Azobenzene and the actuation system

November 6, 2018

by Abhilasha "Abby" Dehankar

Goal: Our goal was to study the switchability of the photoresponsive, hybridized azobenzene-modified-deoxyribonucleic acid (azoDNA) without nanoparticles on our current optical setup. This will provide control experiment information on how the steric hindrance of the nanoparticle affects azoDNA functionality. AzoDNA consists of two complementary pairs of single stranded DNA (ssDNA), one of which is modified with azobenzene molecules. The azo-modified-ssDNA will be used to label gold nanoparticles (AuNPs) and the complementary unmodified-ssDNA will be used to label quantum dots (QDs). This will allow the formation of QD-azoDNA-AuNP conjugates upon DNA hybridization to double stranded DNA (dsDNA), which will yield quenching of QD fluorescence via a Forster Resonance Energy Transfer mechanism. As described in Figure 1, the QD-azoDNA-AuNP conjugates would undergo photoactivated reversible QD fluorescence quenching. When UV light is applied the azobenzene molecules switch from a trans to cis conformation, promoting dissociation of the dsDNA complexes. Thus, QDs are detached from AuNPs, restoring QD fluorescence. These photoswitchable QD-DNA-AuNP conjugates will allow us to employ QDs for STORM imaging because of their predicted and controllable stochastic on/off behavior.

Figure 1: QD-DNA-AuNP conjugate reversible photoswitching mechanism. QD fluorescence quenches
in proximity of AuNPs through a Forster resonance energy transfer (FRET) mechansims. As a result,
QD fluorescence is quenched when the DNA on each NP hybridizes to form QD-DNA-AuNP conjugates.
Materials are photo-responsive because UV wavelength excitation (300-400 nm) applied to QD-DNA-AuNP conjugates
induces trans-azobenzene isomerization to the cis form. This destabilizes the DNA duplex, actuating dehybridization
and re-establishing QD fluorescence. In contrast, visible light (wavelength>400 nm) reverses this process,
promoting trans isomerization, DNA hybridization, and QD fluorescence quenching.

Review of progress-to-date:

  1. We successfully transferred organic QDs purchased from NN-labs into the aqueous phase. We achieved this transfer by exchanging the non-polar ligands (Octadecylamine) on QD surface with polar phytochelatin-3 (PC3) peptides. The resulting QDs (PC3-QDs) are water soluble and terminated with carboxylate (-COOH) and amine (-NH2) groups on their surfaces.
  2. We successfully conjugated dibenzocyclooctyne (DBCO) to -NH2 terminations on the surface of PC3-QDs to form QD-DBCO conjugates using commonly-employed carbodiimide chemistry. Carbodiimide chemistry forms an amide (-CONH2-) bond between –COOH groups and -NH2 groups. This allowed us to employ “click” chemistry to successfully conjugate ssDNA to PC3-QDs through the attached DBCO group. To implement this approach, we mixed the QD-DBCO conjugates with azide (-N3) terminated ssDNA to permit strain promoted alkyne - azide cycloaddition via a click reaction, generating QD-ssDNA conjugates.
  3. In the previous report, we used a DNA origami platform to test the functionality of the ssDNA-QD and ssDNA-AuNP conjugates. For this, we mixed the nanoparticle (NP) – ssDNA conjugates with DNA origami hinges containing corresponding complementary ssDNA sequences only at a specific location on the hinge arm. Successful NP binding is observed only at a specific location on the hinge arms, indicating functional ssDNA-NP conjugates.

Monthly Update:

During the previous report cycle, we tested the functionality of both ssDNA-NP conjugates and observed that they could bind to the corresponding complementary ssDNA on a model DNA origami platform. Therefore, our next goal was to form QD-azoDNA-AuNP conjugates. To achieve this goal, it is extremely important to precisely determine the concentrations of each the individual component required. As mentioned in a prior report (Does DNA sequence matter during QD- DNA conjugation?), in our current protocol, we purify ssDNA-QD conjugates from excess, unconjugated ssDNA using a size-based separation method. Although successful in purification, this method causes significant loss and dilution of QD conjugates (~90-95%). Such a low yield for ssDNA-QDs makes it challenging to accurately determine the concentration of particles employed in our experiments. Therefore, I am currently investigating alternative purification techniques (e.g.,: gel separation, centrifugal filtration, etc.) that could result in higher yields.

Meanwhile, I decided to perform a preliminary test on our optical setup using photoswitchable, hybridized azoDNA without any NPs. Investigating the capabilities of our current setup for the actuation and detection of the photoswitchable azoDNA is important before we proceed with the studies on our QD-azoDNA-AuNP conjugates for two reasons:

  1. We need to confirm that the proposed azo-DNA duplexes are functional.
  2. We need to confirm that our optical analysis method can detect these changes.

To enable this study, I used a duplex consisting of ssDNA modified with a fluorophore (Cyanine 5, Cy5) at one end, and a corresponding, complementary, azobenzene-modified-ssDNA sequence modified with a quencher molecule (Q) at the other end. Thus, the duplex formed would generate a fluorophore-quencher pair separated by the length of the DNA employed (Cy5-azoDNA-Q), similar to the proposed mechanism for our QD-azoDNA-AuNP conjugates (Fig 2A). When the fluorophore (Cy5)-quencher (Q) pair is in close proximity, the fluorescence of the fluorophore should decrease through a mechanism called Forster Resonance Energy Transfer (FRET) (similar to Fig 1). For optimal FRET, the fluorescence spectra of the fluorophore should spectrally overlap completely with the absorbance spectra of the quencher. Therefore, for this study, we used a quencher with an absorbance spectra that completely overlaps the emission spectra of Cy5 (Fig 2B).

Figure 2: (A) Cy5-azoDNA-Q pair photoswitch mechanism, (B) Absorbance of Q and fluorescence of Cy5 showing maximum overlap

Next, the Cy5-azoDNA-Q pair was prepared by mixing Cy5-ssDNA with complementary Q-azo-ssDNA at an equivalent molar ratio (1:1) at a concentration of 125 nM each. Cy5-ssDNA fluorescence was measured by recording its fluorescence spectra (Excitation: 649 nm, Emission: 654-700 nm) before and after addition of the complementary Q-azo-ssDNA (Fig 3). As expected, we observed a significant quenching in the peak fluorescence intensity (~80%) on addition of the quencher. This suggests feasibility of our approach.

Figure 3: Fluorescence quenching of Cy5-ssDNA on formation of Cy5-azoDNA-Q pair.

These initial experiments were conducted on Cy5-azoDNA-Q pairs prepared in natural white/ visible light (400-700 nm). Thus, these data correspond to an azobenzene -trans isomerization and the “off” state. For this report, I focused only on the transition from “off” state to the “on” state. As stated earlier, this can be achieved by actuation in the UV-range (200-400 nm) that dehybridizes the Cy5-azoDNA-Q pair by isomerization of the incorporated azobenzene to a cis form that destabilizes the hybridized state. I used an external UV-lamp (actuation wavelength- 365 nm) as the actuation source and illuminated samples for 5 minutes. Dehybrization was detected by recording the fluorescence spectra of Cy5 (Excitation: 649 nm, Emission: 654-700 nm). We would expect an increase in the peak fluorescence of Cy5 upon UV actuation. However, we observed negligible changes in the peak fluorescence of Cy5 upon UV exposure, even for 10 consecutive minutes. This raises the question of whether the actuation source is powerful enough to drive azobenzene isomerization. Although, we have already modified our fluorometer with an external laser with the help of our Optics team at Georgia, the real question now is whether we need to invest in a new UV-laser to act as a more powerful actuation source.