NMR fingerprint matching

Ideally, structural characterization of a cyclic lipopeptides (CLiPs) includes the analysis of the fatty acid moiety, amino acid sequence and amino acid stereochemistry. While numerous techniques allow determining the amino acid sequence of a novel CLiP, stereochemical characterization is typically done by labour intensive chemical assays, such as Marfey’s analysis, GC-MS or X-ray crystallography. In the case of Pseudomonas CLiPs, genomic prediction of amino acid stereochemistry – by the analysis of dual function C/E domains – has proven to be unreliable due to the presence of epi-inactive C/E domains. Recently, Nuclear Magnetic Resonance (NMR) matching was described as a fast and accurate alternative to these methods.

NMR spectral matching is a technique that exploits the sensitivity of a so-called NMR spectral fingerprint to changes in configuration of individual amino acids to quickly elucidate the stereochemical make-up of the entire CLiP. The approach involves the comparison of the HSQC spectra (see further) of two separate compounds. If both spectra are identical, this means that the two compounds will also be identical, down to the level of the stereochemistry. In case both NMR spectra differ significantly, no conclusions can be drawn except that both compounds are structurally different. In other words, the identity of a CLiP with unknown stereochemistry may be obtained relative to that of a reference compound by matching their NMR spectra recorded under identical conditions. These reference spectra can be downloaded from the Rhizoclip website via the respective CLiP webpages, where information on how the reference CLiP was characterized is also provided.

Once the amino acid sequence of the CLiP has been determined (via NMR or other techniques), NMR fingerprint matching can be used to elucidate the stereochemical make-up of the CLiPs. This only requires the recording of a standard ¹H-¹³C HSQC spectrum of your compound of interest. Next, NMR fingerprint matching can be done either using the NMR spectra themselves, or the ¹H and ¹³C chemical shift values of the relevant nuclei. Both approaches are discussed in detail below.

By making an overlay of the ¹H – ¹³C HSQC spectra of the newly extracted compound and the reference compound, similarities and differences can quickly be identified. By specifically examining the ‘fingerprint region’ of the HSQC spectrum (3.4 – 5.6 ppm for ¹H, 45 – 75 ppm for ¹³C), conclusions can be drawn concerning the stereochemical similarity of the new compound compared to the reference. If both HSQC spectra are identical, this indicates that the structures of both compounds are identical as well, down to the level of the stereochemistry. In case differences are observed between the HSQC spectra, this implies that there are structural differences between the compounds.

On the Rhizoclip website, reference spectra are made available in the form of figures. Since these reference figures feature a transparent background, they can easily be used to make overlays in a large variety of (graphical) softwares such as Photoshop, Inkscape, Edraw, .. or even MS PowerPoint or Paint. Note that the reference figures are available for both the full NMR spectrum (0 – 5.6 ppm for ¹H, 0 – 75 ppm for ¹³C) and the ‘fingerprint region’ (3.4 – 5.6 ppm for ¹H, 45 – 75 ppm for ¹³C). It is important that the spectra that will be compared to the reference have identical chemical shift boundaries. The boundaries of your spectrum can easily be adjusted in the NMR processing software.

Summarizing, to compare a ¹H-¹³C HSQC spectrum of a newly isolated compound with that of a reference, simply make a figure of your spectrum, and overlay this with the reference figure in a (graphical) software programme. Be sure that your spectrum features the same spectral boundaries (lower and upper chemical shift values) and that the spectrum is referenced correctly.

In addition to spectral figures, the Rhizoclip website also makes reference chemical shift tables available. Similarly as for the spectral comparisons, these allow to assess spectral (and therefore structural) similarity between a newly isolated compound and the reference.

The chemical shift tables come with preconfigured x,y scatter plots that mimic the look of an NMR spectrum. A comparison with your NMR spectrum can be done by either introducing a figure of your NMR spectrum as background on the x,y scatter plot, or by introducing your own chemical shift values as a second data series.

Practical aspects of NMR matching

The NMR matching principle works in any solvent, as long as the reference CLiP and the newly isolated CLiP are recorded under identical conditions. However, to establish a standardized protocol, we propose to use DMF-d7 as general solvent. Indeed, although some CLiPs dissolve in e.g. acetonitrile-d3, others will not. DMF-d7 appears to be a good solvent for all structurally different CLiPs. Most of the reference spectra available on the Rhizoclip website are in DMF-d7, but be sure to check this in before your measurement. Additionally, it should be stressed that NMR spectroscopy is a non-destructive technique. Consequently, the solvent can simply be evaporated after the measurement, after which the CLiP can be used for further (biological) testing. However, since DMF-d7 is a solvent with a high boiling point (135°C), lyophilization will be required.

NMR measurements should be performed at the same temperature as the reference spectra. Most of these were recorded at room temperature (20°C, 298K), although there are a few that were recorded at higher temperature (40°C, 328K). Check the reference CLiP before starting your own measurement.

The NMR experiment that allows structural comparison is the so-called “heteronuclear single-quantum correlation” (HSQC) experiment, which shows signals that originate from protons (hydrogen) directly bound to carbon, a so-called ¹J_CH correlation. The combination of all backbone CH^α pairs forms the NMR fingerprint of a CLiP, which will be different for all structurally distinct CLiPs. This is because the chemical shifts (the position of the signal in the ¹H and ¹³C axes) are highly sensitive to the chemical environment of the corresponding nuclei (atoms). Consequently, if the stereochemistry of an amino acid in the CLiP varies, this will lead to changes in the NMR spectra. Finally, the ¹H-¹³C HSQC is a standard experiment that can be measured on almost any NMR magnet. Moreover, it is more sensitive that a one-dimensional ¹³C experiment, and thus takes less time to record (approx. 1-3 hours).

Note: All resonances in the NMR spectra should preferably be assigned before analysing the stereochemical make-up of the CLiP by means of NMR matching, thereby also confirming their amino acid sequence.  This can be done by recording and assigning other 2D NMR spectra such as homonuclear COSY, TOCSY and/or NOESY and heteronuclear HMBC spectra, or by using complementary techniques such as mass spectrometry.

There is an increasing tendency to identify CLiPs originating from novel sources solely through their (high resolution) mass obtained from mass spectrometry. This opens the door to erroneous attributions of identity as there are multiple CLiPs which are isobaric. Typical examples often involve CLiPs belonging to the same (l:m) group such as viscosin/WLIP/massetolide F. These are all (9:7) CLiPs sharing C₅₄H₉₅N₉O₁₆ as molecular formula. Other examples are milkisin and tensin (both from the Amphisin (11:9) group) that both correspond to C₆₇H₁₁₆N₁₂O₂₀. Importantly, isobaric CLiPs can also occur across distinct groups as strikingly demonstrated by cocoyamide (11:5) and putisolvin II (12:4). These quite distinct Pseudomonas CLiPs both share C₆₆H₁₁₅N₁₃O₁₉ as molecular formula yet differ in the total number of amino acids, their number involved in the respective macrocycle and the constitution of the acyl chain. Thus claiming identity requires more advanced characterisation using tandem mass spectrometry (MS/MS or MS²) and 2D-NMR experiments whose combined use allows to establish the peptide primary sequence, the ester cyclisation site and the nature of the fatty acid. Further complications to establish (dis)similarity between CLiPs arise however from the occurrence of D- and L-amino acids in CLiP sequences, meaning that identical primary sequences may still differ in configuration at one or more positions. Indeed, there are several CLiPs from different sources that differ in the stereochemistry of a single amino acid only – for example the pairs viscosin/WLIP (L-Leu5/D-Leu5) or viscosinamide/pseudodesmin A (L-Leu5/D-Leu5) – or where highly homologous primary sequences do not share the same pattern of D and L configurations along the sequence such as syringostatin/syringomycin and (from Bacillus), fengycin/plipastatinTherefore, there is a risk in omitting full configurational analysis by relying on conservation of the pattern of D/L configuration along a sequence based on primary sequence homology, or solely on genomic prediction This again opens the way to erroneous claims of identity (the same CLiP as) or uniqueness (a novel CLiP).

Access to NMR equipment

If you do not have access to NMR equipment at your local institution or through collaborations, the NMR & Structure Analysis unit (NMRSTR) and its NMR expertise centre (EC) at Ghent University (Belgium) provides NMR measurements as a service to the scientific community. If you need measurements or analyses to be performed, contact us via the contact page.