Protein expression monitoring can be a good tool to determine disease onset and progression, and is also used in drug discovery and investigational science (1,2,3,4). However, commonly used assays lack the sensitivity and specificity required for the detection of low-abundance proteins (5,6). Several signal amplification technologies have been designed and applied to protein detection assays, including rolling-circle amplification (RCA) (6,7), tyramide signal amplification (TSA) (8), and enzyme-linked immunosorbent assay (ELISA) coupled with PCR (9). Although all technologies have successfully provided signal amplification to protein detection assays, none has been widely applied to microtiter plate or microarray ELISAs, presumably due to the required additional steps and/or equipment compared with the standard assay.

3DNA 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 of the 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 (11,12,13). Here, we present an application of the 3DNA dendrimer technology to protein detection. Oligonucleotide-conjugated antibodies (for selectivity) and enzymes or dyes (for signal) are hybridized to the outer arms of the dendrimer, which functions as a core. Figure 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 of the dendrimer with solutions of the antibody and enzyme-conjugated oligonucleotides. The buffer is spiked with NaCl to a final concentration of 200 mM to enhance the stability of the UltraAmp complex. After 20 min incubation, the solution is diluted to 6 ng/µL as dendrimer with PBS and 50% v/v final SuperFreeze (Pierce Biotechnology, Inc. Rockford, IL, USA).

Figure 1. Schematic representation of the UltraAmp reagent assembly and output signal for the detection of interleukin-1β (IL-1β). (Click to enlarge)

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. ABB 100; Genisphere) to a final dendrimer concentration of 0.6 ng/µL. Each well was incubated with 50 µL of diluted UltraAmp reagent for 1 h on a reciprocating shaker at approximately 100 rpm at room temperature (RT). The wells were emptied, washed 5–6 times with wash buffer (0.05% Tween 20 in PBS), and blotted dry on paper towels. The wells were then incubated with 100 µL of tetramethylbenzidine (TMB) substrate at RT in the dark. The reaction was stopped after 20 min with 100 µL of 2N H₂SO₄ and the absorbance read at 450 nm.

For interleukines (IL), 5- or 6-point standard curves were run: -1β, -2, -4, -6, and -8. The limits of detection (LOD) were calculated as the average signal of the blank plus two standard deviations (sd) of the blank. Figure 1C shows the standard curves for detection of IL-1 β with UltraAmp and SA-HRP. UltraAmp detection provided approximately threefold signal amplification compared with SA-HRP. To determine intra-assay %CV, four replicas at each concentration were run per plate. Two plates were run to determine inter-assay %CV. Table 1 summarizes the %CVs near the LOD with UltraAmp detection and compares LOD between UltraAmp and SA-HRP. Up to threefold improvement in the LOD was observed when using UltraAmp compared with standard detection, but no significant improvement in the LOD for IL-2 was observed. While we have not determined the source of variation in the signal amplification among cytokines, we hypothesize that the variation may be due to the intrinsic characteristics of each antibody-antigen-antibody complex. Hughes and colleagues (14) showed that for the reaction between erythrocytes and anti-D antibodies there was a 1000-fold increase in the rate of association when the ionic strength was reduced form 0.17 to 0.03. Therefore, it is possible that the buffer used here with the UltraAmp reagents in ELISAs (ionic strength ≥0.2) affects the association rate of some antibody-antigen-antibody complexes. Additional experiments were performed (data not shown) to confirm that the presence of the dendrimer did not sterically hinder interaction between the anti-biotin antibody and the biotinylated secondary antibody.