Protein Environment and DNA Orientation Affect Protein-Induced Cy3 Fluorescence Enhancement

Abstract

The cyanine dye Cy3 is a popular fluorophore used to probe the binding of proteins to nucleic acids as well as their conformational transitions. Nucleic acids labeled only with Cy3 can often be used to monitor interactions with unlabeled proteins because of an enhancement of Cy3 fluorescence intensity that results when the protein contacts Cy3, a property sometimes referred to as protein-induced fluorescence enhancement (PIFE). Although Cy3 fluorescence is enhanced upon contacting most proteins, we show here in studies of human replication protein A and Escherichia coli single-stranded DNA binding protein that the magnitude of the Cy3 enhancement is dependent on both the protein as well as the orientation of the protein with respect to the Cy3 label on the DNA. This difference in PIFE is due entirely to differences in the final protein-DNA complex. We also show that the origin of PIFE is the longer fluorescence lifetime induced by the local protein environment. These results indicate that PIFE is not a through space distance-dependent phenomenon but requires a direct interaction of Cy3 with the protein, and the magnitude of the effect is influenced by the region of the protein contacting Cy3. Hence, use of the Cy3 PIFE effect for quantitative studies may require careful calibration.

Significance

Protein-induced fluorescence enhancement (PIFE) of Cy3-labeled DNA is now a common means used to monitor protein-DNA interactions. In studies of two different proteins interacting with both 3′- and 5′-end labeled single-stranded DNA, we show that the Cy3 PIFE effect requires a direct interaction with the protein, and the magnitude of the effect depends on both the region of the protein that interacts with the Cy3 and the position of the label on the DNA. As such, the quantitative use of PIFE requires careful calibration of the fluorescence signal change.

Introduction

Quantitative studies of protein-nucleic acid interactions can be performed using a variety of techniques. Approaches that monitor spectroscopic changes are particularly convenient and can be used to obtain rigorous equilibrium binding isotherms when the signal is analyzed appropriately (1, 2, 3, 4, 5, 6, 7). Changes in intrinsic protein fluorescence from either tyrosine or tryptophan can often be used to monitor the interactions of proteins with nucleic acids (8). However, it is now common and straightforward to label a nucleic acid with an extrinsic fluorophore to monitor binding via changes in fluorescence intensity, lifetimes, or anisotropy. The fluorescence intensity can be quenched or enhanced depending on the fluorophore and the interaction between the biomolecules. We have shown previously that on the interaction with Eshcerichia coli single-stranded DNA (ssDNA) binding (SSB) protein or E. coli UvrD, the fluorescence intensity of ssDNA labeled with fluorescein is quenched, whereas that of the cyanine dye Cy3 is enhanced (5, 9, 10). Cy3 can be used as a fluorescence donor in Förster resonance energy transfer (FRET) experiments with an acceptor such as Cy5 (11). However, nucleic acids labeled only with Cy3 can also be used to monitor interactions with proteins because of the enhancement of Cy3 fluorescence intensity that often results when Cy3 interacts with a protein. This phenomenon was first shown in a study of the E. coli SSB protein binding to a single-stranded nucleic acid labeled with Cy3 (9). This fluorescence enhancement has been routinely used to monitor the arrival of nucleic acid motor proteins at a site labeled with Cy3 (10, 12, 13, 14, 15, 16, 17, 18, 19, 20) as well as protein conformational changes (21). This phenomenon is now commonly referred to as protein-induced fluorescence enhancement (PIFE) (22, 23, 24).

PIFE effects can complicate FRET studies involving Cy3 but have also been used in conjunction with and to complement FRET studies (25, 26, 27). The excited state of Cy3 can exist as either a cis or trans isomer, with only the trans form having a significant fluorescence quantum yield (28, 29, 30) (Fig. 1 A). Hence, any process that shifts the Cy3 conformational state to favor the trans conformation will result in a Cy3 fluorescence enhancement (23, 24). It has been proposed that the interaction of Cy3 with a protein can sterically hinder the Cy3 isomerization favoring the trans form and thereby enhance its fluorescence intensity (24, 25).

Figure 1.

Chemical structures of Cy3 and Cy3-DNA. (A) Structure of Cy3 in the trans and cis conformations is shown. The trans form has a much higher quantum yield than the cis form. (B) Cy3 is labeled at the 3′ end (3′-Cy3) of (dT)_n. (C) Cy3 is labeled at the 5′ end (5′-Cy3) of (dT)_n. To see this figure in color, go online.

We had shown previously that when human RPA (hRPA) binds to a stretch of 30 oligodeoxythymidylates labeled at its 3′ end with Cy3, 3′-Cy3-(dT)₃₀, the Cy3 fluorescence is enhanced by 72%, and we used this effect to monitor the diffusion of hRPA along poly(dT) (20). Here, we show that the fluorescence enhancement of Cy3 by hRPA and E. coli SSB proteins differs dramatically when the Cy3 is positioned at the 3′ end versus the 5′ end of the ssDNA. The spectral properties of the isolated DNA samples are identical, hence the differential Cy3 fluorescence enhancement is due to effects on the fluorophore within the protein-DNA complex. We also show that this environmentally induced effect increases the Cy3 fluorescence lifetime and influences its rotational freedom. Thus, the magnitude of the Cy3 PIFE effect is dependent on the amino acid environment that contacts Cy3.

Materials and Methods

All oligonucleotides were synthesized using a MerMade 4 DNA Synthesizer (BioAutomation, Plano, TX) with reagents purchased from Glen Research (Sterling, VA). The Cy3 fluorophore was incorporated into the DNA during synthesis. We note that the phosphoramidite available from Glen Research is not sulfonated, as indicated in Fig. 1, because the sulfonated Cy3 is not stable as a phosphoramidite. The distance between Cy3 and the hRPA was systematically varied using a series of oligodeoxythymidylates containing a high-affinity site of (dT)₃₀ flanked by an 18-basepair duplex segment and a stretch of thymidylates in which the backbone polarity connecting each nucleotide alternated between a 3′-3′ and a 5′-5′ linkage. A stretch of alternating polarity (dT)₃₀ has a 450-fold lower affinity for hRPA than a normal (dT)₃₀ (20). Thus, the hRPA will bind exclusively to the normal (dT)₃₀ stretch of DNA as previously described (20). The distance between the high-affinity hRPA binding site and Cy3 was systematically increased by lengthening the alternating polarity segment. Two sets of DNA were made in which the alternating polarity (dT)_N and the duplex region were switched so that the Cy3 was positioned either at the 3′ end or the 5′ end of the high-affinity binding site. The dependence of the Cy3 fluorescence enhancement upon binding one hRPA, E (%), on the number of alternating polarity nucleotides, N, was fitted to the single exponential Eq. 1 to obtain a characteristic length N_c, as described previously (20).

$$\mathrm{E = A\ast }{\mathrm{e}}^{\mathrm{(-N/Nc)}}\mathrm{+ C}\mathrm{.}$$ (1)

The DNA oligomers were purified as described (31). Buffer T is 10 mM Tris (pH 8.1), 1 mM 2-mercaptoethanol, and 0.1 mM Na₂EDTA, with the indicated [NaCl]. hRPA was purified as described (20). E. coli SSB protein was purified as described (32). Fluorescence titrations and analysis were performed as described (5, 20). Briefly, 1.9 mL of 100 nM DNA (Cy3-labeled (dT)₆₈ or (dT)₃₅) was titrated with E. coli SSB in buffer T plus 10 mM or 300 mM NaCl at 25°C in a PTI fluorometer (Photon Technology International, Birmingham, NJ). The Cy3-labeled DNAs were excited at 515 nm (2-nm excitation bandpass), and fluorescence emission was monitored at 570 nm (5-nm emission bandpass). The hRPA titrations were conducted in buffer T plus 500 mM NaCl at 25°C. The DNAs were 95–97% labeled with Cy3. The concentrations of Cy3 and oligomers (dT)_n were determined by absorbance using the following extinction coefficients: ε_Cy3,548 = 136,000 M⁻¹ cm⁻¹, ε_Cy3,260 = 4900 M⁻¹ cm⁻¹, and ε_DNA,260 = 8100 M⁻¹ cm⁻¹ (in nucleotide). Protein concentrations were determined by absorbance using extinction coefficients of ε₂₈₀ = 1.13 × 10⁵ M⁻¹ (tetramer) cm⁻¹ for SSB and ε₂₇₇ = 8.57 × 10⁴ M⁻¹ cm⁻¹ for hRPA.

Fluorescence lifetime and time-resolved fluorescence anisotropy measurements were carried out using the time-correlated single photon counting technique. Samples for time-correlated single photon counting measurements contained 0.5 μM DNA and 1 μM hRPA in buffer T plus 500 mM NaCl at 25°C. The excitation source was a Fianium SC450 supercontinuum laser with a 6 ps pulse width operating at 20 MHz. An acousto-optical tunable filter was used for wavelength selection. The output of the acousto-optical tunable filter was vertically polarized before entering the sample, and fluorescence emission was recorded at 54.7° (the magic angle) with respect to excitation for lifetime measurements and at 0 and 90° with respect to excitation (I_VV and I_VH, respectively) for determining the anisotropy. Emitted photons were detected using a double-grating Jobin Yvon, Gemini-180 monochromator (HORIBA, Kyoto, Japan), and a Hamamatsu microchannel plate photomultiplier tube (R3809U-50) (Shizuoka, Japan). Data was processed using a Becker-Hickl single photon counting card (SPC-830) (Berlin, Germany). The instrument response function (IRF) (FWHM ∼50–60 ps) was measured from the scattering signal of a Ludox (colloidal silica in water) suspension at the excitation wavelength (532 nm). Samples were excited at 532 nm, and emission was detected at 570 nm. The appropriate interference filters for excitation and long band pass filters for the emission were used. Fluorescence intensity decays were fitted with a sum of exponentials using the IRF for deconvolution. Fitting was performed using ASUFIT, an in-house written code running in MATLAB (The MathWorks, Natick, MA). The fluorescence anisotropy decays were calculated as r(t) = (I_VV − I_VHG)/(I_VV + 2I_VHG). The G-factor was determined via the tail-matching method using a solution of Cy3B in water.

Results

E. coli SSB protein binding to Cy3-labeled ssDNA shows different Cy3 fluorescence enhancements depending on which DNA end is labeled

E. coli SSB protein is a stable tetramer and an essential protein that binds tightly to ssDNA intermediates during replication, recombination, and repair (33, 34, 35). We have previously shown that when E. coli SSB tetramers bind to oligodeoxythymidylates labeled on the 3′ end with Cy3, as depicted in Fig. 1 B, the Cy3 fluorescence is enhanced (9). Here, we compare the Cy3 fluorescence enhancement when SSB binds to oligodeoxythymidylates labeled on the 5′ end with Cy3 (Fig. 1 C). Fig. 2 shows the results of a series of experiments in which SSB is titrated into solutions containing either (dT)₆₈ or (dT)₃₅ that were labeled with Cy3 on either the 3′ or 5′ end as depicted in Fig. 1. Titrations were performed in buffer T plus 10 mM NaCl or 300 mM NaCl at 25°C. At high salt (300 mM NaCl), SSB binds in its fully wrapped (SSB)₆₅ mode, in which the ssDNA interacts with all four subunits of the tetramer (36, 37, 38), wrapping with a topology resembling the stitching on a baseball (39). At low salt (10 mM NaCl), SSB binds in a partially wrapped (SSB)₃₅ mode in which an average of only two subunits interact with the ssDNA (36, 37, 38). Hence, at 10 mM NaCl, two SSB tetramers can bind to (dT)₆₈ at saturation (Fig. 2 A), whereas at 300 mM NaCl, only one SSB tetramer will bind to (dT)₆₈ (Fig. 2 B). However, at both 10 and 300 mM NaCl, the enhancement of Cy3 fluorescence is greater when the Cy3 label is on the 3′ end of the (dT)₆₈ (Fig. 2; Table S1). At 300 mM NaCl, we observe 81% vs. 21% Cy3 enhancement for 3′-Cy3 versus 5′-Cy3, respectively, and at 10 mM NaCl, we find 114% vs. 71% for 3′-Cy3 versus 5′-Cy3, respectively. This enhancement differential is also observed for SSB binding to the shorter (dT)₃₅, although there is a quantitative difference. At excess SSB over (dT)₃₅, which populates a single (dT)₃₅ bound per tetramer, the Cy3 fluorescence shows a greater enhancement when the Cy3 label is on the 3′ end (123% at 10 mM NaCl and 98% at 300 mM NaCl) versus the 5′ end (63% at 10 mM NaCl and 24% at 300 mM NaCl). Interestingly, in all cases, the Cy3 enhancement is greater at 10 mM NaCl than at 300 mM NaCl (Table S1). As can be seen from Fig. 1, B and C, the chemical composition of the Cy3-containing DNA is the same regardless of which end is labeled. Furthermore, the Cy3 fluorescence spectra and quantum yield of the Cy3 is independent of the labeling position. The emission spectra of each Cy3-labeled DNA in the absence and presence of SSB are shown in Fig. S1.

Figure 2.

Fluorescence enhancements of Cy3-labeled ssDNA upon E. coli SSB protein binding. (A) Shown is the binding of SSB to (dT)₆₈ with the Cy3 on the 3′ end (red) or 5′ end (blue) in buffer T plus 10 mM NaCl at 25°C, in which two SSB tetramers bind in the (SSB)₃₅ mode. (B) Shown is the binding of SSB to (dT)₆₈ with the Cy3 on the 3′ end (red) or 5′ end (blue) in buffer T plus 300 mM NaCl at 25°C, in which one SSB tetramer binds in the (SSB)₆₅ mode. (C) Shown is the binding of SSB to (dT)₃₅ with the Cy3 on the 3′ end (red) or 5′ end (blue) in buffer T plus 10 mM NaCl at 25°C, in which one SSB tetramer binds in the (SSB)₃₅ mode. (D) Shown is the binding of SSB to (dT)₃₅ with the Cy3 on the 3′ end (red) or 5′ end (blue) in buffer T plus 300 mM NaCl at 25°C, in which one SSB tetramer binds in the (SSB)₃₅ mode. The cartoons illustrate the complexes at the end of the titrations. A summary of the percents of enhancement is shown in Table S1. To see this figure in color, go online.

hRPA binding to Cy3-labeled ssDNA also shows differential Cy3 fluorescence enhancement depending on which DNA end is labeled

The eukaryotic analog of E. coli SSB protein is the replication protein A (RPA) (40, 41). RPA is a heterotrimer consisting of Rpa1 (∼70 kDa), Rpa2 (∼32 kDa), and Rpa3 (∼14 kDa). All three subunits contain oligonucleotide/oligosaccharide binding (OB) folds (42), with Rpa1 containing 4 OB folds (F, A, B, and C), whereas one additional OB fold is contained in each of Rpa2 (D) and Rpa3 (E) (see Fig. 3 A). OB folds A, B, C, and D function in ssDNA binding (43, 44) as shown in a crystal structure of Ustilago maydis RPA bound to ssDNA (dT)₃₂ (Fig. 3 A), in which 25 of the DNA nucleotides are observable (45). The 5′ end of the ssDNA interacts with RPA1, whereas the 3′ end interacts with RPA2/RPA3.

Figure 3.

Fluorescence enhancement of Cy3-labeled ssDNA upon hRPA binding. (A) A crystal structure of U. maydis RPA bound to (dT)₃₂ is shown (45). (B) Human RPA, binding to 3′-Cy3-(dT)₃₀ (red) and (dT)₃₀-Cy3-5′ (blue), is shown. The experiment was conducted in 10 mM Tris (pH 8.1), 0.5 M NaCl, and 1 mM β-mercaptoethanol at 25°C. To see this figure in color, go online.

RPA binding to 3′-Cy3-labeled ssDNA shows a significant Cy3 PIFE effect, and we had used this effect in a single molecule total internal reflectance fluorescence experiment to monitor hRPA diffusion along ssDNA (20). Fig. 3 B shows hRPA titrations of (dT)₃₀ labeled on either its 3′ end or 5′ end with Cy3. These titrations show that upon formation of a 1:1 molar complex of hRPA with (dT)₃₀, the Cy3 enhancement is 77% when Cy3 is on the 3′ end, but only ∼12% when Cy3 is on the 5′ end. This difference is qualitatively similar to that observed with E. coli SSB protein. Again, the 3′-Cy3-(dT)₃₀ and the 5′-Cy3-(dT)₃₀ show identical fluorescence spectra and quantum yields. Because hRPA binds to ssDNA with strict polarity, the Cy3 on the 3′ or 5′ end of the ssDNA interacts with different regions of the hRPA. The Cy3 on the 3′ end of the ssDNA interacts with RPA2/RPA3 subunits, whereas the Cy3 on the 5′ end interacts with the RPA1 subunit (Fig. 3 B). These local environmental differences induce different Cy3 fluorescence intensities in the hRPA-(dT)₃₀ complex.

hRPA induces different Cy3 fluorescence lifetimes depending on the Cy3 position on the DNA

Because the average fluorescence intensity is proportional to the fluorescence quantum yield and therefore proportional to the fluorescence lifetime, the above results suggest that the fluorescence lifetime of the Cy3 at the 3′ end may be larger than that for the Cy3 at the 5′ end. Therefore, we next examined the fluorescence lifetimes and fluorescence anisotropy decay using time-resolved fluorescence. As illustrated in Fig. 4, (dT)₃₀ with Cy3 on the 3′ or the 5′ end shows identical decay properties with the same lifetimes in the absence of hRPA, indicating that the Cy3 fluorescence properties are unaffected by which end is labeled. Upon saturation with hRPA, the amplitude of the short lifetime (∼300 ps) decreases at the expense of an increase in the amplitude of the long lifetime (∼2 ns), indicating a reduction in the fraction of Cy3 molecules that can isomerize efficiently. However, the 3′-Cy3-(dT)₃₀ sample exhibits a longer lifetime compared to (dT)₃₀-Cy3-5′ (Table S2), indicating a greater restriction of photoisomerization in the first case. The average fluorescence lifetime is doubled for the 3′-Cy3-(dT)₃₀, whereas there is a more modest increase (10%) for the (dT)₃₀-Cy3-5′. These results are consistent with those in Fig. 3 B, confirming a higher intensity and a longer fluorescence lifetime for the 3′-Cy3-(dT)₃₀ when bound to hRPA.

Figure 4.

Time-resolved fluorescence lifetime measurements. The black line is the instrument response function (IRF). The fluorescence intensity of the free 3′-Cy3 DNA (red line) is the same as 5′-Cy3 (blue line). The Cy3 on the two oligonucleotides have the same fluorescence lifetimes in the absence of hRPA. Upon hRPA binding, although the lifetimes increase on both complexes, the 3′-Cy3 (red + square markers) has a longer lifetime than the 5′-Cy3 (blue + square makers). The average lifetimes are shown in Table S2. To see this figure in color, go online.

hRPA induces different Cy3 fluorescence anisotropy decays for 3′-Cy3 versus 5′-Cy3

To examine the effects of the local protein environment on the rotational dynamics of Cy3, time-resolved fluorescence anisotropy experiments were conducted. As shown in Fig. 5, 3′-Cy3-(dT)₃₀ and (dT)₃₀-Cy3-5′ display similar anisotropy decay curves. Both curves indicate that fluorescence emission is completely depolarized after ∼12 ns, indicating that the rotational mobility of the dye is not greatly hindered. However, upon binding hRPA, distinct differences were observed for the two DNA molecules. For both hRPA-DNA complexes, the Cy3 displays hindered rotation in the nanosecond timescale, with neither excited states returning to a random orientation on the timescale of the measurement. Thus, the rotational motion of the Cy3 in both DNA molecules is restricted upon hRPA binding, and light remains highly polarized at timescales much longer than the fluorescence lifetime. However, the restriction is higher for the 3′-Cy3-(dT)₃₀ end as compared to the (dT)₃₀-Cy3-5′ (Fig. 5).

Figure 5.

Time-resolved fluorescence anisotropy. In the absence of hRPA, the Cy3 on either end (3′-Cy3 (red) and 5′-Cy3 (blue)) has similar anisotropy decays and exhibits free rotation characteristics (able to return to zero or random orientation). On binding to hRPA, both Cy3 have longer decay times compared to those of unbound DNA and exhibit hindered rotation characteristics. The 3′-Cy3 (red + square markers) has a longer decay or less rotational freedom than that of 5′-Cy3 (blue + square markers). To see this figure in color, go online.

hRPA-induced Cy3 fluorescence enhancement depends on ssDNA length

We had previously shown that the Cy3 PIFE effect can be used to monitor the position of hRPA relative to a 3′-Cy3 label (20). This is because the ssDNA is flexible (46) and dynamic such that the Cy3 end of the DNA can loop back to interact with hRPA, and the probability of looping depends on the persistence length and contour length of the ssDNA (20). To quantitatively determine the relative hRPA position required calibration of the Cy3 fluorescence intensity as a function of the position of hRPA on the ssDNA relative to the Cy3 at the 3′ end. This calibration was performed by using a series of oligodeoxythymidylates containing a high-affinity hRPA site of (dT)₃₀ attached to thymidylate nucleotides that were connected by alternating 3′-3′ or 5′-5′ reversed polarity (RP) linkages. The flexibility of these linkages is the same as for normal poly(dT), but because hRPA binds with polarity to ssDNA, its affinity for the RP DNA is 450-fold weaker, hence the hRPA is constrained to bind exclusively to the normal (dT)₃₀ stretch of DNA (20). The distance between hRPA and Cy3 was systematically increased by lengthening the alternating polarity segment. As shown in Fig. 6, the quantitative extent of the Cy3 PIFE effect upon binding a single hRPA heterotrimer decreased in an exponential manner as the number of alternating polarity nucleotides (N) increased. Also shown in Fig. 6 A are data for DNA in which the Cy3 is placed on the 5′ end of the binding site. The individual titrations used to determine the data in Fig. 6 are shown in Fig. S2. In this case, a Cy3 fluorescence enhancement is still observed, and its amplitude decreases as the Cy3 is moved further away from the hRPA binding site. However, the fluorescence enhancement of Cy3 on the 5′ end is much smaller than when Cy3 is on the 3′ end of the binding site. The results were fit with a single exponential function (Eq. 1) showing a characteristic length of N_c = 17 ± 2 nucleotides for Cy3 on the 3′ end and N_c = 34 ± 7 nucleotides for Cy3 on the 5′ end (Fig. 6 A). The difference in the enhancement amplitudes indicates that the Cy3 local environment influences the Cy3 fluorescence properties. We also note in Fig. 6 B that the two calibration curves do not overlay. This indicates that the relative PIFE effect depends not only on the flexibility and persistence length of the RP-ssDNA but also on which DNA end is labeled.

Figure 6.

The Cy3 fluorescence enhancement induced by hRPA binding depends on the orientation and distance of the Cy3 fluorophore from the hRPA. (A) The two DNAs used possess a high affinity (dT)₃₀ binding site for hRPA attached to an 18-bp duplex on one end and a stretch of N-thymidylates with alternating backbone polarity either on the 3′ end (red) or the 5′ end (blue), which reduces hRPA affinity by ∼450-fold (20). A Cy3 fluorophore (green star) is situated at the end of the alternating polarity stretch of N nucleotides. The data with the N alternating polarity nucleotides on the 3′ end (red) are from Nguyen et al. (20). The solid lines are the best fits of the data to an exponential function (Eq. 1) (B) The normalized exponential fits are compared for both sets of data and show that a different exponential function is needed to describe each set of data (N_c = 17 ± 2 for the 3′-Cy3 (red) and N_c = 34 ± 7 for the 5′-Cy5 (blue)) (see Eq. 1). To see this figure in color, go online.

Discussion

It is well established that the fluorescence properties of fluorophores can be affected substantially upon interaction with biomolecules. There are many examples in which this phenomenon has been used to monitor the binding of proteins to DNA labeled with an extrinsic fluorophore. Protein binding to fluorescein-labeled DNA can result in the quenching of fluorescence (e.g., SSB (5, 10, 47)), enhancement (e.g., human DNA polymerase β (48)), or a complex combination of both quenching and enhancement (e.g., E. coli Rep (49)). The fluorescence of rhodamine red-labeled ssDNA is also quenched upon interaction with UvrD (10). More recently, protein interactions with Cy3-labeled DNA have been monitored by the enhancement of Cy3 fluorescence (9, 21, 47). The first report of Cy3 enhancement was for E. coli SSB binding to Cy3-labeled ssDNA (9). This effect has been used to investigate directional translocation of a number of motor proteins along ssDNA (10, 12, 18, 50, 51, 52, 53, 54) as well as the diffusion and chemomechanical pushing of SSB proteins along ssDNA (20, 55). Cy3 fluorescence enhancement is now commonly referred to as PIFE (22, 23, 24) and has been suggested to be sensitive to the through space distance separating the Cy3 from the protein (22, 23, 56).

In this report, through systematic studies of E. coli SSB and hRPA proteins binding to a series of Cy3-labeled single-stranded oligodeoxythymidylates, we show that the magnitude of the PIFE effect varies with both the protein and the end of the ssDNA to which the Cy3 is attached. The Cy3 PIFE effect differs not only for hRPA and E. coli SSB binding to a Cy3-labeled (dT)_L, but even SSB binding in its different binding modes influences the magnitude of the enhancement. In the absence of protein, the placement of Cy3 on either the 3′ end or the 5′ end of the DNA has no influence on the Cy3 fluorescence lifetime or anisotropy. Hence, the differences observed are due to differences in the Cy3 environment upon contacting the protein in the protein-DNA complex. The different intensities observed upon protein binding to ssDNA with Cy3 at the 3′ or 5′ ends arise from the different Cy3 fluorescence lifetimes and different constraints of Cy3 when in contact with the protein. Cy3 at the 3′ end shows a longer fluorescence lifetime and more restricted rotational freedom upon hRPA binding. This arises from a difference in the steric hindrance induced by the bound hRPA. Clearly, the enhancement amplitude is dependent on the amino acid environment contacting the Cy3, suggesting that the protein environment can modulate the rate of cis/trans photoisomerization and thus its fluorescence properties.

It has been suggested that the Cy3 PIFE effect is a through space distance-dependent phenomenon (22, 23). However, our results do not support this conclusion. The clearest example bearing on this question is the data for binding of SSB to 3′-Cy3-(dT)₆₈ versus (dT)₆₈-Cy3-5′ in its fully wrapped (SSB)₆₅ binding mode. In this case, as indicated in the cartoon accompanying Fig. 2 B, the two ends of (dT)₆₈, and thus the 5′-Cy3 and the 3′-Cy3, will be in approximately the same position in the complex with SSB, yet the PIFE effect differs by a factor of ∼4. This indicates that the Cy3 PIFE effect is dependent on the nature of the interactions established between the protein and the Cy3 fluorophore and is not a distance-dependent phenomenon. We also show that the Cy3 fluorescence can be enhanced upon protein binding to a site on the DNA that does not contain the Cy3 label if the Cy3 label is on a flexible region of the DNA, in this case ssDNA, that is able to loop back and directly contact the protein. In this case, relative quantitative distance information can be obtained but requires careful calibration of the PIFE signal.

Author Contributions

T.M.L., B.N., and A.G.K. designed the fluorescence binding studies. M.A.C. and M.L. designed the fluorescence lifetime studies. B.N. and A.G.K. performed the PIFE binding titrations. M.A.C. performed the fluorescence lifetime studies. B.N., T.M.L., and M.L. wrote the article.

Editor: Doug Barrick.