Vol. 44 ı No. 6 ı 2008 www.biotechniques.com ı BioTechniques ı 815 enchmarks Protein expression monitoring can be a goo too to ete rm ne se ase onset and progression, and is also used in drug discovery and investigational science (1–4). Howe ver, co mm on y us e as sa ys ac t e sensitivity and specificity required for the detection of low-abundance proteins (5,6). Severa s gna amp cat on t ec noog es have been designed and applied to protein detection assays, including rolling-circle amp cat on ( RCA) ( ,7), t yram e s gna amplification (TSA) (8), and enzyme-linked immunosorbent assay (ELISA) coupled w t PCR (9). A t oug a tec no og es have successfully provided signal amplifi- cation to protein detection assays, none has een w e y app e to m crotter p ate or microarray ELISAs, presumably due to the required additional steps and/or equipment compare wt t e stan ar assay . DNA dendrimers are highly branched structures, constructed from seven different single strands of DNA. Some strands are complementary to each other in the middle section, and when incubated under the right conditions they hybridize, forming “X”-like structures called monomers. Monomers are the building blocks of the dendrimer. Different monomers hybridize through their single-stranded regions to form the layers ofthe dendrimer (10). The 3DNA dendrimer technology from Genisphere (Hatfield, PA, USA) has been established as a robust system to provide signal amplification for RNA detection (11–13). Here, we present an application f the 3DNA dendrimer technology to prote n ete ct on. O gonuc eot e-con u- ated antibodies (for selectivity) and nzymes or dyes (for signal) are hybridized to t e outer arms o t e en r mer , w c functions as a core. Figures 1, A and B, show a schematic representation of the end product, called UltraAmp (Genisphere), and its assembly. UltraAmp reagents are prepared by combining a stock solution ofthe dendrimer with solutions of the antibody and enzyme-conjugated oligonucleotides. he buffer is spiked with NaCl to a final oncentration of 200 mM to enhance the stability of the UltraAmp complex. After 20 in incubation, the solution is diluted to 6 g/μL as dendrimer with PBS and 50 % v/v final SuperFreeze (Pierce Biotechnology, Inc. Rockford, IL, USA). Standard sandwich ELISAs were run as recommended by the manufacturer (R&D Systems, Minneapolis, MN, USA). In UltraAmp assays, all standard conditions were maintained, except for the incubation with streptavidin horseradish peroxidase (SA-HRP; R&D Systems), which was replaced with UltraAmp Anti-biotin HRP (Cat. no. AB2080; Genisphere) diluted in Binding Buffer I (Cat. no. ABB100; Genisphere) to a final dendrimer concen- trat on o 0. ng/ μL. Eac we was incubated with 50 μL of diluted UltraAmp reagent for 1 h on a reciprocating shaker at approx mate y 100 rpm at room tempe rature (RT). The wells were emptied, washed 5–6 Protein detection enhanced by 3DNA dendrimer signal amplification Johanna R. Mora 1 , T amara L. Zielinski 2 , Bryce P. Nelson 2 , and Robert C. Getts 1 1 Genisphere, Inc. Hatfield, PA and2 GenTel BioSciences, Inc., Madison, WI, USA BioTechniques 44:815-818 (May 2008) doi 10.2144/000112733 DNA dendrimers, conjugated with both anti-biotin antibodies and up to 350 labeling enti- ties, were designed and adapted to protein microarray and enzyme-linked immunosorbent as- say (ELISA) to improve the limits of protein detection with no additional steps or equipment. Application of conjugated dendrimers to standard ELISA cytokine detection resulted in up to threefold improvement of the limits of detection with no significant increase in the inter- andintra-assay coefficient of variation (CV) compared to streptavidin horseradish peroxidase (SA- HRP) detection. The adaptation of conjugated dendrimers to pr otein microarray cytokine de- tection resulted in up to 10-fold improvement of the limits of detection, but assay conditions would have to be optimized to decrease the intra- and inter-assay %CVs. Figure 1. Schematic representation of the UltraAmp reagent assembly and output signal for the detection of interleukin-1 β (IL-1 β). (A) Diagram of a four-layer 3DNA dendrimer. (B) Magnified representation of two of the single-stranded arms (gray) in the outer layer of the dendrimer that allow ttachment of dye-/HRP- and antibody-conjug ated oligonucleotides by base-pairing interactions. Dye- nd HRP-labeled oligonucleotides are represented in solid black, and the antibody-conjugated oligonu- cleotide is represented in dashed gray. Standard curves obtained for the detection of IL-1 β in sandwich ELISAs (C) using standard (¢, solid line) and UltraAmp (l, dashed line) detection; and in protein microarrays (D) using both standard (p, dashed line) and UltraAmp ( n, solid line) detection. C
Transcript
8/3/2019 Johanna R. Mora, Tamara L. Zielinski, Bryce P. Nelson and Robert C. Getts- Protein detection enhanced by 3DNA de…
Protein expression monitoring can bea goo too to eterm ne sease onsetand progression, and is also used in drugdiscovery and investigational science (1–4).However, common y use assays ac t esensitivity and specificity required for thedetection of low-abundance proteins (5,6).Severa s gna amp cat on tec no og eshave been designed and applied to proteindetection assays, including rolling-circleamp cat on (RCA) ( ,7), tyram e s gnaamplification (TSA) (8), and enzyme-linkedimmunosorbent assay (ELISA) coupledw t PCR (9). A t oug a tec no og eshave successfully provided signal amplifi-cation to protein detection assays, none has
een w e y app e to m crot ter p ate ormicroarray ELISAs, presumably due to therequired additional steps and/or equipmentcompare w t t e stan ar assay.
DNA dendrimers are highly branchedstructures, constructed from seven different
single strands of DNA. Some strands arecomplementary to each other in the middlesection, and when incubated under the rightconditions they hybridize, forming “X”-likestructures called monomers. Monomersare the building blocks of the dendrimer.Different monomers hybridize through theirsingle-stranded regions to form the layers of the dendrimer (10).
The 3DNA dendrimer technology fromGenisphere (Hatfield, PA, USA) has beenestablished as a robust system to providesignal amplification for RNA detection(11–13). Here, we present an application
f the 3DNA dendrimer technology toprote n etect on. O gonuc eot e-con u-
ated antibodies (for selectivity) andnzymes or dyes (for signal) are hybridized
to t e outer arms o t e en r mer, w cfunctions as a core. Figures 1, A and B,
show a schematic representation of the endproduct, called UltraAmp (Genisphere),and its assembly. UltraAmp reagents areprepared by combining a stock solution of the dendrimer with solutions of the antibodyand enzyme-conjugated oligonucleotides.
he buffer is spiked with NaCl to a finaloncentration of 200 mM to enhance the
stability of the UltraAmp complex. After 20in incubation, the solution is diluted to 6
g/ μL as dendrimer with PBS and 50 % v/vfinal SuperFreeze (Pierce Biotechnology,Inc. Rockford, IL, USA).
Standard sandwich ELISAs were run asrecommended by the manufacturer (R&DSystems, Minneapolis, MN, USA). InUltraAmp assays, all standard conditionswere maintained, except for the incubation
with streptavidin horseradish peroxidase(SA-HRP; R&D Systems), which wasreplaced with UltraAmp Anti-biotin HRP(Cat. no. AB2080; Genisphere) dilutedin Binding Buffer I (Cat. no. ABB100;Genisphere) to a final dendrimer concen-trat on o 0. ng/ μL. Eac we wasincubated with 50 μL of diluted UltraAmpreagent for 1 h on a reciprocating shaker atapprox mate y 100 rpm at room temperature(RT). The wells were emptied, washed 5–6
Protein detection enhanced by 3DNAdendrimer signal amplification
Johanna R. Mora1, Tamara L. Zielinski
2, Bryce P. Nelson
2, and Robert C. Getts
1
1Genisphere, Inc. Hatfield, PA and 2GenTel BioSciences, Inc., Madison, WI, USA
DNA dendrimers, conjugated with both anti-biotin antibodies and up to 350 labeling enti-ties, were designed and adapted to protein microarray and enzyme-linked immunosorbent as-say (ELISA) to improve the limits of protein detection with no additional steps or equipment.
Application of conjugated dendrimers to standard ELISA cytokine detection resulted in up tothreefold improvement of the limits of detection with no significant increase in the inter- and intra-assay coefficient of variation (CV) compared to streptavidin horseradish peroxidase (SA-
HRP) detection. The adaptation of conjugated dendrimers to protein microarray cytokine de-
tection resulted in up to 10-fold improvement of the limits of detection, but assay conditionswould have to be optimized to decrease the intra- and inter-assay %CVs.
Figure 1. Schematic representation of the UltraAmp reagent assembly and output signal for thedetection of interleukin-1 β (IL-1 β). (A) Diagram of a four-layer 3DNA dendrimer. (B) Magnifiedrepresentation of two of the single-stranded arms (gray) in the outer layer of the dendrimer that allowttachment of dye-/HRP- and antibody-conjugated oligonucleotides by base-pairing interactions. Dye-nd HRP-labeled oligonucleotides are represented in solid black, and the antibody-conjugated oligonu-
cleotide is represented in dashed gray. Standard curves obtained for the detection of IL-1 β in sandwichELISAs (C) using standard ( ¢ , solid line) and UltraAmp ( l , dashed line) detection; and in proteinmicroarrays (D) using both standard ( p , dashed line) and UltraAmp ( n , solid line) detection.
C
8/3/2019 Johanna R. Mora, Tamara L. Zielinski, Bryce P. Nelson and Robert C. Getts- Protein detection enhanced by 3DNA de…
times with wash buffer (0.05 % Tween 20 inPBS), and blotted dry on paper towels. Thewells were then incubated with 100 μL of
tetramethylbenzidine (TMB) substrate atRT in the dark. The reaction was stoppedafter 20 min with 100 of 2N H SO 4 andthe absorbance read at 450 nm.
For interleukines (IL), 5- or 6-pointstandard curves were run: -1 β, -2, -4, -6,an -8. T e m ts o etect on (LOD) werecalculated as the average signal of the blank plus two standard deviations ( ) of the
an . F gure 1C s ows t e stan ar curvesfor detection of IL-1 β with UltraAmp andSA-HRP. UltraAmp detection providedapprox mate y t ree o s gna amp cat oncompared with SA-HRP. To determine intra-assay % V, four replicas at each concen-trat on were run per p ate. Two p ates wererun to determine inter-assay % V. Table 1summarizes the %CVs near the LOD withU traAmp etect on an compares LODbetween UltraAmp and SA-HRP. Up tothreefold improvement in the LOD waso serve w en us ng U traAmp comparewith standard detection, but no significantimprovement in the LOD for IL-2 waso serve . W e we ave not eterm nethe source of variation in the signal ampli-fication among cytokines, we hypothesize
that the variation may be due to the intrinsiccharacteristics of each antibody-antigen-antibody complex. Hughes and colleagues(14) showed that for the reaction betweenerythrocytes and anti-D antibodies there wasa 1000-fold increase in the rate of associationwhen the ionic strength was reduced form0.17 to 0.03. Therefore, it is possible that thebuffer used here with the UltraAmp reagentsin ELISAs (ionic strength ≥ .2) affects theassociation rate of some antibody-antigen-antibody complexes. Additional experi-ments were performed (data not shown) toconfirm that the presence of the dendrimer
did not sterically hinder interaction betweenthe anti-biotin antibody and the biotinylatedsecondary antibody.
Given that protein microarrays allowanalysis of multiple antigens with smallersample consumption than ELISAs,UltraAmp reagents were also tested in thisplatform. Here, we used ultra-thin nitro-cellulose protein microarray slides fromGenTe B oSc ences (GenTe , Ma son,WI, USA) because of their reported highsensitivity and high precision in quantitativemu t p ex mmunoassays (QMI).
apture antibodies to IL-1 β, IL-2, IL-4, IL-6, and IL-8 (R&D Systems), twopos t ve contro s an a negat ve controwere printed at 0.5 mg/mL in GenTel ArrayBuffer using a GeSim Nanoplotter 2.0e(GeS m, Gro er manns or , Germany) ata relative humidity of 60 %. Six replicatesof each antibody were printed in each of 1 su arrays on eac s e. S es wereassembled in a SIMplex 16-Multiplexingsystem (GenTel) and blocked with 200 μLo GenTe B oc u er n eac we or 1h at RT.
Recombinant human cytokines (R&DSystems) were ser a - ute n GenTewash buffer for a total of seven concentra-tions. Standard curve concentration ranges
were optimized for each detection methodand ranged from 0.32 to 5,000 pg/mLfor standard detection and 0.064 to 1000pg/mL for UltraAmp detection. A blank consisting of wash buffer was also prepared.Immediately following blocking, 100 μL of antigen standard solution was added to eachof the 16 subarrays on a single slide andincubated for 1 h at RT.
Biotinylated detector antibodies (R&DSystems) were diluted in GenTel washbuffer at the following concentrations:IL-1β 150 ng/mL, IL-2 1000 ng/mL, IL-4
00 ng/mL, IL-6 350 ng/mL, and IL-8 100
ng/mL. After antigen binding, the slideswere washed three times with GenTel washbuffer. One hundred microliters of detector
antibody solution was added to each of the16 subarrays for each slide and incubatedfor 1 h at RT. After incubation with detectorantibody, the slides were washed three timeswith GenTel wash buffer. For standard
etection, 100 L of a 1 /mL Dy547-Streptav n (Dyom cs, Jena, Germany)
iluted in GenTel rinse buffer was added toach well and incubated for 1 h at RT. For
U traAmp etect on, en r mer was utein PBS with 2 % BSA and 2 % dextransulfate to a concentration of 1 ng/ μL. Fiftym crograms o en r mer was a e toach well and incubated for 2 h at RT. Slides
were removed from the SIMplex-16 units,r nse w t GenTe r nse u er, an own
ry with nitrogen gas. Arrays were thenscanned using a Tecan LS Reloaded laserscanner (Tecan, Männen or , Sw tzer an )at gain 130 for standard detection and 120for UltraAmp detection.
Seven-po nt stan ar curves were run orall cytokines. Figure 1D shows two standardurves comparing detection with Dy547-
Streptav n an U traAmp. U traAmpprovided fivefold signal amplification. To
etermine intra-assay %CV, two replicas of
ach standard were run per slide, allowingfor data comparison among spots and wells.Three slides were run to determine inter-assay % V. The LOD was calculated as theaverage signal of the blank plus two of the blank. Results are summarized in Table1. A significant increase in the inter-assay%CVs was observed for the UltraAmp
etection. Since the %CV among spotswas considerably smaller, the increasedinter-assay % V may have been due towash inconsistency (number of washes
r strength) among slides. It is likely thatthe wash procedure can be optimized to
Table 1. Summary of %CV Range Observed near the LOD in the Analysis of Cytokines by ELISA and Protein Microarray withDendrimer Detection; and Comparison of LOD between Standard and Dendrimer Detection
Assay ELISA Protein Microarray
ntigen % V
Intra-assay(n = 4)
%CV
Inter-assay(n = 2)
Unamplified
LODa
(pg/mL)
Amplified
LOD (pg/mL)
% V Spot-
to-spot(n = 6)
% V Well-
to-well(n = 6)
% V
Inter-ssay
(n = 3)
Unamplified
LOD(pg/mL)
Amplified
LOD(pg/mL)
IL-1β 7.2 0 13.0 7.00 .1 18.8 .6 1.27 0.14
IL-2 3.6 9.2 58.43 60.32 .8 15.0 8.3 11.30 0.92
IL-4 1.0 1.7 4.77 1.42 .9 12.8 4.9 4.98 0.56
IL-6 2.2 0.8 14.12 6.56 .3 19.6 9.6 7.55 0.96
LOD, limits of detection.
8/3/2019 Johanna R. Mora, Tamara L. Zielinski, Bryce P. Nelson and Robert C. Getts- Protein detection enhanced by 3DNA de…
achieve more acceptable CVs for proteindetection assays. Also, preliminary experi-ments indicate that doubling the volume of UltraAmp solution used in the assay resultsin slightly tighter CVs.
By adapting the use of UltraAmpreagents to cytokine detection using proteinmicroarrays, we achieved LOD below thereported mean values for those cytokinesin serum and better than the reported valuesachieved with RCA protein microarraydetection (6). Table 1 also shows betterLOD achieved with protein microarray thanwith microtiter plate detection. This may beexplained by the Ekins’ theory (15,16). Eventhough capture antibodies are present inlower amounts in protein microarrays thanin microtiter plates, the coating density is
000 to 6000 × larger. This allows the proteinmicroarray to function as a concentration-sensing device if the number of capturedmolecules is <0.1/K ffinity . The higher coatingdensity of capture antibodies might alsobe beneficial to UltraAmp detection. Onecan ypot es ze t at t e U traAmp reagent
needs to anchor itself through the biotins of nearly located secondary antibodies. In ahigh density–coated microspot, even at low-antigen concentrations, the targets may belose enough for UltraAmp to anchor itself.
This would explain why the application of UltraAmp reagents to protein microarrayassays gave better signal amplification(fivefold) than their application to ELISAs(threefold). In the ELISA, at the same low-antigen concentration, the targets may be toospread out, making it difficult for UltraAmpto attach itself through to biotins from
ifferent targets. A more simple explanationmight be related to the number of labelsattached per dendrimer. Dendrimers usedfor protein detection had 350 fluorophores,while reagents used in ELISAs had 200
HRPs per dendrimer. Further, the averagesize of the dendrimers used for fluorescent
etection were significantly smaller thanthose attached to the larger HRP molecules,thus limiting the packing density of
endrimers for detecting neighboring biotinmo ecu es.
In conclusion, UltraAmp reagentsprovide signal amplification to a variety
f cytokine-detection ELISAs with nosignificant increase in inter- and intra-assay % Vs and without additional steps
r equipment. Even though the UltraAmpreagents seem to provide greater signalamplification to cytokine detection usingprotein microarrays, further optimization isrequired to improve the intra-assay % Vs.
ACKNOWLEDGMENTS
The authors would like to thank Paul W. Rhyne for insightful comments regarding the findings presented here and James Kadushin
nd Lori Getts for the manufacturing of na-
ed and fluorophore-labeled dendrimers.
COMPETING INTERESTSSTATEMENT
J.R.M. declares no competing interests. R.G. works at Genisphere, the company
Focus on miRNAs that are most relevantto your study.FlexmiR Select miRNA arrays are the perfectsolution for application-speci c assays such astissue-speci c studies, pathway-speci c screening,nd biomarker validation.
Multiplex 2-50 miRNA targets of interestin a single wellxMAP® technology and simple FlexmiR work owprovide accurate results in 4-5 hours without samplemanipulation or bias. Quantitative data can bettained with standard curves.
����� sa -le-l -t-7-����sa --- ---���� - -7-���sas ---- -----���s -- -����� -- ---����� ss --- --1����hs --- ---1- a
s a---- ---���� ---m -- a��� - i-- ---- 3b���� -- ii --14 --����s - -���� -- -1--15b
���sa - --����a-m --- *����hss -- i- -1--
hs - i -- a����s ---m 1--- 8a*��� - i - 8���� --miR--11 9����s -mi ----199 b��� s -- --- a0a����� a- i- -- 0����s -m --
that makes the UltraAmp reagents. T.Z. and B.N. work at GenTel Biosciences, the manu- facturer of protein microarrays. Both com- panies could potentially benefit from the publication of this manuscript.
REFERENCES
1. hambers, A.F. and B.C. Vanderhyden. 200 .Ovar an cancer omar ers n ur ne. C n ca Ca.Res. 12:323-327.
2. Rosalki, S.B., R. Roberts, H.A. Katus, E.Giannitsis, J.H. Ladenson , and F.S. Apple. 2004.Cardiac biomarkers for detection of myocardialinfarction: perspectives from past to present. Clin.Chem. 50:2205-2213.
3. Mathers, A.R. nd .F. Cuff. 2004. Role of inter-leukin-4 (IL-4) and IL-10 in serum immunoglobulinG antibody responses following mucosal or system-c reov rus n ect on. J. V ro . 78:3352-33 0.
4. Nguyen, K.T., T. Deak, S.M. wens, T. Kohno,M. Fleshner, L.R. Watkins, and S.F. Maier. 1998.Exposure to acute stress induces brain interleukin-1β protein in the rat. J. Neurosci. 18:2239-2246.
5. Gross, M., I. Top, I. Laux, J.Katz, J.Curran, C.Tindell, and D. Agus. 2007. β 2-microglobulin isan androgen-regulated secreted protein elevated inerum of patients with advanced prostate cancer.
Clin. Cancer Res. 13:1979-1986.
6. Shao, W., Z. Zhou, I. Laroche, H. Lu, . Zong,D.D. Patel, S. Kingsmore, and S.P. Piccoli. 2003.
ptimization of rolling-circle amplified proteinmicroarrays for multiplexed protein profiling. J.Biomed. Biotechnol. 003 :299-307.
7. Schweitzer, B., S. Wiltshire, J. Lambert, S.
’Malley, K. Kukanskis,Z. hu,S.F.Kingsmore,P.M. Lizardi, et al. 2000. Immunoassays w t ro -ng crc e DNA amp cat on: a versat e pat orm
for ultrasensitive antigen detection. Proc. Natl.Acad. Sci. USA 97 :10113-10119.
8. Varnum, S.M., M.G. Warner, B. Dockendorff,N.C. Anheier, J. Lou, J.D. Marks, L.A. Smith ,J.D. Feldhaus , et al. 2006. Enzyme-amplified pro-tein microarray and a fluidic renewable surface fluo-rescence immunoassay for botulinum neurotoxin
etection using high-affinity recombinant antibod-es. Ana. C m. Acta 570 :137-143.
9. Sano, T., .L. Smith, and .R. antor. 1992.Immuno-PCR: very sensitive antigen detectionby means of specific antibody-DNA conjugates.Science 258 :120-122.
10. Nilsen, T.W., J. rayzel, and W. Prensky. 1997.Dendritic nucleic acid structures. J. Theor. Biol.187 :273-284.
11. Stears, R.L., R.C. Getts, and S.R. Gullans. 2000.A novel, sensitive detection system for high-densitymicroarrays using dendrimer technology. Physiol.Genom cs 3:93-99.
12. Goff, L.A., M. Yang, J. Bowers,R.C. Getts, R.W.Padgett, and R.P. Hart. 2005. Rational probe op-
timization and enhanced detection strategy for mi-croRNAs using microarrays. RNA Biol. 2:e9-e16.
13. Mora, J.R. nd R.C. etts. 2006. Enzymatic mi-croRNA detection in microtiter plates with DNAdendrimers. BioTechniques 41:420-424.
14. Hughes-Jones, N.C., B. Gardner, and R.
Telford. 19 4. T e e ect o pH an on c strengton t e react on etween ant -D an eryt rocytes.Immuno ogy 7 :72-81.
15. Saviranta, P., R. Okon, A. Brinker, M.Warashina, J. Eppinger, and B.H. eierstanger.2004. Evaluating sandwich immunoassays in micro-array format in terms of the ambient analyte regime.Clin. Chem. 50:1907-1920.
16. Ekins, R.P. 1998. Ligand assays: from electro-phoresis to miniaturized microarrays. Clin. Chem.44:2015-2030.
Received 5 November 2007; accepted 17December 2007.
Address correspondence to Johanna R. Mora, Bristol-Myers Squibb, P.O. Box 4000,Princeton, NJ 08543, USA. e-mail: [email protected]
