TATA DNA Sequence/Activity Correlation [Strahs & Schlick, 2000; Qian et al., 2001]

Our global bending framework and A-tract findings [Strahs & Schlick, 2000] helped test/refine the hypothesis that dynamic differences in DNA deformability are related to the transcriptional activity of the TATA-box binding protein (TBP) [Patikoglou et al., 1999]. An excellent example of single base-pair modulation of transcriptional activity emerges in the work of Steve Burley [Patikoglou et al., 1999], whose group demonstrated remarkable structural similarity, but different transcriptional activity, for 11 TBP/DNA complexes differing by one base pair (see table).

Label Sequence Efficiency (%)
WB GC    -31TATA·AAAG-24  GGCA 100  
A31 GC         AATA·AAAG       GGCA 14  
T30 GC         TTTA·AAAG       GGCA 25  
A29 GC         TAAA·AAAG       GGCA <1  
C29 GC         TACA·AAAG       GGCA 20  
G28 GC         TATG·AAAG       GGCA <1  
T28 GC         TATT·AAAG       GGCA 14  
T27 GC         TATA·TAAG       GGCA 35  
T26 GC         TATA·ATAG       GGCA 6  
G26 GC         TATA·AGAG       GGCA 18  
C25 GC         TATA·AACG       GGCA 6  
T25 GC         TATA·AATG       GGCA 100  
T24 GC         TATA·AAAT       GGCA 40  

Table 2 from [Qian et al., 2001]: Selected DNA sequences and their transcriptional efficiencies [Patikoglou et al., 1999]. The TATA octamers are flanked by GC on the 5' side and by GGCA on the 3' side. The adenovirus 2 major late promoter (AdMLP) TATA element sequence, a particularly efficient transcription promoter, serves as the control (or `wildtype') sequence (WB). Single position variants (red characters) are indicated relative to WB, and labeled according to the replaced base and position with respect to the transcription initiation site. Transcriptional efficiencies (TEs) for A29 and T24 are estimated [Qian et al., 2001].

Our analyses of dynamic trajectories of 13 TATA variants delineated systematic, sequence-dependent structural, energetic, and flexibility properties tailored to TBP interactions on a larger system of DNAs than studied to date [de Souza & Ornstein, 1998; Pastor et al., 1997; Pastor et al., 2000]. These factors important to activity include overall deformability, minor groove widening (with roll, rise, and shift increases) at TATA ends, untwisting within the TATA element (with large rolling at ends), and relatively low maximal water densities around the DNA (see figure). These factors work with the severe deformation induced by the minor-groove binding protein, which kinks the TATA element at the ends and displaces local waters to form stabilizing hydrophobic contacts [Burley & Roeder, 1996].

Though certain variants bend in directions consistent with their activity/inactivity (e.g., wildtype (`WB') and inactive A-tract (`A29') [Qian et al., 2001; Strahs & Schlick, 2000]), intrinsic bending is largely unrelated to activity. Our hypotheses on sequence/properties and their relevance to transcription assembly are gaining support in molecular dynamics simulations of the complexes between each of the 13 TATA variants and TBP.



Figure 4 from [Qian et al., 2001]. Illustrated are some of the factors correlated in our simulations with the transcriptional efficiency (TE) of various TATA elements: solvation, ion atmosphere, and local deformations.

A, top: Maximal water density for 13 TATA variants plotted against TE. A, bottom: Illustration of the water oxygen density for the four TATA variants A29, T28, T27, and WB drawn at a contour density of 0.075 molecules per Angstrom3. Note that the water density decreases with increasing TE. In the complex with TBP, the TATA element is almost completely desolvated; bound waters, represented by regions of higher density, are more difficult to remove and would destabilize complex formation.

B, top: This figure illustrates the probability of adopting bent, underwound structures. We calculated this probability by analyzing 300 structures sampled over 1.8 ns at a frequency of 6 ps. Specifically, we count snapshots that are underwound throughout the TATA element and are specifically bent at the first and last base pair step. Conformations of free DNA that are underwound and bent are potential substrates for TBP, since TBP unwinds the TATA element and inserts phenylalanines at the 5' and 3' ends to bend the DNA.

A TBP/TATA complex (PDB code 1CDW) is illustrated in the center with the phenylalanines intercalated at the 5' and 3' ends indicated with blue and green CPK models, respectively. Examples of C25 and T25 (T25 is bent and unwound) are shown at the right side of the figure with blue and green triangles indicating the 5' and 3' ends, respectively, of the TATA element (red).

B, bottom: The average number of ion contacts per snapshot within a 5.5 Angstrom radius of major groove base atoms is plotted against TE. Note that the number of ion contacts seems to describe a Gaussian distribution, suggesting an optimal ion density within the major groove associated with transciptional activity. This may be due to the electrostatic potential that develops around the TBP complex; this potential, illustrated in yellow at a contour of -5kBT/e, indicates that regions of higher potential develop near the bent regions at the ends of the TATA DNA. The right side of the figure illustrates specific regions of ion density at a contour of 0.02 molecules per Angstrom3 relative to the average structures of T25 and T30.




Relevant Links






Click to go back to Research