We provide chemical analysis for partners at the Faculty of Agriculture and for collaborators outside the faculty. Basic and advanced techniques of analytical chemistry that we offer include targeted analyses of soluble compounds by HPLC with different optical and MS detectors and non-targeted analysis (metabolic profiling) by HPLC with ESI and full-scan MS. Analysis of volatile compound by GC-FAIMS, spectrofluorometry and different kinds of electrophoresis are also available. Flash chromatography with gradient elution, preparative HPLC and Clevenger distillation apparatus are used for purification. Rotary evaporators, vacuum concentrators and setups for preparative TLC and SPE complement the equipment of our analytical lab.
Head of the lab: Prof. Dr. Petr Karlovsky (phone 0551-3912918) In charge of HPLC-MS systems: PD Dr. Franz Hadacek (phone 0551-3914418) and Dr. Anna Rathgeb (phone 0551-3913230)
Usage policies for our HPLC-MS systems have recently been formalized in a policy document. No formal policies have been established for other chemical-analytical methods provided by our lab. Drop us a call or an email if you are interested in any of these methods.
HPLC-MS/MS in MRM mode is our standard multi-mycotoxin method. Because we work with diverse matrices such as grains, plants stems, single spikelets and bodies of insects, we calibrate our analysis with matrix-matched standards. Isotopically labelled internal standards are used to estimate in checks of method performance. Analytical standards not available commercially are prepared by purification from fungal cultures. We have been using a triple quadrupole (1200L) and an ion trap (500-MS) for mycotoxin analysis, as shown in this example from 2005. Since 2005 then we were continuously extending protocols to further mycotoxins and matrices. The acquisition of a two HPLC-MS/MS systems scheduled for July 2017 will substancially increase the sensitivity and throughput of our mycotoxin analysis.
Our oldest mass spectrometers from inside:
On the left you see our ion trap and on the right the triple quadrupole during maintenance by Katharina. LC-MS users seldom see the inside of their machine.
This is an ion trap disassembled into smallest parts. Users are not expected to try this if they intend to use the machine again but Katharina does it regularly. The skillful and highly motivated team of our workshop supports her when troubleshooting is needed.
Although HPLC-MS/MS in MRM is a golden standard for multi-mycotoxin analysis today, HPLC-FD is well suitable for certain mycotoxins; in certain situations it is preferable to HPLC-MS/MS. For instance, we use HPLC-FD for the analysis of zearalenone in our own projects on detoxification as well as in collaborations. Ochratoxin A is another mycotoxin suitable for HPLC-FD; we published a solid bar microextraction protocol with HPLC-FD detection for ochratoxin A in wheat and maize (Toxins (2016) 7:3000-3011) and used HPLC-FD in projects on zearalenone and aflatoxins.
Detection of ochratoxin A by HPLC-FD (Toxins (2016) 7:3000-3011)
HPLC-DAD proved useful in some mycotoxin projects, too, for instance in a recent work of Dr. Rosine Suchfort on enniatins. Another application of HPLC-DAD was a study of transformation products of fusaric acid, which for narrow peaks requires mobile phase additives that are not compatible with MS:
Transformation products of fusaric acid resolved by HPLC-DAD (Rasoul Abousaeedi)
Examples of publications relying on our mycotoxin analysis by HPLC-MS/MS:
1. Guo Z, Pfohl K, Karlovsky P, Dehne HW, Altincicek B (2016) Fumonisin B1 and beauvericin accumulation in wheat kernels after seed-borne infection with Fusarium proliferatum. Agric Food Sci 25: 138-145. OPEN ACCESS
3. Trümper C, Paffenholz K, Smit I, Kössler P, Karlovsky P, Braun HP, Pawelzik E (2016) Identification of regulated proteins in naked barley grains (Hordeum vulgare nudum) after Fusarium graminearum infection at different grain ripening stages. J Proteomics 133: 86-92.
3. Amato B, Pfohl K, Tonti S, Nipoti P, Dastjerdi R, Pisi A, Karlovsky P and Prodi A (2015) Fusarium proliferatum and fumonisin B1 co-occur with Fusarium species causing Fusarium Head Blight in durum wheat in Italy. Journal of Applied Botany and Food Quality 88: 228-233. OPEN ACCESS
HPLC-DAD is well suitable for most plant metabolites. With the help of pure standards every metabolite absorbing UV light can easily be quantified. HPLC-DAD is however useful for nontargeted analysis, too. UV absorption spectra combined with relative retention time are offten sufficient to point at the chemical identity of such metabolite even when standards are not available, as shown on the following figure. Dr. Franz Hadacek maintains a library of UV spectra that allows us to identify many plant metabolites without mass spectra. Using related compounds with similar chromatographic behavior and known absorption coefficients allows rough concentration estimates for metabolites without standards; when we need accurate quantification of metabolites for which standards that are not commercially available, we make them by purification.
Metabolites identified in root exudates of Arabidopsis thaliana (Dr. Pervin Akter and Dr. Franz Hadacek)
HPLC-DAD can often reveal differences among complex metabolites mixtures, such as in the following example:
HPLC-DAD chromatograms of root exudates of maize under normal conditions (left) and under stress (right) (Dr. Pervin Akter and Dr. Franz Hadacek)
We use a modified ether extraction protocol, MRM and internal isotopically labelled standards. Analysis of phytohormones is also available in Prof. Feußner's lab.
Examples of recent papers with our phytohormone analysis:
1. Ulferts S, Delventhal R, Splivallo R, Karlovsky P, Schaffrath U (2015) Abscisic acid negatively interferes with basal defence of barley against Magnaporthe oryzae.BMC Plant Biology 15/7.OPEN ACCESS
2. Häffner E, Karlovsky P, Splivallo R, Traczewska A, Diederichsen E (2014): ERECTA, salicylic acid, abscisic acid and jasmonic acid modulate quantitative disease resistance of Arabidopsis thaliana to Verticillium longisporum. BMC Plant Biology 14/85. OPEN ACCESS
3. Ratzinger A, Riediger N, von Tiedemann A., Karlovsky P (2009): Salicylic acid and salicylic acid glucoside in xylem sap of Brassica napus infected with Verticillium longisporum. Journal of Plant Research 122:571-579. Download OPEN ACCESS.
So far our lab provided pesticide analysis for two divisions of the Faculty of Agriculture and one division of the Faculty of Forest Sciences and Forest Ecology. We were however notified that the need for pesticide analysis will increase in future. Apart from routine analysis of pesticide residues, we can support studies of enzymatic transformation of pesticides in metabolism-based resistance.
The analytes are separated by HPLC on RP and detected in full-scan mode after electrospray ionization. The approach is similar to metabolomics but we use a different data processing strategy because the purpose is to identify MS signals intensified by a treatment rather than to record and identify all metabolites in a sample. Our data processing pipeline relies on a combination of commercial software and custom-made Perl scripts:
1. Noise reduction by CODA and peak detection We use CODA algorithm as implemented ACD/Labs software, raw data are imported after conversion to NetCDF.
2. Chromatogram alignment We used Perl software that adjust the retention time of each signal to compensate for shifts due to temperature variation and column aging
3. Normalization of peak intensities and comparative analysis We use a normalization algorithm that eliminates bias caused by the effect of strongly induced or suppressed metabolites on the normalization factor. Normalized intensities of sample groups (controls vs. treatment) are compared according to a rich set of customizable criteria such as variance, maximum fraction of controls with a signal, minimum fraction of samples without a signal, minimum distance between controls and treatments etc. The outcome of the analysis is the set of Rf-m/z values for candidate metabolites. These ideas were integrated into a MATLAB-based tool for MS data developed in collaboration with bioinformaticians and plant biochemists.
Metacolomics platform XCMS was used in some of our projects. XCMS is the best OA metabolomics platform we have tested but familiarization with the system takes time. For most PhD students in the past, our ACD/Labs & Perl pipeline was easier to learn. In small collaborative projects we may carry out the entire data analysis; for larger efforts, however, we recommend students to learn the system and process their data themselves. Comparing different parameter setting often helps to extract maximum information from the data.
If the metabolite has already been described, its identity can often be discerned based on m/z value of its molecular ion and fragments in MSn spectrum and UV absorption. Dr. Franz Hadacek maintains a database of UV spectra of metabolites that we have been studying, which allows us to tentatively identify these metabolite when we encounter them again based solely on their retention time and UV spectrum. Below you see an example of the analysis of maize root exudates by HPLC-DAD:
HPLC-DAD chromatograms of root exudates of maize under normal conditions (left) and under stress (right) with UV spectra recorded at major peaks (Dr. Pervin Akter and Dr. Franz Hadacek)
Information about biosynthetic capacity of the strain/species used as a source is helpful. Our Perl scripts search available databases automatically but unfortunately most metabolites relevant for the projects that we have been supporting are underrepresented in databases.
Often purification is necessary for the identification of compounds responding to a treatment (see below). Because the efficiency of ionization and thus intensity of the MS signal varies widely with structure of the analyte, wrong targets for purification were often selected in the past, resulting in purified metabolites in amounts insufficient for structure elucidation and biological assays. The evaporative light scattering detector (ELSD) will solve this problem because ELSD signal is roughly proportional to the mass of the analyte irrespective of its structure.
Examples of papers in which we used nontargeted metabolic profiling:
1. Chatterjee S, Kuang Y, Splivallo R, Chatterjee P, Karlovsky P (2016) Interactions among filamentous fungi Aspergillus niger, Fusarium verticillioides and Clonostachys rosea: fungal biomass, diversity of secreted metabolites and fumonisin production. BMC Microbiology 16/83 (13 pp).OPEN ACCESS
2. Döll K, Chatterjee S, Scheu S, Karlovsky P, Rohlfs M (2013): Fungal metabolic plasticity and sexual development mediate induced resistance to arthropod fungivory. Proceedings of the Royal Society B 280:20131219. FREE ACCESS
3. Khorassani R, Hettwer Z, Ratzinger R, Steingrobe S, Karlovsky P, Claassen N (2011): Citramalic acid and salicylic acid in sugar beet root exudates solubilize soil phosphorus. BMC Plant Biology 11/111. OPEN ACCESS.
This service task gradually developed to support projects in which mycotoxins or other chemicals are transformed by microorganisms or purified enzymes into other compounds. We know the core structure of the products and can often identify MS signals of transformation products using a technique called neutral loss scan (NLS) on a triple quadrupole or by its substitution by the analysis of MS2 data generated in automatic, data-dependent acquisition. The new HPLC-MS system that we expect to arrive in the lab this year will greatly improve our ability to estimate structures of transformation products based on their MS spectra. Analysis of UV spectra of transformation products provides useful information for some mycotoxins and transformations.
MS2 usually provides sufficient information to generate hypotheses (hydroxylation, glycosylation etc.). MS3 and higher fragmentation levels are available on the ion trap when needed. Untypical modifications such as amidation with GABA recently discovered in our laboratory require purification.
Detoxification of enniatins: Dr. Rosine Suchfort carried out the transformation and Dr. Kirstin Feußner from Prof. Ivo Feußner's lab generated HR-MS spectra
For many metabolites commercial standards for HPLC analysis are not available. We purify such standards from plant and fungal extracts using flash chromatography with gradient elution, preparative HPLC, Clevenger apparatus and preparative TLC. The progress of purification is monitored by HPLC. The protocols used vary depending on the source of material and on the purpose of the purification. A sequence of several chromatographic steps is usually used. Typically flash chromatography on silica gel with a cyclohexane - ethyl acetate - methanol elution gradient is the first step, followed by chromatography on Sephadex LH20 and/or preparative HPLC on C18. Analytical as well as preparative TLC on normal and reverse phase is often used, too. Upon the initiative of Zana Jamal Kareem we recently established a Clevenger apparatus as a simple yet powerful system for gentle distillation of essential oils from plant material. Zana currently uses Clavenger in his work with henbane (Hyoscyamus niger) and sesame (Sesamum indicum).
Apart from the preparation of standards for analytical methods, purified metabolites are often needed for biological experiments, for instance for studies of toxicity and in enzymatic transformations. Below you see purification of enniatins as an example:
Enniatins from Fusarium tricinctum as a raw extract (left) and after purification (right) (Dr. Rosine Suchfort)
Preparative separation of individual enniatins from Fusarium tricinctum (Dr. Rosine Suchfort)
Flash chromatography is the fundamental method for natural product purification. We usually start with normal phase eluted with a gradient of cyclohexane, ethyl acetate and methanol. Reverse phase and other materials are also available; we use ready-made single-use cartridges in the first step but to achieve a better separation we often fill longer columns ourselves.
Flash chromatography: in a fume hood you see pumps on the bottom (up to 5 MPa), fraction collector on the top and a column on the right. The signal of the detector is displayed on a laptop screen.
TLC is a simple yet very useful method, especially in its preparative version. Preparative TLC is often used in our lab in combination with flash chromatography and/or preparative HPLC:
TLC box in a foome hood, with Dr. Yi Kuang preparing samples for loading
Some plant and fungal metabolites are volatile. These compounds are sampled from the gaseous phase and analyzed by GC. Directly in the lab we use a GC-FAIMS system which allows us to distinguish among isomers. GC-FAIMS is a young method; the choice of software for data processing is therefore limited. We developed our own software that works with raw data from the detector, allowing us to optimize the parameters for specific tasks. On the following figures you see raw FAIMS data collected by Johannes Ott for fungal VOCs and chromatograms extracted from these data by our software:
Raw histogram of GC-FAIMS with positive ionization (Johannes Ott)
Raw histogram of GC-FAIMS with negative ionization (Johannes Ott)
Normalized GC-FAIM chromatograms for positive (red) and negative (green) ionization; notice signals in neg. ionization hardly visible in raw histogram (Johannes Ott and Petr Karlovsky)
In many projects with VOCs we however rely on state of the art GC-MS equipment in the laboratory of Prof. Ivo Feußner. Since a suspension bridge connecting our labs was constructed as you see below, moving forth and back between our labs is comfortable even on rainy days.
Bridge connecting our analytical chemistry lab (left) with the lab of Prof. Ivo Feußner (right), photo by Petr
Bridge from the other side, photo by Franz
The selection of target metabolites for purification was tricky in the past because it was based on the results of comparative metabolic profiling by HPLC-MS. Strongly ionizable metabolites with nice MS signals might be present in very low concentrations, leading to insufficient amounts after purification. With a new evaporative light scattering detector, which we acquired in August 2017, we are able to roughly estimate concentrations of metabolites of unkonwn structures as long as we can separate them by HPLC, and can thus decide whether purificaiton is meaningful.