined that the probability for a sequence to bind atarget improves with increasing structural complexity [16]This means unstructured sequences or oligonucleotidesthat form simple structures such as those in a randomoligonucleotide library have reduced potential to show anytype of function Constituents of the random pool consist ofmostly unpaired regions combined with short (low stability)stem-loop structures and the probability of containing anabundance of more complex high-affinity aptamers in thestarting random library is low [17] Several groups haveexplored this issue experimentally by optimizing the SELEXstarting library to contain more complex partially structuredsequences [18ndash20] These studies showed that the partiallystructured libraries provided more sequences that bindthe target andor sequences with higher binding affinitiescompared to a completely random starting library

The same principle can be broadened to biasing a mi-croarray starting library to contain sequences with increasedstructural complexity and thus enhanced potential for targetbinding In previous works [9ndash11] starting libraries consistedof 102ndash104 initial sequences and evolved aptamers throughin silico genetic algorithms in multiple chip generations(rounds) These studies used either naturally fluorescenttargets utilized a target with well-studied aptamer bindingmotifs or took into account the characteristics of knownaptamers for library design

In this work we applied DNA library patterning in orderto circumvent the microarray density problem and rapidlyidentified an aptamer to the target molecule thrombin ina single round employing a pattern which considered nostructural information from previously reported thrombinaptamers We show that this aptamer does not form thewell-known G-quartet structure reported in early iterationsof thrombin aptamers by utilizing a combination of iminoproton NMR and 2D NMR for structural characterizationthat is simpler than legacy methods of establishing sequence-specific assignments The NMR results also raise questionson whether the aptamer identification platform (surface-immobilized or solution-based) may significantly influencethe binding mechanism of the final aptamer These resultsgenerally demonstrate a promisingmethod for rapidly identi-fying and characterizing aptamers whichmay directly benefitthe field of aptamer biosensors where immobilization ofsolution-selected aptamers on a sensing platform has provento be challenging

2 Materials and Methods

21 Chemicals and Equipment The chemicals used were IgE(Fitzgerald) Thrombin (Haematologic Technologies Inc)BSA and HSA (Sigma) neuropeptide Y (Pheonix Pharma-ceuticals) IllustraNAP-25 desalting columns andCy3Mono-Reactive Dye Pack (GEHealthcare) NanoDrop (Thermo Sci-entific) nuclease free water (Gibco) Microarray equipmentconsisted of the following custom 8times15 k DNAmicroarrays8 times 15 k gasket slides ozone barrier slides hybridizationchambers scanner cassettes hybridization oven and High-resolution Microarray Scanner (all Agilent) and slide rackand wash dishes (Shandon) and Kimtech polypropylenewipes (Kimberly-Clark) All DNA was purchased throughIDT

4A018 GGTTGGTTTTTCAATCAGCGATCGCG-GAATCCAGGGTTAGGCGGCCAACC (with andwithout 31015840-T

10-Biotin moiety)

TFBS GGTTGGTGTGGTTGG

Buffers Binding [PBSMTB]-1x PBS (81mMNa2HPO4

11 mM KH2PO4 27mM KCl 137mM NaCl pH

74) + 1mM MgCl2+ 01 Tween-20 and 1 BSA

Washing [PBSM]-1x PBS (81mM Na2HPO4 11mM

KH2PO4 27mMKCl 137mMNaCl pH 74) + 1mM

MgCl2 Rinse [R]-14 dilution of PBSM and nuclease

free water

22 Protein Labeling Thrombin was labeled with Cy3 using aCy3 Mono-Reactive Dye Pack (GE Healthcare) Protein wasdiluted to 1mgmL in 1mL 01M Na

2CO3buffer (pH = 93)

and incubated for 30 minutes The product was purified withTexas Red protein labeling size exclusion column (MolecularProbes) and the dye-to-protein ratio (DP) was calculated as08DP using the manufacturerrsquos instructions with UVVisdetection





10spacer













0100002000030000400005000060000700008000090000

100000

TFBS

THBS

4A01

81A

175

1A52

32A

353

4A20

20A

044

2A13

33A

645

3A32

51A

529

3A58

91A

926

0A41

80A

189

4A20

8SA

Sequence ID

Fluo

resc

ence

inte

nsity

02000400060008000

10000120001400016000


Fluo

resc

ence

inte

nsity










119889(400 120583M)



119889of the




0 2 4 6 8 10 12

800

600

400

200

0

Rela

tive r

espo

nse (

RU)

ThrombinBSA

HSANPY

[Sample] (120583M)

(a)

0 2 4 6 8 10 12

800

1000

1200

1400

600

400

200

0

Rela

tive r

espo

nse (

RU)

ThrombinBSA

HSANPY

[Sample] (120583M)

(b)






(a)

(b)

16 15 14 13 12 11






0

1

2

3

4

5

6

778 75


Chem

ical

shift

(ppm

)

GH8

GH1998400

H8H6H2998400 H2

998400998400

(a)

78 75


0

1

2

3

4

5

6

7

Chem

ical

shift

(ppm

)

H8H6H2998400 H2

998400998400

(b)







4 Conclusions








Acknowledgments


References









































Journal of

Chemistry


Advances in

Physical Chemistry



Journal of

Volume 2014














Journal of

Spectroscopy



Journal of


Quantum Chemistry






CatalystsJournal of





10spacer













0100002000030000400005000060000700008000090000

100000

TFBS

THBS

4A01

81A

175

1A52

32A

353

4A20

20A

044

2A13

33A

645

3A32

51A

529

3A58

91A

926

0A41

80A

189

4A20

8SA

Sequence ID

Fluo

resc

ence

inte

nsity

02000400060008000

10000120001400016000


Fluo

resc

ence

inte

nsity










119889(400 120583M)



119889of the




0 2 4 6 8 10 12

800

600

400

200

0

Rela

tive r

espo

nse (

RU)

ThrombinBSA

HSANPY

[Sample] (120583M)

(a)

0 2 4 6 8 10 12

800

1000

1200

1400

600

400

200

0

Rela

tive r

espo

nse (

RU)

ThrombinBSA

HSANPY

[Sample] (120583M)

(b)






(a)

(b)

16 15 14 13 12 11






0

1

2

3

4

5

6

778 75


Chem

ical

shift

(ppm

)

GH8

GH1998400

H8H6H2998400 H2

998400998400

(a)

78 75


0

1

2

3

4

5

6

7

Chem

ical

shift

(ppm

)

H8H6H2998400 H2

998400998400

(b)







4 Conclusions








Acknowledgments


References









































Journal of

Chemistry


Advances in

Physical Chemistry



Journal of

Volume 2014














Journal of

Spectroscopy



Journal of


Quantum Chemistry






CatalystsJournal of


0100002000030000400005000060000700008000090000

100000

TFBS

THBS

4A01

81A

175

1A52

32A

353

4A20

20A

044

2A13

33A

645

3A32

51A

529

3A58

91A

926

0A41

80A

189

4A20

8SA

Sequence ID

Fluo

resc

ence

inte

nsity

02000400060008000

10000120001400016000


Fluo

resc

ence

inte

nsity










119889(400 120583M)



119889of the




0 2 4 6 8 10 12

800

600

400

200

0

Rela

tive r

espo

nse (

RU)

ThrombinBSA

HSANPY

[Sample] (120583M)

(a)

0 2 4 6 8 10 12

800

1000

1200

1400

600

400

200

0

Rela

tive r

espo

nse (

RU)

ThrombinBSA

HSANPY

[Sample] (120583M)

(b)






(a)

(b)

16 15 14 13 12 11






0

1

2

3

4

5

6

778 75


Chem

ical

shift

(ppm

)

GH8

GH1998400

H8H6H2998400 H2

998400998400

(a)

78 75


0

1

2

3

4

5

6

7

Chem

ical

shift

(ppm

)

H8H6H2998400 H2

998400998400

(b)







4 Conclusions








Acknowledgments


References









































Journal of

Chemistry


Advances in

Physical Chemistry



Journal of

Volume 2014














Journal of

Spectroscopy



Journal of


Quantum Chemistry






CatalystsJournal of


0 2 4 6 8 10 12

800

600

400

200

0

Rela

tive r

espo

nse (

RU)

ThrombinBSA

HSANPY

[Sample] (120583M)

(a)

0 2 4 6 8 10 12

800

1000

1200

1400

600

400

200

0

Rela

tive r

espo

nse (

RU)

ThrombinBSA

HSANPY

[Sample] (120583M)

(b)






(a)

(b)

16 15 14 13 12 11






0

1

2

3

4

5

6

778 75


Chem

ical

shift

(ppm

)

GH8

GH1998400

H8H6H2998400 H2

998400998400

(a)

78 75


0

1

2

3

4

5

6

7

Chem

ical

shift

(ppm

)

H8H6H2998400 H2

998400998400

(b)







4 Conclusions








Acknowledgments


References









































Journal of

Chemistry


Advances in

Physical Chemistry



Journal of

Volume 2014














Journal of

Spectroscopy



Journal of


Quantum Chemistry






CatalystsJournal of


0

1

2

3

4

5

6

778 75


Chem

ical

shift

(ppm

)

GH8

GH1998400

H8H6H2998400 H2

998400998400

(a)

78 75


0

1

2

3

4

5

6

7

Chem

ical

shift

(ppm

)

H8H6H2998400 H2

998400998400

(b)







4 Conclusions








Acknowledgments


References









































Journal of

Chemistry


Advances in

Physical Chemistry



Journal of

Volume 2014














Journal of

Spectroscopy



Journal of


Quantum Chemistry






CatalystsJournal of







Acknowledgments


References









































Journal of

Chemistry


Advances in

Physical Chemistry



Journal of

Volume 2014














Journal of

Spectroscopy



Journal of


Quantum Chemistry






CatalystsJournal of






















Journal of

Chemistry


Advances in

Physical Chemistry



Journal of

Volume 2014














Journal of

Spectroscopy



Journal of


Quantum Chemistry






CatalystsJournal of

Research Article Single-Round Patterned DNA Library...

Documents

Transcript of Research Article Single-Round Patterned DNA Library...