Introduction

The object of this tutorial is to guide you through the structure solution of cimetidine, exploring the problems of different molecular models and cis/trans isomerism (this tutorial assumes that you have completed Tutorial 1). In this tutorial, you will learn how to:

• Handle a structure solution where the molecule has different stereoisomers.

• Investigate the success rate with different sources of molecular models.

• Learn a bit more about the Pawley fitting process.

There are possible cis/trans stereoisomers, and we will refer to cis- when the cyano group is on the same side of the C=N double bond as the sulphur chain.

Data

The data set Tutorial_3.xye is a synchrotron X-ray diffraction data set collected on Station 8.3 of the Daresbury SRS. The incident wavelength was 1.5285 Å.

Stage 1: Reading the Data

• Open DASH and select the directory where the data resides.

• Select View data / determine peak positions and click Next >.

• Select the file Tutorial_3.xye using the Browse... button.

• Click Next >.

• Check that the wavelength and radiation source have been set correctly and click Next >.

Stage 2: Examining the Data

The data spans 8.01 to 56.0° 2θ. Truncate the data to 2.0 Å resolution and subtract the background using the default window value of 100. You can examine the background curve (green) in detail in the usual way by zooming in on regions of the profile. Since this data has a very uniform low-level background there are no problems, click Next >.

Stage 3. Fitting the Peaks to Determine the Exact Peak Positions

Select the first twenty peaks using the method described in Tutorial 1.

Here is a guide to the positions (° 2θ) of the first 20 peaks:

9.33	9.97	12.84	13.42	14.20
14.58	16.37	16.57	16.70	17.65
18.27	18.66	18.72	18.93	19.45
19.51	19.72	20.00	20.35	22.92

• Click Next >.

• Select Run> to run DICVOL or use another indexing program as described in Tutorial 1: Step-by-Step Structure Solution of Hydrochlorothiazide.

Stage 4. Indexing

If the selected peaks were very close to those given in the previous stage then the DICVOL program returns a monoclinic cell with a = 10.3846 Å, b = 18.7995 Å, c = 6.8201 Å, beta = 106.44°, and volume = 1277.04 ų with figures of merit M(20) = 80.4 and F(20) = 198.7. Only one other cell was suggested, also monoclinic with almost identical volume, b and c axes, and alternative a = 10.687 and beta = 111.29°.

Stage 5. Stop and Think

Does the cell make sense? In this case we estimate the molecular volume to be ~320 ų, from the fact that there are 17 non-Hydrogen atoms in the molecule, each volume approximately 15 ų and 16 H each approximately 5 ų, so (17 x 15 ų) + (16 x 5 ų) = 323 ų. Therefore, given the unit cell volume of ~1277 ų we know from this very rough approximation that the cell is most likely to accommodate 4 molecules. At this point, your knowledge of space group frequencies should suggest that P2₁/c is a strong possibility.

Stage 6. Checking the Cell and Determining the Space Group

The space group P2 will automatically have been selected. Although the most likely space group is number 14, you should now check through the systematic absences by scrolling through all the space groups, b-axis unique, which have Z = 4. You will have to decide which setting of space group 14 is correct, P2₁/a, P2₁/c, or P2₁/n. For example, look at the peaks in the region 12 to 15°, P2₁/n creates a tick mark at 13.8 where there is no peak and has no tick mark at 13.42 where there definitely is a peak. P2₁/c also has no tick mark at 13.42 but creates a tick mark at 8.1 where there is no intensity. You should examine other peaks and tick marks to confirm the choice of space group as P2₁/a.

Stage 7. Extracting Intensities

Choose 8 isolated peaks from across the pattern (e.g. 9.33, 12.84, 14.58, 17.65, 18.27, 19.72, 23.63, 26.10). Fit these peaks using the method described in Tutorial 1, then carry out the Pawley refinement. The initial 3 cycles of least squares refinement only involve the terms corresponding to the background (which actually has been removed, so notice only 2 polynomial terms are used). This should give a Pawley χ² of about 76; accept these three cycles. The next 5 cycles of least squares refinement involve the terms describing background, intensities, unit cell and zero point. This should bring Pawley the χ² down to about 35.

Up to this point the Peak Shape parameters have not been refined. To refine these you fill in the tick boxes Sigma(size), Sigma(strain), Gamma(size), Gamma(strain); it is best to try these options just one at a time. DASH does not allow you to refine both sigma parameters simultaneously, or both gamma parameters simultaneously. It’s a good idea to examine the values of the peak shape parameters (Select Peak Widths from the View menu) and refine only the parameters that have large coefficients, as these are the ones that impact upon the profile fit. In the case of cimetidine, refining just Gamma-strain and setting the number of cycles to 10 rather than the default value of 5, produced a χ² of 20.0. Your final Pawley χ² should be in the range 20 - 25.

If the χ² increased considerably after a refinement, select Reject and try refining with a different peak shape parameter.

Accept your best Pawley fit, making a note of the χ², and save it as Tutorial_3.sdi.

Stage 8. Molecule Construction

Construct a 3D molecular description of the molecule using your favourite modelling software and save it in pdb, mol or mol2 format. If you do not have a model building program to hand, there is a file supplied with the tutorial, Tutorial_3-cis.mol2. (This model for the cis-isomer was created using the Spartan program using default minimisation settings.) Select the file Tutorial_3-cis.mol2.

Stage 9. Setting up the Structure Solution Run

• Start DASH as before and select Simulated annealing structure solution from the Wizard.

• Select the Tutorial_3.sdi file.

• Click on the icon and select Tutorial_3-cis.mol2 (the file that you created in Stage 8); a Z-matrix file called Tutorial_3-cis.zmatrix will be generated automatically.

• Read in the Tutorial_3-cis.zmatrix file.

• At this point DASH will confirm that there are 14 independent parameters. These parameters are listed when you click on Next >. There are 3 parameters describing the positional coordinates, 4 (of which 3 independent) describing the molecular orientation within the unit cell and 8 variable torsion angles. Note that in this model we are keeping the cis-configuration fixed.

• Click Next > to proceed to the Simulated Annealing Protocol window. Leave the parameters set at the default values, click Next > again, then click Solve >. This will take a little longer than the earlier Tutorial examples as there are more torsion angles allowed to vary: 8 compared with 6 in Tutorial 2.

Stage 10. Monitoring Structure Solution Progress

The progress of the structure solution can be followed by monitoring the profile χ² and the difference plot.

At some point in the run you should see a dramatic fall in the χ² value from about 1000 to around 200. At this point you can investigate if a local minimisation produces an improvement - the answer will almost certainly be ‘Yes’, so accept this improved point. Have a look at the structure with the View button. You should see that the H-bonding of groups is now quite plausible; always look first for unsatisfied H-donor atoms. You will see also that the molecule has coiled around to form an intra-molecular H-bond from the NH near the end of the chain to the acceptor N on the imidazole ring.

Stage 11. Examining the Output Structure

View the structure using the View button in the Results from Simulated Annealing window. All should look reasonable, there should be no abnormal close contacts between molecules, except perhaps for some H-atoms. The H-atoms contribute such a small percentage of the total scattering power of the molecule that they have very little effect on the value of χ². The positions of methyl H atoms in particular are poorly determined, as they have been placed in calculated positions and not allowed to rotate. There is a crystal structure for this cis-isomer in the CSD (CIMEDT03); the H-bonding scheme matches this exactly. There may appear to be an extra H-bond to the cyano-N in your solution, but this will be just within the arbitrary limit set for distance scanning for H-bonds in the visualiser. (The Mercury visualiser allows you to easily examine the H-bond network just by clicking on the H-bonds.)

Stage 12. Experiments Altering the cis/trans Isomer

It is interesting to do the experiment of trying to solve the structure with the trans-isomer. You can either build a model with the trans-configuration, save it as Tutorial_3-trans.mol2 (a trans-isomer model file is provided with this Tutorial as Tutorial_3-trans.mol2), or use a more advanced feature of DASH to allow the model to rotate about the relevant C=N double bond.

You will find that if you take a trans-isomer model the SA solution process will not get very far. Using the multiple-run feature of DASH a typical set of 5 runs had final χ² of 367, 452, 647, 975, 624, with hopelessly tangled close contacts.

You can modify the previous cis-cimeditidine model to cover both cis- and trans- configurations. To do this you need to make one small change to the Z-matrix file, Tutorial_3-cis.zmatrix. Look at the labelling of the source molecular model provided with this Tutorial, Tutorial_3-cis.mol2, shown below, hydrogens omitted.

• The relevant torsion angle we now want to vary is C3-N2-C1-N11. If you examine the Z-matrix you will see the line that specifies this torsion angle, looking at the labels on the right, count up 15 from bottom. The torsion angle (column 6) is set at 1.4448765; the number 0 that follows is a control indicator to tell DASH that this is a fixed torsion angle; if you change this to 1 this indicates a variable torsion angle.

• Make this change and save your Z-matrix as a new file, e.g. cistrans.zmatrix. Now begin a structure solution run loading this new matrix; look down the list of parameters in the Parameter Bounds window. This angle is now freely variable from 0 to 360°, and the total DASH parameter count is now 15. Go to the Simulated Annealing Protocol window, where there is the option of performing multiple runs. If you do a set of 5 runs as before, with the maximum number of moves per run 3,000,000 and χ² multiplier 3.0, you will get probably about 2 or 3 correct solutions out of 5.

• A typical set of runs gave χ² = 329, 200, 76, 199, 74. The low values 74 and 76 are very satisfactory solutions; with variable torsion angle C3-N2-C1-N11 values around -6.0°, which is the cis-conformation.

(The Z-matrix format is described in Appendix F: Extinction Symbols and their Space Groups)

Stage 13. Effect of Molecular Models on Simulated Annealing

Another interesting experiment is to see how much the fine detail of the model building affects the chance of solution with DASH. Experiments were carried out with three models, the files are provided with this Tutorial as Tutorial_3-ModelA.zmatrix, Tutorial_3-ModelB.zmatrix, Tutorial_3-ModelC.zmatrix.

• Model A was prepared using the ISIS/Draw sketcher and WebLabViewer, with no energy minimisation, exported as a mol file.

• Model B was prepared using Spartan to sketch and then do a simple energy minimisation, exported as a mol2 file.

• Model C was the CIMETD03 structure taken from the CSD as a mol2 file. H-atom positions were recalculated at ideal geometry using Rpluto.

Each model was used for a set of 30 SA runs, with Max. number of moves / run 3,000,000 and Profile chi-sqd multiplier 3.0. The Pawley-fit χ² was 22.06. The solutions had profile χ² as follows:

A 18 x ~149*, 11 x ~282, 1 x 443

B 19 x ~103*, 11 x ~241

C 20 x ~83*, 10 x ~268

The solutions marked * are correct - with good H-bond patterns, and torsional geometry close to the CSD. This gives us some confidence that the solution of structure with 8 torsion angles can be carried out with a good likelihood of success. It is interesting that Model A, which had not been subjected to energy minimisation, still gave correct solutions, but with a higher χ² than the other solutions. You should try a multiple run with your own constructed cis-isomer model file.

Stage 14. Rietveld Refinement

In order to demonstrate the utility of the built-in rigid-body Rietveld refinement module, a refinement on simulated annealing solutions generated from Model B will be outlined. A similar process could be carried out using one of the interfaces to an external refinement package. Model B was generated in PCSpartan Pro and a simple energy minimisation performed and therefore can be expected to have bond angles and bond lengths that are only roughly in agreement with the crystal structure values.

• Carry out multiple simulated annealing runs with the molecular model described by Tutorial_3-ModelB.zmatrix and random seeds 314 and 159. Once the simulated annealing is complete and the Analyse Solutions dialogue box is displayed, click on the Rietveld button corresponding to the best solution.

• The refinement of the best solution from these simulated annealing runs should take the structure towards the solutions obtained for a refinement on Model C, the model generated from a single crystal structure.

• In the runs performed here, the best solution had a profile χ² of 52.88 and an intensity χ² of 41.70. After allowing all the sets of parameters to refine, individually (for example Global isotropic temperature factor, then torsion angles, then angles, then bond lengths and zmatrix) the profile χ² and intensity χ² had reduced to 34.48 and 21.14, respectively.

• The following table lists selected angles in Model B solutions before and after Rietveld refinement. The values found in the single crystal structure are also given for comparison. It can be seen that the solution for Model B is moving towards the single crystal structure during Rietveld refinement.

Angle	Before RR (o)	After RR (o)	Crystal Structure (o)
N11:C1:N5	111.53	117.54	123.89
N11:C1:N2	127.43	124.36	117.52
N2:C3:N4	176.98	170.13	170.34
N25:C24:C30	129.70	120.79	121.81
C20:S19:C16	96.47	99.07	105.50

• Not many refinement cycles are performed before changes in the χ² values become very small. Like Tutorial 1, small changes can be brought about in the value of intensity χ² by repeatedly refining the bond angles and bond lengths for example. However, given the resolution of the data, these changes do not represent an improved set of coordinates for this structure.

Stage 15. Conclusion

• One can clearly distinguish between cis- and trans- isomers in this case.

• It is possible to use DASH to allow cis/trans as a variable torsion angle, giving the correct solution.

• The accuracy of the molecular model does matter.

References

*DICVOL Program:
*D. Louer & M. Louer (1972) J. Appl. Crystallogr. 5, 271-275.
A. Boultif & D. Louer (1991) J. Appl. Crystallogr. 24, 987-993.

*Model Builders:
*PC Spartan Pro Version 1.0.5 (16/8/2000) Copyright (1996-2000) Wavefunction, Inc.

Single crystal structure cis-cimetidine (CSD reference code CIMETD03*):
*R.J. Cernik, A.K. Cheetham, C.K. Prout, D.J. Watkin, A.P. Wilkinson, B.T.M. Willis, (1991) J. Appl. Crystallogr. 24, 222-226.

Single crystal structure trans-cimetidine [CSD reference code CIMETD01*].
*L. Parkanyi, A. Kalman, B. Hegedus, K. Harsanyi, J. Kreidl (1984) Acta Crystallogr. C40, 676-679.