Hoogsteen base pair coordinate frame

[lvelve0901]

Hi, xiangjun,

I hope you have a good time.

Recently, I am analyzing a 12-mer DNA (MD trajectory) with N1-methyladenine in ADE16 which adopt a syn conformation and form Hoogsteen base pair with THY9. When I use X3DNA to analyzing it, I find some parameters are kinda wired. I analyze the base pair parameters in 3-22, 4-21, 5-20, 6-19, 7-18, 8-17, 9-16, 10-15 bps and step and helical parameters in 3-4, 4-5, 5-6, 6-7, 7-8, 8-9. 9-10.

First, let's just look at the analysis of first snapshots in this MD simulation (A6me_50ps.pdb).
(1) Base pair parameters: There are clear differences in the 9-16 step (HG base pairs) like shear, stretch and especially opening.
(2) Step and helical parameters: There are some sort of difference in tilt, twist, inclination.

My first question is: Where are these difference come from. I know that you define a coordinate frame for idealized DNA before analyzing, but is this definition also applied to Hoogsteen base pair? Do we need to redefine a coordinate frame for Hoogsteen base pair when we want to analyzing base pair parameters (or just use the idealized coordinate frame)? Are the step and helical parameters also intrinsically influenced by the coordinate frame of Hoogsteen base pair so we should expect to see these difference?

I also attach the full MD trajectory analysis for your reference. Some parameters' error bar is very large like x-displacement, y-displacement, that could be due to some MD issue which I plan to double check that.

Many thanks,

Best,
Honglue

[xiangjun]

Hi Honglue,

You've touched on a common yet "complicated" issue on analyzing structures with non-cannoical base pairs (Hoogsteen in your case). 3DNA/DSSR takes the following technical details into consideration in deriving the base-pair and step parameters:

• Each nucleotide is specified by its base reference frame in 3D space. In 3DNA, you can get a list of the base reference frames by:
  

Code: [Select]

find_pair -s A6me_50ps.pdb
The output file "ref_frames.dat" contains contents like this:

Code: [Select]

   24 bases
...     1 C   # -:...1_:[CYT]C
  -13.1598     7.3845   -11.5069  # origin
    0.8356     0.4997     0.2281  # x-axis
   -0.1418    -0.2049     0.9684  # y-axis
    0.5307    -0.8416    -0.1003  # z-axis
...     2 G   # -:...2_:[GUA]G
  -10.7299     3.2207   -10.6198  # origin
    0.1231     0.8647     0.4870  # x-axis
   -0.7648    -0.2301     0.6018  # y-axis
    0.6324    -0.4465     0.6330  # z-axis
......


• Check the M+N vs M-N relative base-orientation in the pair. If M-N, as in Watson-Crick pairs, the y-and z-axes of N is reversed. If M+N, as in the Hoogsteen pair, the N base reference frame is unchanged.
• The six base-pair parameters are calculated following the CEHS scheme as detailed in $X3DNA/doc/tech-details.pdf. The derived base-pair reference frames are then used to calculate step and helical parameters (again, see $X3DNA/doc/tech-details.pdf).
• As such, the parameters of all base-pairs (canonical or otherwise) are calculated using exactly the same algorithm. As shown in the 3DNA NAR2003 and NP2008 papers, these parameters are defined rigorously to allow for an exact reconstruction of base pair geometry. For example, run the following command:
  
  

Code: [Select]

find_pair A6me_50ps.pdb | analyze
rebuild -atomic bp_step.par 3dna-rebuilt.pdb

You will see that "3dna-rebuilt.pdb" file can be superposed to "A6me_50ps.pdb" with an RMSD of < 0.1 Å. Alternatively, you can analyze "3dna-rebuilt.pdb" to get virtually identical parameters as those from "A6me_50ps.pdb".[/li]


• These fundamental features of the base-pair parameters were introduced in the 3DNA NAR2003 paper, and more recently re-emphased in the DSSR NAR2015 paper:

  "By virtue of the definition of the standard base reference frame, Shear, Stretch, and Opening are all close to zero for Watson-Crick pairs. Moreover, every other type of pair has a set of characteristic parameters. For example, the wobble G–U pair is characterized by an average Shear of –2.2 Å, and the Hoogsteen A+U pair is distinguished by a Stretch of approximately –3.5 Å and an Opening of near 66°."
• Depending on the type and context of non-canoncial pairs, however, the associated step/helical parameters may look weird. These cryptic numbers are due to irregular base-pair reference frames of non-canonical pairs. In extreme cases, as reported in the 3DNA Forum a while ago, Rise could be zero (when the two bp reference frames are perfectly anti-parallel along the z-axis)!
• To get "reasonable" numerical values for structures with non-canonical pairs, 3DNA v2.3 and DSSR introduced a set of "simple" base-pair and step parameters. See the note in 3DNA v2.3 output:
  
  

  The simple parameters are 'intuitive' for non-Watson-Crick base pairs and
  associated base-pair steps, where the above corresponding 3DNA parameters
  may appear cryptic. Note that the following sets of simple parameters are
  for structural description only, not to be fed into the 'rebuild' program.
  Overall, they complement the rigorous characterization of base-pair
  geometry, as exemplified by the original 3DNA analyze/rebuild programs.

I know this is a confusing area, and have planed to write something on the topic. The basic idea is that "no one size fits all". I am glad that you are interested in understanding what the numbers (from DSSR) mean.

HTH,

Xiang-Jun

[lvelve0901]

Hi, xiangjun,

Thank you so much.

I have two more questions:

(1) So I just checked each frame in my MD simulation. It seems like the x-displacement and y-displacement are not varied a lot. But you do see some big error bar (standard deviation) in my full MD analysis. I am wondering could this also be due to the coordinate frame issue (maybe it's more sensitive to some distance in this frame? I don't know) or is it just my MD issue?

(2) For the simple base pair and helical parameters you mentioned above, I am wondering that where is these parameters.
So what I did is:

3dna-dssr --json --more -i=A6me_50ps.pdb -o=A6me_50ps.json

And I manipulate the json file to get all the parameters I need. But I cannot find the simple base pair and helical parameters. Could you tell me where is it? And could you give me some explanation of this simple parameters or where to find the explanation?

Thank you again.

Best,
Honglue

[xiangjun]

(1) So I just checked each frame in my MD simulation. It seems like the x-displacement and y-displacement are not varied a lot. But you do see some big error bar (standard deviation) in my full MD analysis. I am wondering could this also be due to the coordinate frame issue (maybe it's more sensitive to some distance in this frame? I don't know) or is it just my MD issue?

It is hard to say. However, the helical parameters are expected to be sensitive to local distortions during MD simulations.

(2) For the simple base pair and helical parameters you mentioned above, I am wondering that where is these parameters.
So what I did is:

3dna-dssr --json --more -i=A6me_50ps.pdb -o=A6me_50ps.json

They are exported as part of helices/stems, with only 4 parameters listed right now. Run the following command:

Code: [Select]

jq .helices A6me_50ps.json
You will see:

Code: [Select]

    "pairs": [
      {
        "index": 1,
        "nt1": "CYT1",
        "nt2": "GUA24",
        "bp": "C-G",
        "name": "WC",
        "Saenger": "19-XIX",
        "LW": "cWW",
        "DSSR": "cW-W",
        "bp1_params": [
          0.377,
          0.209,
          -1.05,
          44.947,
          -28.633,
          6.26
        ],
        "step_params": [
          -1.605,
          -0.545,
          4.072,
          -14.647,
          23.526,
          35.342
        ],
        "heli_params": [
          -3.539,
          0.357,
          3.487,
          33.19,
          20.664,
          44.637
        ],
        "bp2_params": [
          -0.413,
          0.024,
          -0.502,
          -2.768,
          -10.782,
          0.182
        ],
        "simple_Rise": 4.072,
        "simple_Twist": 37.532,
        "simple_hRise": 3.55,
        "simple_hTwist": 46.355
      },

[lvelve0901]

Hi, xiangjun,

(1) About the x and y displacement issue
Here I attach the overlay of some MD snapshots of 8-10 and 15-17 steps (9-16 bp is Hoogsteen base pair).
In the pse file I put some extreme example,

obj01: MD snapshot with x-displacement: 72 Å (9-10 step)
obj02: MD snapshot with x-displacement: -101 Å (9-10 step)
obj03: MD snapshot with y-displacement: 97 Å (9-10 step)
obj04: MD snapshot y-displacement: -192 Å (9-10 step)
obj05: fiber idealized b form DNA

These base steps looks fine to me but the output value is very huge. I think it is also due to the coordinate from of Hoogsteen base pair, right?

(2) About the simple parameters

Could you give me some introduction of these simple parameters? How do you define these parameters? The simple rise seems reasonable but I am not sure how to do manipulate the twist to make it normal?



Best,
Honglue

[xiangjun]

Hi Honglue,

Thanks for your followup.

(1) About the x and y displacement issue

I see the huge differences in numerical values of x-displacements in dinucleotide steps involving Hoogsteen pairs. Even in DNA structures with only Watson-Crick pairs, local distortions may also lead to a large variation of helical parameters. You're right in thinking that such dissimilarity is "also due to the coordinate from of Hoogsteen base pair".

To really understand why 3DNA/DSSR report the numbers you see, I'd suggest you to actually work out some examples. For example, you should be able to get the reported values for the following two cases by following $X3DNA/doc/tech-details.pdf:

obj01: MD snapshot with x-displacement: 72 Å (9-10 step)
obj02: MD snapshot with x-displacement: -101 Å (9-10 step)

You could first start with the base-pair (bp) references frames to derive the step and helical parameters. Thereafter, try to derive the bp parameters and bp reference frames from the base reference frames. Or the other way around, i.e., base reference frames ---> bp parameters + bp frame ---> step + helical parameters.

By really working out the examples, you'll have a deep understanding how 3DNA/DSSR is working. I will help you through, step by step, if you meet any difficult in the process.

(2) About the simple parameters

Could you give me some introduction of these simple parameters? How do you define these parameters? The simple rise seems reasonable but I am not sure how to do manipulate the twist to make it normal?

See my short communication "Characterization of base pair geometry" in Computational Crystallography Newsletter (2016). 7, 6–9 (also included in $X3DNA/doc/CCN201601-base-pair-geometry.pdf).

See also by blogpost "Details on the simple base-pair parameters" and other posts referred therein.

Depending on users feedback, I will add more in the future to clarify the issue.

Xiang-Jun

[lvelve0901]

Hi, xiangjun,

Thank you so much for your help. I fully understand the difficult task I have for now.

However, I think I don't need to address this issue for now. If I really need to solve this problem I will let you know and hopefully we can work with this together and conquer the task. 

Thank you again for your time。

Best,
Honglue

[xiangjun]

Thanks for your questions on this common, confusing issue. 3DNA/DSSR already has means on this very topic, for quite a long time. I will consolidate the various docs and blog posts etc in one ms to fully address the issue. However, for those who want to get to the bottom of the technical details, there is no substitute than to work through real examples. And that's what I had suggested you to do.

Xiang-Jun