Solid-state NMR Course - Part One

1. Introduction

General Aims

This tutorial aims to give the reader an introduction to using CcpNmr Analysis software (version 2.2.1 and above) for the analysis of magic angle spinning (MAS) solid state NMR spectra, although much of the material is also relevant for other NMR techniques which do not observe 1H resonances, such as 13C direct detection. Some of the points covered here will also be discussed in other CCPN software courses, but this course aims to bring together a few of the most useful bits for an MAS solid state project. The emphasis is naturally on manipulating and assigning the spectra, although with significant attention paid to the isotope labelling strategies commonly employed in solid state samples.

For this part of the tutorial the CCPN project can be downloaded at CcpnCourseMas1a. and spectra can be downloaded at solidStateSpectra.tgz. Unpack these files using the command  "tar -xzf fileName".

GENERAL NOTE: this tutorial extends the one Vicky Higman created first and which is available at http://www.protein-nmr.org.uk. Supplemental figures, also from Vicky, can be downloaded in PDF form here.

 

Solid State Experiment Types

When reading in a new spectrum CcpNmr Analysis will always ask you to specify the Experiment Type. The solid-state experiment types have been overhauled for version 2.2.1 and most solid-state experiment types are now included in the CcpNmr Analysis software. Note that the experiment types do not correlate with pulse sequences, since some pulse sequences will give near identical spectra (e.g. PDSD, DARR and RFDR) and in other cases a simple change of the mixing time will substantially alter the peaks observed (e.g. DARR at different mixing times). Instead the experiment type should be chosen in accordance with the kind of spectrum observed, e.g. CC (one-bond), CC (relayed) or CC (through-space). A full table of all solid-state experiment types is available on the CCPNMR Wiki site. Additional experiments can be added by users or by the CCPN upon request. The advantage of setting experiment types when you read in your spectra is that the program can then help you to filter assignments, for instance.

 

Isotope Labelling

Several isotope labelling schemes have been incorporated into CcpNmr Analysis. Along with the standard 15N, 13C and 15N,13C,2H schemes, this tutorial will make particular reference to the 1,3-13C-glycerol and 2-13C-glycerol labelling schemes which are useful for solid-state MAS NMR spectroscopists. Knowledge of how much of a given atom site, or a combined pair of atom sites, contains spin-active nuclei is used at many points in CCPN software for the analysis of spectra. To name a few examples: predictions can be made about which assignments are possible and which are not, synthetic peak lists can be generated according to a given scheme and peak based distance restraints can be filtered according to NMR sample.

 

2. Loading and Configuring MAS Solid State Spectra

Loading a PDSD spectrum

Start CcpNmr Analysis by typing "analysis" at the command line and press <Return>:

> analysis

Or if you are using Windows simply double click the icon.

When the Analysis menu bar has appeared select M:Project:Open Project. Navigate within the file browser to find and then select the CcpnCourseMas1 project (the directory initially will be in green and then when you have selected it in deep purple). Then click [Open]. You might get a warning that various files have moved location, which is fine; just click [OK].

Dependent on how the project files and directories have been unpacked, you might also get a dialogue panel with a list of spectrum file paths, because these also have moved location. If any file location is in red, Analysis cannot find the corresponding spectrum data file, but you can tell Analysis where it is by double clicking the path cell and navigating to the correct location (although you can leave things red and accept that that particular spectrum will not have its contours displayed). Note that Analysis will be able to find the spectra for this tutorial if the masSpectra download has been extracted alongside (in the same parent directory) as the CcpnCourseMas1 CCPN project folder.

For our newly opened CCPN project you will see that a spectrum window appears with pre-loaded spectra. Next load another spectrum, a 2-dimensional 13C-13C PDSD (proton-driven spin diffusion), into this project. To do this go to M:Experiment:Open Spectra and in the resultant Open Spectra popup set the File format option at the top to UCSF (Sparky spectrum format). Then navigate to the ucsf/ directory (where the tutorial spectrum data was unpacked )and select sh3_uni_pdsd100.ucsf. Note that the the other files in this directory are either spectra that are already open, or those that will be needed later.

 

Enter a name for the selected experiment by double clicking on the experiment name cell, Expt_7, and entering the name and then pressing <Return> or clicking outside the box. Set the name of the experiment to "PDSD100_Uni" or similar; giving a visual clue to the mixing time and uniform isotope labelling.  The spectrum name can also be altered, and in this case it is probably best to save space and set the name to "SH3", or something similarly simple, given that the experiment name already contains most relevant info. Finally click [Open Spectrum].

 

Setting Solid State Experiment Types

You will next see a popup dialogue that ask for confirmation of the spectrum details. The first Verify Spectra popup will ask for verification of the spectrum referencing (how the spectrum data points relate to the ppm scale) and the spectrum file details. For the tutorial spectrum these will all be correct, so simply click on [Commit].

The next popup requires a little more attention and prompts us to specify the experiment type. In CCPN software it is important to set the experiment type (and you will be nagged at load time if you don't). The experiment type specification is used at many points in the software and makes things like assignment and restraint generation much simpler and less error prone. For example, knowledge of which experimental dimensions are connected by one-bond transfers allow many assignment possibilities to be discarded.

In the upper table of the Specify Experiment Types tab scroll down until you can see the row for the PDSD100_Uni experiment just loaded. For this row double click in the Type Synonym column and select through-space subsection CC NOESY; CC (through-space). Note that although this spectrum was specifically from a PDSD experiment we are merely specifying that the experiment has two carbon dimensions with a through-space correlation, i.e. the CCPN experiment type specification is a fairly simple one, aiming to predict what may be seen in the spectrum, and not a not a detailed quantum mechanical description of magnetisation/coherence transfer mechanisms.

The Experiment Dim-Dim Transfers panel at the lower left of the popup is worth a special note. This table allows the user to adjust which experimental dimension from the data file refers to which dimension in the reference description. For example, if Analysis loads an NCACX experiment the program does not know for sure which 13C dimension is which, i.e. which is CA and which is CX. If the initial setting is incorrect, the mapping can be altered by double clicking in the First Dim or Second Dim columns. Naturally, for the NCACX experiment the 15N dimension should have a "onebond" relationship to which ever 13C dimension represents the CA.

With the experiment type set for the new spectrum select [Close - All Done].

 

Setting Contours and Window Appearance

The newly loaded PDSD spectrum should now appear in the open spectrum window. If you cannot see the window for any reason then it can be re-opened by selecting M:Window:CC PDSD. Next, in order to clearly distinguish this spectrum from other spectra (of samples with special isptope labelling), set the contour colour of the spectrum. Thus at M:Experiment:Spectra in the Display Options tab double click in the Positive Colours cell for the new PDSD100_Uni experiment and choose black.

Also, you may wish to adjust the contour levels for the spectrum. There are a couple of ways of doing this, but the easiest is to do it directly via the spectrum window. Accordingly, at the top of the PDSD window select the [Spectra] button, and then press the coloured buttons for all of the displayed spectra except PDSD100_Uni.  This will toggle the contour display for all the other spectra to off. Then select the [Contours] button at the top and press the green up and down arrows until the contour level looks acceptable (perhaps showing a just tiny amount of noise). When the levels are set you can toggle the other spectra back on; by having them off we ensured that only the visible spectrum would have its levels adjusted.

 

3. Spectrum Windows

Navigating in Loaded Spectra

With the spectrum loaded it is worth having a little reminder of how we can navigate through the spectra (especially if this is the first CCPN tutorial you've done). Looking at the 2-dimensional PDSD window, you can move the spectra using either

You can zoom in and out in the spectrum windows by using

Move around the PDSD spectra noting that along the X axis they cover 13C chemical shifts values from the methyl region right through to the carbonyl region around 175 ppm.

If you open a 3-dimensional window by going to M:Window: CCN: NCACX Nz you will see that we get an extra scrollbar at the bottom (compared to the 2D) which allows navigation through depth planes parallel to the display, which in this case will be the 15N dimension. Also, by typing a ppm value into the bottom left corner of a 3-dimensional window you can go to the spectrum planes that position. To change the 3D plane with the mouse use the mouse wheel while pressing <control>.

Note that the spectra in this project  are colour coded according to the isotopic labelling of the samples they were recorded on:

Also note the nomenclature of the spectrum names – they indicate spectrum type, mixing time and sample labelling, e.g:

 

Marks and Crosshairs

Adjust the spectrum windows so that you can see both the PDSD and NCACX Nz spectrum windows.  Note that when moving the mouse over either of the windows crosshair lines will appear at the current cursor location and that these are at the same ppm values in both window. Wherever possible Analysis will aim to show crosshair lines at equivalent positions in all visible windows.

Because these spectrum windows show 13C-13C planes a diagonal line is drawn to indicate where ppm values are the same on both the X & Y axes. When the crosshair lines are drawn in such a window, note how that a total of four lines are drawn: there is one cross at the current cursor location and another on the opposite side of the diagonal reciprocal ppm location, e.g. if there is one cross at 40, 70 ppm another will be drawn at 70, 40. The effect of this is that you can easily find reciprocal peaks and so that you can find further peaks that derive from two correlated resonances.

Observe how, by pressing the <m> key, fixed marker lines can be drawn at the current crosshair location. In the same manner as the moving crosshair, these also show equivalent, reciprocal positions on the other side of a diagonal. To remove marker lines simply press the <n> key.

Note that to get a double crosshair mouse the window must be setup appropriately at M:Window:Windows, although we have done this in advance for the tutorial project. When you click on a window in the upper table, you will see the axes displayed in the lower table. One of the columns is headed Panel Type. By default these are all set to be different, so for a 13C-13C window, they will be called C1 and C2. If however you set the panel types to be the same, then you will obtain a double crosshair mouse.

 

Spinning Sidebands

If you zoom out in the PDSD window you will see that there are two minor diagonal lines of signal in the spectra, about 40 ppm away from the main diagonal. Make sure that all PDSD spectra are toggled on so that you can see this more clearly. The position of these is determined by the sample's spinning speed (i.e. spinning in a rotor at the magic angle). Many of the peaks along these spinning sideband diagonals are not from resonances that we want to use for assignment, hence it is a good idea if we can mark these locations. We can do this in two ways: by adding extra sideband diagonal lines and/or by marking the peaks themselves with a warning annotation.

To add sideband diagonals to the spectra we only need to specify the sample spinning speed, so that Analysis can determine the correct location. Accordingly, open M:Experiment Experiments and choose the Experimental Details tab. Here double click in the Spinning Rate column for the three PDSD experiments and set the value to 8000 Hz. Returning to the spectrum window you will now see that there are extra dashed diagonal lines repeat across the spectra to indicate the sideband positions.

To add a visible warning to a peak location the peak must first be picked (its centre found). As an example, locate some obvious sideband diagonal peaks near 0 ppm. Pick these peaks by holding down <Shift> and <Control> keys while you click and drag the mouse over peak the contours. (We deliberately require pressing keys for this so that it is difficult to accidentally pick peaks). The peaks will be picked when you can see diagonal crosses at the positions of the signal maxima. To mark a single peak with a warning, right click over it and select R:Peak:Set merit:0.0. Hopefully you will see an exclamation mark ! at the start of the peak annotation telling you that it is dodgy. To set the merit value quickly for multiple peaks you can select them with and click and drag then press <s> to get the peak selection table. In the popup Peaks:Selected Peaks table set the Merit value of the first peak to 0.0 then select all peaks (using mouse click + <Shift> on the relevant rows or <Control+A>). Ensure that the peak with zero merit is the darker, last selected row (use mouse click + <Shift>) and then press [Propagate Merit] at the bottom. Hopefully all the selected peaks now carry the warning annotation.

If you want to change the warning symbol you can do this at M:Peak:Draw Parameters in the Merit Symbols tab, e.g. entering "*BAD*" for the Poor merit entry. If you change this setting you must press [Set symbols] for the changes to take affect and press [Update Full Annotations] in the Annotation Style tab to update existing peak annotations. The effect of all this is to specify which peaks we think are useful and which are not. Any peaks with a merit value set to zero (and hence have the warning symbol) will be excluded from things like chemical shift calculations and generating structural restraints.

 

4. Isotope Labelling Schemes

We now move on to thinking about the isotope labelling in the samples that gave rise to the spectra in our CCPN project. For the black PDSD100_Uni spectrum the situation is simple: given that its sample is uniformly 13C and 15N labelled no resonances are excluded; which is the default situation in Analysis. For the other spectra however the samples were made using cultures grown with either 1,3-13C or 2-13C glycerol as the carbon source (but still have uniform 15N).

This 13C labelling means that these spectra only show a subset of the total resonances compared to the uniform case. Also, because we have two samples from mutually exclusive glycerol carbon labelling, the resulting isotopic incorporation (and thus observed resonances) will be mutually exclusive; if an atom is labelled in one glycerol-fed sample it will be unlabelled in the other. However, it should be noted that because many NMR spectra rely upon the coincidental labelling of two or more atoms at the same time, the glycerol-label spectra themselves will not add up to the uniformly labelled situation. Fortunately we do not have to calculate the levels of isotopic incorporation for particular atoms or atom pairs, because Analysis can do that automatically if we associate a specific labelling scheme with an experiment. Have a look at Vicky's supplemental figures to see illustrations of the resulting labelling patterns.

Setup of 1,3-13C and 2-13C Glycerol Samples

Analysis is already aware of the labelling patterns that result from 1,3-13C and 2-13C glycerol labelling. These patterns are stored in the form of a scheme, which describes which atoms are labelled in the various labelled isotopomer forms of each kind of amino acid residue. We can use this information in two ways in Analysis: we can either switch between the different labelling scheme information manually when we make assignments etc., or we can associate a given set of experiments with a given sample, which in turn is associated with a specific labelling pattern. If we do the latter, the labelling pattern is automatically known when we use one of the experiment's peaks.

To setup the labelling go to M:Molecule:Isotope Labelling. You will note the the resultant popup is relatively empty, but is aware of the SH3 molecule, the sequence for which is already entered into the CCPN project. Here we will associate two labelling patterns with the SH3 molecule, one for each of the two kinds of glycerol-fed sample. The setup is quite straightforward: click [New Sample] twice (the SH3 will molecule automatically be selected as it is the only one in the project) so that two rows are filled in the upper Labelled Samples table. Next we have to specify what each of these two samples contains, both in terms of the labelling pattern and which NMR data they relate to.

Click on the row of SH3 sample number 1, then in the bottom right hand corner select "13Glycerol" from the pulldown menu and press [New Pattern From Scheme:]. Hopefully you will see pattern "A" appear for this sample in the lower table. This sample will contain 100% of this pattern, but if we wanted to specify mixed sample we could add multiple patterns and specify their relative proportions. Next double click in the Experiments column in the top table for the first row and check the boxes for all the experiments with names ending in "13C". This first sample is now setup: all the required experiments are associated with the 1,3 glycerol pattern.

Now repeat the above procedure for the second sample, remembering to select the "2Glycerol" scheme at the bottom right and experiments ending in "2C". Once set, we are ready to use the labelling information in CCPN for assigments etc. Note that if the isotope labelling of your sample does not match the standard schemes that CCPN supplies. You can either tweak existing schemes; in the Labelling Patterns tab you can make adjustments and mixed residue patterns (where the labelling scheme changes in the sequence), or you can make a new reference labelling scheme at M:Molecule:Reference Isotope Schemes, which may be saved into your CCPN project.

 

5. Resonance Assignment of MAS Spectra

For this part of the course, if anything went wrong with the setup of the spectra and labelling you can load the CCPN project at CcpnCourseMas1b to get to the required point for the remainder of this section.

We now go on to do some resonance assignment, associating peaks with both atoms and residues. However, it should be noted that we are only illustrating one possible assignment strategy, naturally there are others depending on what kind of spectra have been recorded.

The SH3 domain of chicken alpha-spectrin contains three threonines. This section will show you how to identify them based on their chemical shifts, pick their peaks, generate resonances and spin systems for them, and assign their atom and amino acid types. Then you will use the 3D spectra and glycerol labelling pattern to sequence specifically assign the threonines.

For this section you may need to refer to the following Figures in the supplemental PDF file:

 

Finding Threonine Spin Systems in 13C-13C spectra

Based on their characteristic chemical shifts, try to identify the CA-CB, CA-CG2, CN-CG2, CA-CO and CB-CO cross peaks of the three threonines in the uniform (black) PDSD100_Uni spectrum and peak pick them (except possibly the CB-CY2 peaks which are not very well separated). Drawing marks through your peaks will help you connect them.

In the PDSD window, find the three CA-CB cross peaks in the black PDSD100_Uni spectrum below the diagonal (x-axis near 63 ppm; y-axis near 71 ppm). Use Figure 6a as a guide, if you like.

Begin by peak picking the left hand cross peak (<shift> + <control> + click and drag) and then press <m> with the cursor over its centre to place marker lines through it. Then navigate to the right to look at the threonine CG2 region at about 20 ppm; hopefully you will see how the marker lines for both the CA and CB resonances allow us to find pairs of matching peaks. There is only one chemical shift (22.1 ppm) at which there are peaks which go through both s lines (i.e. the CA-CG2 and CB-CG2 cross peaks).

Peak pick the CA-CG2 cross peak  at 22.1, 65.0 ppm. Unfortunately this cross peak actually has two maxima (perhaps because of the sideband) so if you get two peak crosses just delete one of them by selecting it an pressing <Del>.

Now look at the PDSD CO window, this shows the carbonyl region of the spectrum. Again there is one chemical shift (175.8 ppm) at which there are peaks which go through both marks. These are the CA-CO and CB-CO cross peaks so peak pick both of these. You have now identified four resonances for one threonine spin system. Now remove your marks with <n> and try the same procedure for the other two threonine spin systems.

If you are having trouble, then use this chemical shift table to help you find the peaks:

 

Adding Resonances and Atom Type

With a few peaks located for these threonines we will now start the formal assignment process. In CCPN this means that we will associate resonances with the dimensions of each peak. In this context a resonance means making a link to state which atom gives rise to a particular signal. Initially however we won't know exactly where in the molecule a resonance comes from, but we can certainly indicate when several peaks relate to the same resonance and also say what kind of resonance is present, in terms atom and residue type.

To do some initial assignment, go to the leftmost of the CA-CB cross peaks, place the mouse on it and either press <a> or right-click the mouse and go to M:Assign:Assign peak id. This will bring up the Assignment Panel:

Firstly, we will add a new resonance number to each peak dimension. To do this click on the [New] buttons for both the upper and lower sections (which represent the two dimensions). When you do this you will see that resonance numbers of the form [1] and [2] appear to identify the signals, both in the tables and spectra. The actual values of these numbers is not important, they are simply bookkeeping numbers for you project.

Next we will specify that these two resonances belong to the same spin system, which effectively means that they are part of the same residue in the sequence. To do this we simply add another bookkeeping number by pressing [Set Same Spin System]. Hopefully the spin system annotation {1} will appear in the peak assignment table and in the spectrum annotation.

If you know nothing else about your resonances, you could leave the assignments in this blank state until you gained further information. However, in this case we know that your resonances are in a threonine spin system and that they have CA and CB atom types. Accordingly to set the atom type, select the top {1}[1] resonance (left hand side) first and then click on [Set Type] which will bring up the Atom Browser panel. Here you will first of all have to toggle on the carbon atoms by clicking on [C] button at the top. Then select any CA button in the sequence (it doesn't matter which!). Now do the same for the lower resonance, but selecting CB as the atom type. At the end of this the two resonance annotations should be {1}CA[1] and {1}CB[2], i.e. in the same spin system and with atom types.

Next we can label the spin system (both resonances) as coming from a threonine amino acid, because of the very distinctive chemical shift combination. Naturally we cannot always do this at such an early stage. To do this click on a resonance in the left hand panels and click [Set Type] once again. This time click on any "Thr" in the Residue column (i,e, not an atom), and you should see the resonance annotation change to {1}ThrCA[1] etc.

Next we will add resonances for the CA-CG2 peak (near the sideband diagonal at 22, 65 ppm). If you bring up the Assignment Panel for this peak (<a> with the cursor over the peak) you will see that this time the tables are not completely empty. The {1}ThrCA[1] resonance we just created is listed in the lower panel because it closely matches the ppm value of this peak dimension. Click on the {1}ThrCA[1] row and the assignment will be added to the left hand panel and the spectrum. We still have to add a CG2 resonance however, so click [New] for the top panel, then [Set Same Spin System] and finally select {1}Thr[3] at the top left and [Set Type] to be CG2.

Repeat the procedure again for the carbonyl peaks peak 176 ppm  in the PDSD CO window. Remember to select an existing resonance from the right hand panels wherever possible and when you [Set Type] select "C" which is the IUPAC name for carbonyl carbon.

As an exercise repeat this procedure for the other two threonine spin systems; start from the CB-CA peak and find the four carbon resonances for each spin system, setting the atom type as you go.

Assingning Isoleucine 30

Next we will locate and assign an isoleucine spin system. Here we are lucky because isoleucines have a distinctive set of 13C chemical shifts and there is only one isoleucine in the whole sequence. The Ile CD1 chemical shift at 11.1 ppm is particularly distinctive in the PDSD spectra. Pick the five Ile peaks the line up at 11.1 ppm on the X axis, which are respectively at 11.1, 18.0, 26.6, 35.7 and 58.0 ppm on the Y axis. Looking at the supplementary figures, or in Analysis at M:Chart:Reference Chemical Shifts, selecting Residue Code: Ile and Atom Type: Heavy, you will see that the peaks almost certainly represent CD1, CG2, CG1, CB and CA resonances, going from top to bottom.

With these assignments in mind we will assign all of these peaks. As before, press <a> over any one of the new peaks and press [New] as needed for the two peak dimensions; noting that the diagonal peak with have the same resonance on both dimensions. Because the resonance in the first dimension (CD1) relates to all peaks we can quickly spread the assignment to all peaks that line up at 11.1 ppm. Hence, select all the Ile peaks with click + drag, then in the right mouse menu select R:Assign:Propagate assignments and observe how all the peaks are assigned to the same resonance along the X-axis.

Because we only have a single Ile in the sequence, rather than assigning merely the types of atom and amino acid, here we can specify exactly which atom the resonance comes from. So, press <a> over any one of the new peaks, click on the top left resonance number and press [Assign [x]], now when the Atom Browser opens, set Residue: to "Ile" and click on the CD1 cell for 30 Ile. Hopefully you will see that all the Ile peaks are now all assigned to the same, precise atom and residue. All the peaks were affected because we had connected them to the same resonance.

Now for these peaks, add Ile resonances to the second dimensions and assign directly to the correct atoms in 30Ile, using the reference chemical shifts as a guide to which atom is which. Note that if you already set the atom or spin system type for these peaks, this information is superseded when you make the complete assignment.

Peak Differences from Isotope Labelling Patterns

The 30Ile assignments allow us to illustrate the effect of isotope labelling patterns on the appearance of peaks in the other spectra, i.e. the blue PDSD50_13C and red PDSD50_2C.

Clear all existing marks with <n> and mark all four non-diagonal 30Ile cross peaks aligned at 11.1 ppm (Cg2, Cg1, Cb and Ca) using the <m> . Note how we get the expected vertical lines for all the carbon resonances. Now select the [Spectra] button at the top of the spectrum window and toggle the black uniform-label spectrum off and toggle the red and blue spectra on. See how that for the whole spin system, where the marker lines intersect, some peaks are missing in one or both spectra compared to the uniform labelling.

Pick the red peak, at 11.1, 35.7 ppm and assign it to 30IleCd1,Cb. Looking at the supplemental isotopomer diagrams can you see why this peak is visible in the 2-13C glycerol labelled sample but not in the 1,3-13C  glycerol labelled sample? With the assignment panel open (<a>) for this red peak change the Isotope Labelling pulldown to Scheme: 13Glycerol, thus overriding our automated setup. Note how the  30IleCb resonance disappears as a labelling possibility under this scheme (and also considering the adjacent Minimum Isotope Fraction setting). Resturn the Isotope Labelling pulldown to Automatic from sample and the option should re-appear.

Next add a fake peak where there are no red or blue contours at the intersection of the CG2 and CB lines.  Hold <Control> and click the mouse where the two lines cross at 18.0, 35.8 ppm. This will actually add two peaks, to both the red and blue spectra at this position, because both are visible. Now turn the red spectrum off at the top, select the fake peak and open the assignment panel with <a>. Note how for the blue spectrum the CB resonance is not present in the labelling pattern. Next switch to the red spectrum, turning the blue one off, and look at the assignment options for the other peak at this position. Here the CB resonance is present, but the CG2 is not. Thus because both labelled samples do not have both CB and CG2 atoms labelled at the same time, no peak is seen in either coloured spectrum.

The outcome of all this is that you should be aware of which atoms assignments are possible in a given isotope labelled situation. When making the initial assignments you can filter the Atom Browser by selecting a labelling scheme in the Options panel. And once you have found some resonances Analysis is fairly smart at saying which possibilities are most likely present, and which are not.

 

Finding 15N Resonances in 3D spectra

For the next section we aim to identify the threonine nitrogen chemical shifts using the 2- 13C glycerol NCACX

spectrum with 500ms mixing time; NCACX500_2C.

One way to do this is to go to open the NCACX CAz window and type the threonine CA chemical shift value into the bottom left hand box (e.g 65.0).  Alternatively, you can navigate automatically by holding the mouse cursor over the ThrCA,CB in the PDSD window and press the right button to select R:Navigate:(F2)13C - (F1)13C in NCACX CAz, being careful to choose this option which will set the CA (F1) shift as the depth position in the 3D window. Now mark the CA and CG2 crosspeaks for this residue in the PDSD window and look at the lines that appear in the 3D window.

You should be able to find a horizontal strip which contains a strong N-CA-CA 'diagonal' cross peak as well as a slightly weaker N-CA-CG2 cross peak. If you look at the glycerol labelling scheme you will see that N-CA-CO and N-CA-CB cross peaks should not be observable for the threonine in the NCACX spectrum of the 2-13C glycerol sample, as the CA and CO or CA and CB are never simultaneously labelled.

Peak pick your N-CA-CA2 cross peaks and assign the nitrogen chemical shifts using

the procedure you used above for the carbon resonances: Bring up the Assignment Panel with <a>. Select the CA and CA2 resonances suggested as assignment options (if the CG is not shown then try selecting Double Tolerances to relax the position matching a little).

Click on [New] in the nitrogen dimension and then [Set Same Spin System]. Select the nitrogen dimension resonance and click [Set Type] and then select [N] atom in the Atom Browser and click on any "N" cell.

Sequential Residue Assignments

Using the NCOCX500_13C and NCOCX200_2C spectra and the graph of standard chemical shifts (Figure 1 or M:Chart:Reference Chemical Shifts), can you work out which threonine is preceded by a valine, which by a leucine and which by a serine (see Figure 5)?

Go through the threonines one at a time. Choose one, and mark all its chemical shifts. Now use the NCOCX Nz window and set the depth plane chemical shift to that of your threonine nitrogen chemical shift; ether type into the bottom corner or use the right mouse option  R:Navigate:(F1)13C (F2)13C 15N in NCOCX Nz. Make sure that only the NCOCX500_13C and NCOCX200_2C spectra are visible. Use supplementary  Figure 5 (and Figures 1 and 4) to think about what you would expect to see in an NCOCX of the 1,3-13C glycerol sample or the 2- 13C glycerol sample if the threonine were preceded by a valine, leucine or serine.

In essence we would potentially see the carbonyl, CA and CB resonances of the current and previous residue, subject to the labelling pattern. In the nitrogen plane of the threonine which is preceded by a valine, you should only be able to see cross peaks in the NCOCX spectrum of the blue 1,3-13C glycerol sample. The CG resonances of a valine (~ 20 ppm) and the CA, CB, possibly also CG2 resonance of your threonine should be visible.

For the threonine preceded by the leucine, you should only see peaks in the 2-13C glycerol sample: the CB and CG resonances of a leucine (~42 and ~27 ppm, respectively) and the CA perhaps also CB and CG2 of the threonine.

For the threonine preceded by a serine, cross peaks will again only be visible in the 1,3-13C glycerol sample: The CB of the serine (~ 64 ppm) and the CA, CB, possibly also CG2 resonance of your threonine. Once you have found the cross peaks in the NCOCX spectra which arise from your threonines, peak pick these.

You can confirm whether the peaks which you think may arise from valine/leucine/serine residues really do, by marking them and checking whether you can see any corresponding cross peaks in the PDSD spectra (see Figure 6 for reference).

For valine you should see a CG1-CG2 cross peak (near the diagonal at ~ 20 ppm) in the 1,3-13C glycerol spectrum and CG1/2-CA (~ 20/62 ppm) and CG1/2-CB (~ 20/33 ppm) cross peaks in the uniformly labelled spectrum and a strong CA- CB cross peak (~ 62/33 ppm) in the 2-13C glycerol spectrum. It is easy to find the valine CA and CB peaks because they show strongly in the NCACX spectrum of the neighbouring threonine.

For leucine your CB and CG should form a strong cross peak in the 2-13C glycerol spectrum (~ 42/27 ppm), which you can see in the PDSD and these resonances are also clear in the NCACX spectrum of the neighbouring threonine.

For serine you would expect a CA-CB cross peak (~ 58/64 ppm) in the uniformly

labelled spectrum.

In order to assign the cross peaks in the 3D NCOCX spectra, bring up the Assignment Panel by holding the mouse over a cross peak and pressing <a>. Your threonine resonances should be provided as assignment options. But first, generate your new spin systems for the previous valine, leucine or serine.

Create a new resonances by clicking on [New]. Generate a new valine/leucine/serine spin system or add resonances to a spin system by clicking on [Set Same Spin System]. Assign the atom type by selecting the resonance and clicking on [Set Type] and then choosing the correct atom type in the

Atom Browser.

Further Assignment of the EVTMK Motif (time allowing)

Try to identify the methionine which follows this threonine. Use the NCOCX COz window and go to the plane of the threonine CO. In the 2-13C glycerol spectrum (200ms) you should be able to see links to the threonine CB and methionine CA. To confirm that the CA you have found really is compatible as a methionine CA – see if it matches up with a possible CA-CG cross peak in the 2-13C glycerol PDSD50_2C spectrum (see Figure 6 for reference). You should be able to find the nitrogen chemical shift of the methionine residue by looking for the N-CA-CG peak in the 2-13C glycerol NCACX at the CA plane (53.8 ppm). At each stage, pick your peaks and assign your resonances and spin systems.

Now see if you can find links from the methionine to the neighbouring lysine. First find the Met CO via a CO-CG peak in the PDSD100_U and an N-CA-CO peak in the NCACX200_13C (note that correlations to the Thr CO are also visible in these spectra!). Then find the Lys N and CA in the NCOCX200_2C spectrum. You can confirm whether you have identified the correct resonances by looking for a CA-CG peak in the PDSD50_2C spectrum and an N-CA-CG peak in the NCACX500_2C spectrum.

Next: Solid-state NMR Course - Part Two