<p>U ovoj doktorskoj disertaciji analizirano je bra&scaron;no strnih žita (p&scaron;enice, spelte, ječma,<br />raži, tritikalea, ovsa), prosolikog žita (kukuruza) i pseudocerealija (heljde i<br />amarantusa) primenom gasne hromatografije sa masenom spektrometrijom (GC-MS) u<br />kombinaciji sa multivarijantnom analizom, u cilju ispitivanja mogućnosti za<br />određivanje autentičnosti navedenih biljnih vrsta. U tu svrhu, iz bra&scaron;na navedenih<br />cerealija i pseudocerealija ekstrahovani su mali molekuli (lipidi i &scaron;ećeri). Lipidne<br />komponente bra&scaron;na ekstrahovane su heksanom. Nakon vi&scaron;estrukog obezma&scaron;ćivanja<br />istih uzoraka bra&scaron;na heksanom i su&scaron;enja, usledila je ekstrakcija &scaron;ećernih komponenata<br />96%-tnim etanolom. Dobijeni heksanski i etanolni ekstrakti derivatizovani su<br />odgovarajućim reagensima. Za derivatizacaju lipidnih komponenata kori&scaron;ćen je<br />rastvor TMSH (trimetilsulfonijum hidroksida, 0,2 M u metanolu). Time se masne<br />kiseline prevode u odgovarajuće metil-estre. Za derivatizaciju &scaron;ećernih komponenata<br />kori&scaron;ćen je etanolni rastvor natrijum-hidroksida i hidroksilamin-hidrohlorida u<br />kombinaciji sa BSTFA (bis-(trimetilsilil-trifluoroacetamidom). Na taj način su prosti<br />&scaron;ećeri prevedeni u odgovarujuće trimetilsilil-oksime. Ovako pripremljeni ekstrakti<br />uzoraka bra&scaron;na su analizirani na GC&ndash;MS uređaju.<br />Dobijeni hromatogrami međusobno pokazuju veliku sličnost, a naročito u okviru<br />uzoraka iste botaničke vrste. U obradi hromatograma i masenih spektara, kori&scaron;ćen je<br />MSD Productivity ChemStation program uz primenu Wiley 275 biblioteke masenih<br />spektara. Dobijeni podaci analizirani su na tri načina. Prvi način podrazumeva<br />kreiranje numeričkih matrica u modu ukupne jonske struje (TIC). Drugi postupak<br />podrazumeva kreiranje numeričkih matrica izolovanjem karakterističnih fragmentnih jona sa hromatograma (kod lipida 74 m/z, a kod &scaron;ećera 73 m/z u kombinaciji sa bar<br />jednim od sledećih jona: 204 m/z, 217 m/z, i 361 m/z). Kod trećeg načina obrade<br />podataka, analiza je urađena kreiranjem matrica primenom binarnog sistema (1/0), gde<br />&bdquo;1&ldquo; označava prisustvo određene komponente, a &bdquo;0&ldquo; označava njeno odsustvo u<br />posmatranom uzorku. Tako dobijeni podaci podvrgnuti su multivarijantnoj analizi<br />primenom statističkih programa &ndash; PAST i STATISTICA. U sva tri slučaja ispitivani<br />uzorci grupisani su u međusobno jasno odvojene grupe, prema odgovarajućoj biljnoj<br />vrsti. Uočeno je jasno razlikovanje pseudocerealija (heljde i amarantusa) i prosolikog<br />žita (kukuruza) od grupe strnih žita (p&scaron;enice, spelte, ječma, raži, tritikalea i ovsa).<br />Primenjenim metodama postiže se manje selektivna klasifikacija biljnih vrsta u okviru<br />strnih žita, sa izuzetkom uzoraka ovsa i spelte. Ovi rezultati pokazuju da je novim<br />navedenim postupcima moguće utvrditi autentičnost bra&scaron;na različitih botaničkih vrsta<br />cerealija i pseudocerealija.<br />Na isti način tretirani su i uzorci proizvedenog hleba (kore i sredine) u kojima je<br />p&scaron;enično bra&scaron;no supstituisano sa 0, 20, 40, 50, 60, 80 i 100% heljdinog bra&scaron;na.<br />Multivarijantnom analizom GC-MS podataka dobijaju se jasno odvojeni klasteri u<br />skladu sa porastom udela bra&scaron;na heljde u kori i sredini hleba. Najbolji rezultati<br />postignuti su koreliranjem sastava i sadržaja ugljenohidratnih komponenata u kori<br />hleba.<br />Krajnji ishod razvoja ovih novih originalnih metoda je mogućnost njihove praktične<br />primene u cilju rutinskih kontrola kvaliteta bra&scaron;na i gotovih pekarskih proizvoda u<br />laboratorijama, proizvodnim pogonima i inspekcijskim službama. Posebna pogodnost<br />primene predloženih metoda je to &scaron;to su one semi-kvalitativnog i semi-kvantitativnog<br />karaktera, pa se mogu izvoditi bez tačne identifikacije eluirajućih malih molekula</p> / <p>In this doctoral thesis various flour samples of small grains (wheat, spelt, barley, rye,<br />triticale, oats), corn and pseudocereals (buckwheat and amaranth) were analyzed using<br />gas chromatography - mass spectrometry (GC-MS) in combination with multivariate<br />analysis, in order to examine the possibilities for determining the authenticity of the<br />plant species listed. Small molecules (lipids and sugars) were extracted from flour<br />samples of the above-mentioned cereals and pseudocereals. The lipid components of<br />the flour were extracted with hexane. The extraction of sugar components from the<br />same flour samples was performed with 96% ethanol solution, after multiple defatting<br />with hexane and drying of the samples. The obtained hexane and ethanol extracts were<br />derivatized with appropriate reagents. For the derivatization of the lipid components, a<br />solution of TMSH (trimethylsulfonium hydroxide, 0.2 M in methanol) was used. Thus,<br />the fatty acids were converted into the corresponding methyl esters. For the<br />derivatization of the sugar components, an ethanol solution of sodium hydroxide and<br />hydroxylamine hydrochloride combined with BSTFA (bis- (trimethylsilyl<br />trifluoroacetamide) was used. Thus, the free sugars were converted into the<br />corresponding trimethylsilyl oximes. The derivatized extracts prepared in this manner<br />were analyzed on a GC -MS device.<br />The obtained chromatograms were very similar, and especially within the samples of<br />the same botanical species. In the processing of the chromatograms and mass spectra,<br />MSD Productivity ChemStation program was used with the Wiley 275 Mass Spectra<br />Library. The obtained data were processed in three ways. The first method involves the<br />creation of numerical matrices in the total ion current mode (TIC). The second method<br />involves the creation of numerical matrices by isolating characteristic ion fragments from the TIC chromatograms (74 m/z for the lipid components, and 73 m/z in<br />combination with at least one of the following ions: 204 m/z, 217 m/z, and 361 m/z, for<br />the sugar components). In the third way of data processing, the analysis was done by<br />creating a matrix using a binary system (1/0), where &ldquo;1&rdquo; denotes the presence of a<br />particular component, and &ldquo;0&rdquo; denotes its absence in the observed sample. The data<br />thus obtained were subjected to multivariate analysis using statistical programs - PAST<br />and STATISTICA. In all three cases, the analyzed samples were grouped in clearly<br />separated groups according to the appropriate plant species. A clear distinction was<br />observed between pseudocereals (buckwheat and amaranth), corn and the group of<br />small grains (wheat, spelt, barley, rye, triticale, oats). The applied methods achieved a<br />less selective classification of plant species within the group of small grains, with the<br />exception of oats and spelt samples. These results show that it is possible to determine<br />the authenticity of flour of various botanical species of cereals and pseudocereals,<br />applying new methods described in this doctoral thesis.<br />The samples of produced bread (crust and crumbs) in which wheat flour was<br />substituted with 0, 20, 40, 50, 60, 80 and 100% of buckwheat flour were treated in the<br />same way. Multivariate analysis of GC-MS data provided clearly separated clusters in<br />accordance with the increase in the share of buckwheat flour in bread crusts and<br />crumbs. The best results were achieved by correlating the composition and content of<br />the carbohydrate components in bread crusts.<br />The ultimate outcome of the development of these new original methods is the<br />possibility of their practical application for the purpose of routine quality control of<br />flour and bakery products in laboratories, production facilities and inspection services.<br />A particular advantage of applying the proposed methods is that they are semiqualitative<br />and semi-quantitative. Therefore, they can be performed without the exact<br />identification of eluting small molecules (lipids and free sugars), nor their<br />quantification using analytical standards.</p>

Ziele: Die Arbeit hatte die Etablierung und Validierung einer robusten, spezifischen und sensitiven GC-NCI-MS Methode zum Ziel, um Testosteron und 17α-OH-Progesteron im Serum zuverlässig quantifizieren zu können. Anschließend wurde die Methode mit dem Siemens ADVIA Centaur Immunoassay zur Testosteronanalyse verglichen. Hintergrund der Arbeit ist die mangelnde Richtigkeit und Sensitivität der Steroidhormonanalyse mit Immunoassays, inbesondere bei Hormonanylsen in geringen Konzentrationen, wie z.B. Testosteron bei Frauen und Kindern. Methode: Die Methode umfasste die Zugabe von deuterierten Internen Standard zu 1 ml Serum, gefolgt von Flüssig-Flüssig-Extraktion mit Ethylacetat und einem Clean-up mittels Festphasenextraktion. Die angereicherten und aufgereinigten Proben können mit Pentafluor¬benzyl¬hydroxylamin-hydrochlorid und MSTFA/TMCS (99:1) erfolgreich derivatisiert und im Anschluss in die GC-MS injiziert werden. Die Methode zeigt eine exzellente chromatographische Trennung. Testosteron und 17-OH-Progesteron wurden im Selected Ion Monitoring detektiert, die Quantifizierung erfolgte durch den Vergleich der Verhältnisse der Peakflächen zwischen Internen Standard und Analyten. Für den Methodenvergleich, zwischen der GC-MS Methode und dem Siemens ADVIA Centaur Immunoassay, wurde Testosteron in 10 Proben von Männern und 20 Proben von Kindern und Frauen analysiert. Ergebnisse: In dem für Männer physiologischen Konzentrationsbereich (2,62 -9,29 ng/ml) konnte eine sehr gute Übereinstimmung der Analysenergebnisse beider Methoden gezeigt werden (r= 0,97). Im Gegensatz dazu fiel im niedrigen Konzentrationsbereich aus Seren von Frauen und Kindern (0,05 - 0,51 ng/ml) eine geringe Übereinstimmung (r= 0,77) der Messergebnisse auf, wobei hier mit dem untersuchten Immunassay im Mittel um 158,9 % höhere Konzentrationen erzielt wurden. Schlussfolgerung: Der Methodenvergleich spiegelt die aktuelle Problematik in der Testosteronanalytik wider. So werden Messdifferenzen zu massenspektrometrischen Referenzmethoden auf eine mangelnde analytische Spezifität und Sensitivität, als auch ungenügende Validierung der Immunoassays zurückgeführt. Wenn niedrige Konzentrationen erwartet werden, wie bei der Testosteronanalyse in Serum von Frauen und Kindern, oder wenn zweifelhafte Analysenergebnisse vorliegen, sollte auf massenspektrometrische Referenzmethoden zurückgegriffen werden.