《天然药物化学》课程教学资源（文献资料）Structure Determination of Natural Products by Mass Spectrometry

团购合买资源类别：文库，文档格式：PDF，文档页数：21，文件大小：2.53MB

1.INTRODUCTION In the fall of 1945,I entered the University of Innsbruck(Austria),the city I was born in nineteen years earlier,to study pharmacy,a tradition in my family.By the time I finished with a master's degree in February 1948,I had realized that I did not want to spend my life in an apothecary. I had become interested in organic chemistry,so I continued in that field,which was quite easy,because for both disciplines the same laboratory courses,and lectures taught in the same classroom,were required.So I had to take only a few additional courses and exams.and then continue with ate work the only student in that category,because men my age and older either did not surviv World War II or were still in prisoner-of-war camps,a fate I had avoided at my own great risk.A that time,women rarely studied chemistry;they generally went into pharmacy.When I graduated in February 1951 with a PhD in organic chemistry,I was appointed Instructor and taught a course in the analysis of pharmaceuticals;I was also charged with running the organic chemistry laboratory for the chemis try and pharmacy students. My PhD thesis was carried out under the direction of Professor Hermann Bretschneider,an organic che ustry,first in Hungary and then in Germany.He moved rckerWord War Itohead the organicchemisty department theniversity there. He himself had studied in Vienna under Professor Ernst Spath.His research was in the design and synthesis of organic molecules that could be of therapeutic use.Consequently,my graduate work was in the same field,and,following the hierarchical principles of academia at that time, continued so after my graduation. thesis,I be restless aftera few years By chan e,I noticed one day at the dean's off Cambridge that was offering to host voung scientists and engineers over the summer of 1954. The application process was simple,just a personal letter outlining a prospective research project, publications,if any,and proof of an adequate command of the English language.I already had six papers with H.Bretschneider,and,fortunately,had five years of English in high school.Ad- ditignally,I was at that time in the process per Bretschneider's request of translating an organic chemistry tbook from English in My application s approved and at the end of May 1954 I was on my way,by boat fro Rotterdam (The Netherlands),to Boston and Cambridge,via New York.The program at MIT had been conceived and was run by a group of undergraduate students who had spent the final vears of World War l in the armed forces.They had witnessed the devastation of many cities and damage to universities,and thus wanted to help by providing facilities to carry out work that we could not do at home.I was assigned to the research group of Professor George Buichi,an organie chemist trained at the fidg ssische Technische Hochschule,Zurich(Switzerland)and wh research ntere was the structure and syn thesis o about modern instrumentation,such as ultraviolet (UV)and infrarec spectros opy,and about topics like reaction mechanisms.While still in Innsbruck,I discovered the 1954 Annual Congres of the American pharmaceutical association would be held in Boston the following august.so l took a chance and submitted an abstract about my work on the synthesis of a pyridine analog of the antibiotic chloramphenicol (1).To my surprise.it was accepted and I presented my work on August 26. When the ended rted me as a ral fell w un en whe sa exp ired.On ce home my brief stay in America had not had the blessing of Professor Bretschneider,and he practically

AC08CH01-Biemann ARI 10 June 2015 13:25 1. INTRODUCTION In the fall of 1945, I entered the University of Innsbruck (Austria), the city I was born in nineteen years earlier, to study pharmacy, a tradition in my family. By the time I finished with a master’s degree in February 1948, I had realized that I did not want to spend my life in an apothecary. I had become interested in organic chemistry, so I continued in that field, which was quite easy, because for both disciplines the same laboratory courses, and lectures taught in the same classroom, were required. So I had to take only a few additional courses and exams, and then continue with graduate work. I was the only student in that category, because men my age and older either did not survive World War II or were still in prisoner-of-war camps, a fate I had avoided at my own great risk. At that time, women rarely studied chemistry; they generally went into pharmacy. When I graduated in February 1951 with a PhD in organic chemistry, I was appointed Instructor and taught a course in the analysis of pharmaceuticals; I was also charged with running the organic chemistry laboratory for the chemistry and pharmacy students. My PhD thesis was carried out under the direction of Professor Hermann Bretschneider, an organic chemist in the pharmaceutical industry, first in Hungary and then in Germany. He moved to Innsbruck after World War II to head the organic chemistry department at the university there. He himself had studied in Vienna under Professor Ernst Spath. His research was in the design ¨ and synthesis of organic molecules that could be of therapeutic use. Consequently, my graduate work was in the same field, and, following the hierarchical principles of academia at that time, continued so after my graduation. Although this work was interesting and productive, and I learned a great deal of organic synthesis, I became restless after a few years. By chance, I noticed one day at the dean’s office an announcement for a summer program at the Massachusetts Institute of Technology (MIT) in Cambridge that was offering to host young scientists and engineers over the summer of 1954. The application process was simple, just a personal letter outlining a prospective research project, publications, if any, and proof of an adequate command of the English language. I already had six papers with H. Bretschneider, and, fortunately, had five years of English in high school. Additionally, I was at that time in the process, per Bretschneider’s request, of translating an organic chemistry textbook from English into German. My application was approved and at the end of May 1954 I was on my way, by boat from Rotterdam (The Netherlands), to Boston and Cambridge, via New York. The program at MIT had been conceived and was run by a group of undergraduate students who had spent the final years of World War II in the armed forces. They had witnessed the devastation of many cities and damage to universities, and thus wanted to help by providing facilities to carry out work that we could not do at home. I was assigned to the research group of Professor George Buchi, an organic ¨ chemist trained at the Eidgenossische Technische Hochschule, Zurich (Switzerland) and whose ¨ research interest was the structure and synthesis of natural products. During this time, I learned about modern instrumentation, such as ultraviolet (UV) and infrared spectroscopy, and about topics like reaction mechanisms. While still in Innsbruck, I discovered the 1954 Annual Congress of the American Pharmaceutical Association would be held in Boston the following August. So I took a chance and submitted an abstract about my work on the synthesis of a pyridine analog of the antibiotic chloramphenicol (1). To my surprise, it was accepted and I presented my work on August 26. When the MIT program officially ended September 15th, Professor Buchi supported me as a ¨ postdoctoral fellow until the end of November, when my visa expired. Once home, I realized that my brief stay in America had not had the blessing of Professor Bretschneider, and he practically 2 Biemann Annual Rev. Anal. Chem. 2015.8:1-19. Downloaded from www.annualreviews.org Access provided by 45.58.110.168 on 08/27/16. For personal use only

ignored me after my return.I was never asked what I did or learned;I had to continue to put my myacademic career would be what was referred to as the Habl process through which one has to demonstrate independent research.After a while,Bretschneider would permit me to publish some work under my own name,to fulfil these requirements.During my briefstay at MIT,I had learned about the American way of academic life,so different from the autocratic system prevailing at that time in Austria(and Germany).It took me only a few months to decide that I would be better off in the former than in the latter. When Geo orge Buchi heard about my desi e to continue to tudy in the Us he offered me a new postdoctoral position,which Iaccepted an started October1,1955,in the by then familia environment at MIT.The major project I worked on was the synthesis of muscopyridine for the purpose of proving a structure that George had proposed on biogenetic grounds.This compound had been isolated previously from the perfume gland of the musk deer.Such a proof of structure must begin with a compound of known structure and must involve reactions of well-established mechanisms.The svnthesis george had designed involved eleven steps.of which six had been ried out by a graduate ntil he ut of material it wa ugh to earn him a PhD.It was now my job repe. at his work on a larg cale and carry ou the ning five ns.At each step,the product of the reaction has to be isolated,purified,and fully characterized by taking melting points and infrared and UV spectra and burning a few precious milligrams for elemental analysis.The product of the final step was indeed identical to the natural product,thus proving the proposed structure(2). In the fall of 1956,while I was still working on this synthesis,the head of the Department of Chemistry at MIT,Professor Arthur C.Cop an eminent o addan nic chemist analytical divi analysis of pl fnr器t the U with my work in Buchi's lab.He offered me a position as an instructor-at the time the lowest rung of the academic ladder-starting July 1,1957.Although it meant a rather dramatic change from the synthetic chemistry I was trained in,I accepted.I now had to think of an area of research that could benefit from that training but involved analytical chemistry. 2.PEPTIDES,PROTEINS,AND MASS SPECTROMETRY It was just three years since Fred Sanger at the University of Cambridge(United Kingdom)had determined the first structure of a protein,insulin(3).This was accomplished by separating the two disulfide-linked chains,cleaving them by partial acid hydrolysis into a mixture of small peptides. and separating them by paper chromatography.Sanger had developed a method for marking the N-terminal amino group of peptides by reaction with 2,4-dinitro-fuorobenzene.Upon complete acid hydrolvsis,follo red by acids can be ified.For paper chrom A D A or A-C dipep que fora tripeptide,there are can De d for overlaps. Clearly,marking also the C-terminal amino acid would greatly facilitate protein sequencing and simultaneously combine svnthesis with analvsis.It so happened that my final work in Innsbruck involved the synthesis of 3-amino-1,2,4-triazoles from carboxylic acid hydrazides(4).Carrying out this reaction on a peptide hydrazide,produced by partial hydrazinolysis of a protein,would label the C-terminal amino acid with a ver stable marker.Combined with Sange ent,a tripe tide,'A-B-C reas for a te the proposal ww.annualrecieesorg.Strucure Determination of Natural Prodncts 3

AC08CH01-Biemann ARI 10 June 2015 13:25 ignored me after my return. I was never asked what I did or learned; I had to continue to put my nose to the grindstone and just keep quiet. The next step in my academic career would be what was referred to as the Habilitation, a process through which one has to demonstrate independent research. After a while, Bretschneider would permit me to publish some work under my own name, to fulfil these requirements. During my brief stay at MIT, I had learned about the American way of academic life, so different from the autocratic system prevailing at that time in Austria (and Germany). It took me only a few months to decide that I would be better off in the former than in the latter. When George Buchi heard about my desire to continue to study in the US, he offered me a ¨ new postdoctoral position, which I accepted and started October 1, 1955, in the by then familiar environment at MIT. The major project I worked on was the synthesis of muscopyridine for the purpose of proving a structure that George had proposed on biogenetic grounds. This compound had been isolated previously from the perfume gland of the musk deer. Such a proof of structure must begin with a compound of known structure and must involve reactions of well-established mechanisms. The synthesis George had designed involved eleven steps, of which six had been carried out by a graduate student, until he ran out of material. It was enough to earn him a PhD. It was now my job to repeat his work on a larger scale and carry out the remaining five reactions. At each step, the product of the reaction has to be isolated, purified, and fully characterized by taking melting points and infrared and UV spectra and burning a few precious milligrams for elemental analysis. The product of the final step was indeed identical to the natural product, thus proving the proposed structure (2). In the fall of 1956, while I was still working on this synthesis, the head of the Department of Chemistry at MIT, Professor Arthur C. Cope, an eminent organic chemist of his day, decided to add an organic chemist to the analytical division. He knew that I had taught a course in qualitative analysis of pharmaceuticals at the University of Innsbruck, and apparently was quite impressed with my work in Buchi’s lab. He offered me a position as an instructor—at the time the lowest ¨ rung of the academic ladder—starting July 1, 1957. Although it meant a rather dramatic change from the synthetic chemistry I was trained in, I accepted. I now had to think of an area of research that could benefit from that training but involved analytical chemistry. 2. PEPTIDES, PROTEINS, AND MASS SPECTROMETRY It was just three years since Fred Sanger at the University of Cambridge (United Kingdom) had determined the first structure of a protein, insulin (3). This was accomplished by separating the two disulfide-linked chains, cleaving them by partial acid hydrolysis into a mixture of small peptides, and separating them by paper chromatography. Sanger had developed a method for marking the N-terminal amino group of peptides by reaction with 2,4-dinitro-fluorobenzene. Upon complete acid hydrolysis, followed by paper chromatography, the marked and unmarked amino acids can be identified. For a dipeptide, the sequence ∗A-B is unique; for a tripeptide, there are two possibilities, ∗A-B-C or ∗A-C-B, which still can be used for overlaps. Clearly, marking also the C-terminal amino acid would greatly facilitate protein sequencing and simultaneously combine synthesis with analysis. It so happened that my final work in Innsbruck involved the synthesis of 3-amino-1,2,4-triazoles from carboxylic acid hydrazides (4). Carrying out this reaction on a peptide hydrazide, produced by partial hydrazinolysis of a protein, would label the C-terminal amino acid with a very stable marker. Combined with Sanger’s reagent, a tripeptide, ∗A-B-C∗∗, would be uniquely identified, whereas for a tetrapeptide two possible sequences would remain. I prepared an application for a research grant from the NIH on the basis of this proposal. www.annualreviews.org • Structure Determination of Natural Products 3 Annual Rev. Anal. Chem. 2015.8:1-19. Downloaded from www.annualreviews.org Access provided by 45.58.110.168 on 08/27/16. For personal use only

Then something happened that would dramatically change my future career.Firmenich&Cie a preeminent Swiss firm active in flavor and fragrances that also funded my position in Biichi' group,asked me to attend a conference on food flavors to be held in Chicago in late spring 1957 and to provide a report.Although I was not interested in that topic,attending provided me with the opportunity to fly to Chicago.Of the papers presented,one in particular caught my attention:that by William H.Stahl of the research labo of the Us Ar y ouarterma ter corns in natick MA,not far fro as acetone,ethyl butyrate,butylacetate,fruit mass spectrometry (MS).It was done by comparing the mass spectra of the compounds isolated with the spectra of authentic compounds. Once I had sent off my report to Firmenich,I looked up what was known about the mass spectra of organic molecules.Most of it was about hydrocarbon analysis,because the method was widely used in the s on aliphatic alcohols (5). aldehydes(⑥，ker s (7),me ster(⑧，and amines(9,to more imp They all dealt with orrelation be twee nown mass spectra to establish fragmentation rules.It soon became clear to me that mass spectrometry is particularly informative about the structure of linear molecules containing heteroatoms-and peptides are such molecules!But I also had learned that to obtain a mass spectrum,the compound has to be vaporized into the ion source,usually held at 250C at low pressure.However,peptides decompose rather than vaporize upon heating due to their zwitterionic character,which is caused by the pre ence of an acidic and a basic p in the well as nds between the and my training me in handy.The carboxyl group can b e converte the amino group can be acylated,and the carbonyl groups can be reduced to CH2 by lithium aluminum hydride,producing a polyamino alcohol that retains the backbone and position of the side chains of the original peptide(Figure 1). 1.MeOH/HCI 2.Ac20/py R. R LiAID./glyme CHCDNH-CH-CONH-CH-CO3-NH-CH-CD-OH Figure 1 sidechains ofa from Reference 46

AC08CH01-Biemann ARI 10 June 2015 13:25 Then something happened that would dramatically change my future career. Firmenich & Cie., a preeminent Swiss firm active in flavor and fragrances that also funded my position in Buchi’s ¨ group, asked me to attend a conference on food flavors to be held in Chicago in late spring 1957 and to provide a report. Although I was not interested in that topic, attending provided me with the opportunity to fly to Chicago. Of the papers presented, one in particular caught my attention: that by William H. Stahl of the research laboratory of the US Army Quartermaster Corps in Natick, MA, not far from Cambridge. He reported on the identification of simple, small molecules, such as acetone, ethyl butyrate, butyl acetate, etc., in fruit extracts using a method I had never heard of: mass spectrometry (MS). It was done by comparing the mass spectra of the compounds isolated with the spectra of authentic compounds. Once I had sent off my report to Firmenich, I looked up what was known about the mass spectra of organic molecules. Most of it was about hydrocarbon analysis, because the method was widely used in the petroleum industry. However, there were also papers on aliphatic alcohols (5), aldehydes (6), ketones (7), methyl esters (8), and amines (9), to name the more important ones. They all dealt with the correlation between the known structure of reference compounds and their mass spectra to establish fragmentation rules. It soon became clear to me that mass spectrometry is particularly informative about the structure of linear molecules containing heteroatoms—and peptides are such molecules! But I also had learned that to obtain a mass spectrum, the compound has to be vaporized into the ion source, usually held at 250◦C at low pressure. However, peptides decompose rather than vaporize upon heating due to their zwitterionic character, which is caused by the presence of an acidic carboxyl group and a basic amino group in the same molecule, as well as hydrogen bonds between the carbonyl groups and amide hydrogens. Here my training in synthetic organic chemistry came in handy. The carboxyl group can be converted to an ester, the amino group can be acylated, and the carbonyl groups can be reduced to CH2 by lithium aluminum hydride, producing a polyamino alcohol that retains the backbone and position of the side chains of the original peptide (Figure 1). Figure 1 Reaction scheme for the reduction of peptides to polyamino alcohols. LiAlD4 was used to prevent the sidechains of aspartic acid and serine, and of glutamic acid and threonine, from becoming isobaric. Reprinted from Reference 46. 4 Biemann Annual Rev. Anal. Chem. 2015.8:1-19. Downloaded from www.annualreviews.org Access provided by 45.58.110.168 on 08/27/16. For personal use only

A more formidable obstacle was the fact that MIT did not have a mass spectrometer(MS),nor ates at that time.When ,he replied,"It takesa full-time electric I enginee to keep it running."I remarked that I could do this myself and explained my plans.He listene carefully and responded,"Ifyou promise the instrument will not collect dust,I promise to provide the money."We both kept our word.At that time,federal agencies did not easily fund expensive instruments like mass spectrometers,which cost more than S50.000.What I did not know was that Cope had recently negotiated with MI'Ts president James R.Killiana fund of $250,000 to be r the next t ade the chemistry depar tment.In a letter dated July 22,1964. to the then president of MITJulius A.Stratton,professor Cope stated that the purchase of my mass spectrometer was one of the best uses of the funds:"Subsequently he has become recognized as the foremost person in the world in the application of mass spectrometry to the determination ofstructures of organic compounds..."(10).In addition to the $50,000 for the mass spectrometer Firmenich Cie.provided $10,000 with the understanding that I would provide mass spectra of some of its research products,including their interpretation. IorderedaCEC21-103Cma er.the utinely usedin the from Co orporation CE C)in Pasa a,CA,which deliv early May 1958.In the meantime,the NIH grant I had applied for was approved.It allowed changing the approach to a more promising one,so I was permitted to use the funds for the mass spectrometric sequencing of peptides.It also provided funds for a postdoctoral associate,a position I offered to Josef("Sepp")Seibl,who had recently obtained his PhD in Bretschneider's group and whom Iknew well.He accepted and arrived one day before CEC'sengineer Hank DeQuasie began installing the instrum ent.It took six eks,includine g two weeks of instruction for operation nd intena e test.Thi ole irbons. As I was not interested in hydrocarbons,or ir quantitative analyses,I asked that it be a mixture of the corresponding alcohols.CECmanagement at first refused my request,but Hank persuaded them.The experiment worked satisfactorily and we took official possession of the mass spectrometer. A National Science Foundation (NSF)grant I had been approved for also provided a post- up,Fritz Ga who ed a few nths later at which the three of us started our rk o oject.As Ihac the pept d pre cted,the po yaminoalcohol ery go ss spectra and odue to the preferential clevage at the NHCH(R-CH,N This feature was very important,because we could not use the common method of identifying the unknown by comparison with the mass spectrum of a standard reference compound.There are 20 different amino acids in mammalian proteins;as such,there are 400 different dipeptides,8,000 peptides,et Therefore,it was practically impossible to compilea brary of authent di-,tri-,ete nino alcohols for c and we had to in terpret the cratch.A rt c was the first pa er reporting the in peptide (and prot n)ch try.These were suf ently vo to gas chromatography(GC)(12),thus allowing the separation of the complex mixtures expecte from partial hydrolyses of proteins.Over time.the chemical procedure was improved with the use of trifluoroacetylation(13)and silylation(14),which allowed its extension to penta-or even hexapeptides.With the interfacing of the GC to the MS(15)and to an IBM 1800 computer (16), eof心en ss spect also sed for osteocalcin fron chicken b one,a fifty ar o acid】 ong protein. cont s three y-car oxy-glutami which were converted to y-dideutero-glutamic acid before acid or enzymatic hydrolysis(18). ww.Structure Determination of Natural Prodnc

AC08CH01-Biemann ARI 10 June 2015 13:25 A more formidable obstacle was the fact that MIT did not have a mass spectrometer (MS), nor did any academic chemistry department of a university in the United States at that time. When I asked Professor Cope why we did not have one, he replied, “It takes a full-time electrical engineer to keep it running.” I remarked that I could do this myself and explained my plans. He listened carefully and responded, “If you promise the instrument will not collect dust, I promise to provide the money.” We both kept our word. At that time, federal agencies did not easily fund expensive instruments like mass spectrometers, which cost more than $50,000. What I did not know was that Cope had recently negotiated with MIT’s president James R. Killian a fund of $250,000 to be spent over the next ten years to upgrade the chemistry department. In a letter dated July 22, 1964, to the then president of MIT Julius A. Stratton, professor Cope stated that the purchase of my mass spectrometer was one of the best uses of the funds: “Subsequently he has become recognized as the foremost person in the world in the application of mass spectrometry to the determination of structures of organic compounds...” (10). In addition to the $50,000 for the mass spectrometer, Firmenich & Cie. provided $10,000 with the understanding that I would provide mass spectra of some of its research products, including their interpretation. I ordered a CEC 21-103C mass spectrometer, the type routinely used in the petroleum industry, from Consolidated Electrodynamics Corporation (CEC) in Pasadena, CA, which was delivered early May 1958. In the meantime, the NIH grant I had applied for was approved. It allowed changing the approach to a more promising one, so I was permitted to use the funds for the mass spectrometric sequencing of peptides. It also provided funds for a postdoctoral associate, a position I offered to Josef (“Sepp”) Seibl, who had recently obtained his PhD in Bretschneider’s group and whom I knew well. He accepted and arrived one day before CEC’s engineer Hank DeQuasie began installing the instrument. It took six weeks, including two weeks of instruction for operation and maintenance and for running an acceptance test. This was to be the quantitative analysis of a mixture of low–molecular weight hydrocarbons. As I was not interested in hydrocarbons, or in quantitative analyses, I asked that it be a mixture of the corresponding alcohols. CEC management at first refused my request, but Hank persuaded them. The experiment worked satisfactorily and we took official possession of the mass spectrometer. A National Science Foundation (NSF) grant I had been approved for also provided a postdoctoral position. I offered it to another recent graduate from Bretschneider’s group, Fritz Gapp, who arrived a few months later, at which point the three of us started our work on the peptide project. As I had predicted, the polyamino alcohols produced very good mass spectra and sequence information due to the preferential cleavage at the NHCH(R)-CH2NH bonds (Figure 2). This feature was very important, because we could not use the common method of identifying the unknown by comparison with the mass spectrum of a standard reference compound. There are 20 different amino acids in mammalian proteins; as such, there are 400 different dipeptides, 8,000 different tripeptides, etc. Therefore, it was practically impossible to compile a library of authentic di-, tri-, etc., amino alcohols for comparison, and we had to interpret these spectra from scratch. A subsequent short communication (11) was the first paper reporting the use of mass spectrometry in peptide (and protein) chemistry. These derivatives also were sufficiently volatile to be amenable to gas chromatography (GC) (12), thus allowing the separation of the complex mixtures expected from partial hydrolyses of proteins. Over time, the chemical procedure was improved with the use of trifluoroacetylation (13) and silylation (14), which allowed its extension to penta- or even hexapeptides. With the interfacing of the GC to the MS (15) and to an IBM 1800 computer (16), a powerful system was created for determining, finally, the first primary structure of a protein, subunit I of monellin, entirely by mass spectrometry (17). It was then also used for osteocalcin from chicken bone, a fifty amino acid long protein, which contains three γ-carboxy-glutamic acids which were converted to γ-dideutero-glutamic acid before acid or enzymatic hydrolysis (18). www.annualreviews.org • Structure Determination of Natural Products 5 Annual Rev. Anal. Chem. 2015.8:1-19. Downloaded from www.annualreviews.org Access provided by 45.58.110.168 on 08/27/16. For personal use only

236 CH HCH3 278 CH2 16 CH3 2 CH3CD2-NH-CH CD2 1NH -CH1CD2-N CD20H 677 17518 26031 MW293 20 236 260 .294 m/e 100 140 180 220 260 3U0 Figure 2 s spectrum of the derivative of tripeptide (Leu-Ala-Pro).Figure reprinted from Reference 12 3.ALKALOIDS Naturally occurring substances,mainly from tropical plants,had been used for centuries in folk medicine and later in academic studies.Atropine,morphine,and yohimbine are good examples In the first half of the twentieth century,the pharmaceutical industry began the systematic search for active substances.An early success was an alkaloid isolated in 1932 from the roots of Rauolfia serpentina and named reserpine.Its structure was not determined until 12 years later.Reserpine was one of the first antihypertensive drugs and became a huge financial success for CIBA(Basel, Switzerland)in the early 1950s.Research laboratories in both the pharmaceutical industry and academia raced to dupl e this achi names and melting points,but determining their structures was a laborious,time-consuming process. In the course ofmy work on muscopyridine Ihad learned much about alkaloidsand the problems involved in the determination of their molecular structures.I remembered one case,where three independent laboratorieshad proposed Structure Ifor the alkaloid sarpagine,but only one ofthem suggested proving it by correlation with a degradation product(Structure ID ofanother alkaloid of ajmaline.However,three years hadg one by without success.This was probably elation,the to the one it should be comparedh buthssevera eperimentallc steps.It now occurred to me that by using mass spectra for the comparison,one needed to prepare only a similar comparison product.The reasoning was that because these indole alkaloids consist of an aromatic system and a polycyclic unit,only the latter will fragment;the former will remain relatively intact.Therefore,the mass spectra of molecules with identical polycyclic structures,but differently substituted aromatic portions,will display the same patterns will be shifted by the mass differen of the s substituentso the aromatic por ry to convert sarpagin (Structure mpond of sterure .hicoed on thre very simple chemical raction Checking the product of each step by mass spectrometry eliminated the need for conventional 6

AC08CH01-Biemann ARI 10 June 2015 13:25 Figure 2 Mass spectrum of the derivative of a tripeptide (Leu-Ala-Pro). Figure reprinted from Reference 12. 3. ALKALOIDS Naturally occurring substances, mainly from tropical plants, had been used for centuries in folk medicine and later in academic studies. Atropine, morphine, and yohimbine are good examples. In the first half of the twentieth century, the pharmaceutical industry began the systematic search for active substances. An early success was an alkaloid isolated in 1932 from the roots of Rauwolfia serpentina and named reserpine. Its structure was not determined until 12 years later. Reserpine was one of the first antihypertensive drugs and became a huge financial success for CIBA (Basel, Switzerland) in the early 1950s. Research laboratories in both the pharmaceutical industry and academia raced to duplicate this achievement, and shelves began to fill with alkaloids, which had names and melting points, but determining their structures was a laborious, time-consuming process. In the course of my work on muscopyridine I had learned much about alkaloids and the problems involved in the determination of their molecular structures. I remembered one case, where three independent laboratories had proposed Structure I for the alkaloid sarpagine, but only one of them suggested proving it by correlation with a degradation product (Structure II) of another alkaloid of known structure, ajmaline. However, three years had gone by without success. This was probably the case because for such a correlation, the sarpagine molecule has to be converted to a compound identical to the one it should be compared with, but this involves several experimentally difficult steps. It now occurred to me that by using mass spectra for the comparison, one needed to prepare only a similar comparison product. The reasoning was that because these indole alkaloids consist of an aromatic system and a polycyclic unit, only the latter will fragment; the former will remain relatively intact. Therefore, the mass spectra of molecules with identical polycyclic structures, but differently substituted aromatic portions, will display the same patterns, except that some peaks will be shifted by the mass difference of the substituents on the aromatic portions of the molecules. According to this reasoning, it was only necessary to convert sarpagine (Structure I) to a compound of Structure III, which involved only three very simple chemical reactions. Checking the product of each step by mass spectrometry eliminated the need for conventional 6 Biemann Annual Rev. Anal. Chem. 2015.8:1-19. Downloaded from www.annualreviews.org Access provided by 45.58.110.168 on 08/27/16. For personal use only

AC08CH01-Biemann ARI 10 June 2015 13:25 Structures I–III characterization, as described for the muscopyridine work. The structure of ajmaline had been determined by Professor Robert B. Woodward at Harvard University (Cambridge, MA), just up the river from MIT. In the course of that work he had prepared compound II. I knew him very well and he gladly provided me with a sample. Woodward, the preeminent natural products chemist of the time (and a 1965 Nobel laureate) was curious about how my novel approach to structure proof would work out. The mass spectra of compounds II and III indeed showed almost identical patterns with the predicted shift in mass (Figure 3) (19). This method then became known as the mass spectrometric shift technique. I presented this work at the International Symposium on the Chemistry of Natural Products in August 1960 in Melbourne, Australia. Apparently, word of my work had gotten around and the room was so crowded that the eminent Bob Woodward had to sit on the steps of the lecture room. After my talk, Professor Carl Djerassi of Stanford University (Stanford, CA) came over and invited me to come to his laboratory to help set up his mass spectrometer, which was already on order, and teach his students and postdocs how to interpret the spectra. I agreed and spent January and February of 1961 at Stanford. Later, some of my colleagues commented that it was a great mistake to help a powerful, well-funded man like Djerassi become my competitor. However, I had thought that if one develops a useful new methodology, one should also help to spread it around. Carl indeed competed with me in the alkaloid field, but we became good friends in later years. In a retrospective article (20) he paid tribute to my contributions, referring to a seminar talk I gave while at Stanford in 1961: “It was the elegant rationalization by Biemann et al. of the mass spectral fragmentation behavior of alkaloids of the aspidospermine class that stimulated a serious effort at Stanford on organic chemical applications of mass spectrometry” (p. 1341). A www.annualreviews.org • Structure Determination of Natural Products 7 Annual Rev. Anal. Chem. 2015.8:1-19. Downloaded from www.annualreviews.org Access provided by 45.58.110.168 on 08/27/16. For personal use only

198 295 160 180 200 220 240 260 280 300 182 237 279 m/e120 140 160 180 200 220 240 260 280 Figure3 ssspectra of the conversion product III()of sarpagine and that (Structure I of ajmaline ()Reprinted from Reference19. book I wrote soon thereafter (1)helped spread this info mation much more effectively.It was later considered so seminally significant that it was republished in 1998 by the American Society for Mass Spectrometry. 4.HIGH-RESOLUTION MASS SPECTROMETRY By the mid-1950s,John Beynon at Imperial Chemical Industries(Manchester,UK)had built a double-focusi ng mas of theN which he demonstrate the value of high-resolution mas tion of molecules and their fragments.CEC had built a double-focusing mass spectrometer of the Mattauch-Herzog geometry with a spark ion source for inorganic analyses.It used a photographic 贴ate to reordthe n amwhic独8 focal plane,not just one focal point.Insp orga rt from the NIH and NSF.I purchased the CEC 21-110B high-resolution mass spectrometer in 1962.The pho ophtepealediomc, because one could record a complete mass spectrum in a minute or less at high resolution and mea sure it later at millimass accuracy,independent of the spectrometer.For this purpose,we adapted a commercially available microdensitometer to semiautomatically scan the photoplate and punch line position and density onto IBM cards,which then were processed to yield elemental composi- tion and abundance at MII's computer center.This wasa great advantage over Beynon's method of carefully me ring the rati sively focus a single ion and th a reference ion onto thec tor slit(peak mat ching),or segment of the spectrum and recording it with an oscillograph recorder. ainga very shor When we reported ou method and associated computer algorithms for the interpretation of the data at the E-14 meeting in Montreal,Canada.in 1964 (23).it caused a scramble for any method that could accomplish the same with the Nier-Johnson geometry of Beynon's instrument,which Associated Electronics

AC08CH01-Biemann ARI 10 June 2015 13:25 Figure 3 Mass spectra of the conversion product III (top) of sarpagine and that (Structure II) of ajmaline (bottom). Reprinted from Reference 19. book I wrote soon thereafter (21) helped spread this information much more effectively. It was later considered so seminally significant that it was republished in 1998 by the American Society for Mass Spectrometry. 4. HIGH-RESOLUTION MASS SPECTROMETRY By the mid-1950s, John Beynon at Imperial Chemical Industries (Manchester, UK) had built a double-focusing mass spectrometer of the Nier-Johnson geometry, with which he demonstrated the value of high-resolution mass spectrometry for organic compounds (22). This approach appealed to me for my work on alkaloids, because of the ability to determine the elemental composition of molecules and their fragments. CEC had built a double-focusing mass spectrometer of the Mattauch-Herzog geometry with a spark ion source for inorganic analyses. It used a photographic plate to record the ion beams, which is possible with the Mattauch-Herzog geometry given its focal plane, not just one focal point. Inspired by Beynon’s work, they also fitted it with an electron ionization source for organic materials. With support from the NIH and NSF, I purchased the CEC 21-110B high-resolution mass spectrometer in 1962. The photoplate appealed to me, because one could record a complete mass spectrum in a minute or less at high resolution and measure it later at millimass accuracy, independent of the spectrometer. For this purpose, we adapted a commercially available microdensitometer to semiautomatically scan the photoplate and punch line position and density onto IBM cards, which then were processed to yield elemental composition and abundance at MIT’s computer center. This was a great advantage over Beynon’s method of carefully measuring the ratio of accelerating voltages necessary to successively focus a single ion and then a reference ion onto the collector slit (peak matching), or of scanning a very short segment of the spectrum and recording it with an oscillograph recorder. When we reported our method and associated computer algorithms for the interpretation of the data at the E-14 meeting in Montreal, Canada, in 1964 (23), it caused a scramble for any method that could accomplish the same with the Nier-Johnson geometry of Beynon’s instrument, which Associated Electronics 8 Biemann Annual Rev. Anal. Chem. 2015.8:1-19. Downloaded from www.annualreviews.org Access provided by 45.58.110.168 on 08/27/16. For personal use only

Industries of Manchester,UK,had commercialized as the MS-9.Recording onto magnetic tape interface the gas chromatograph to the mass spectrometer in an efficient and practical way (15) Previous efforts were hampered by the lack of a suitable way of recording the spectra.Gohlke (24)had used a time-of-fight mass spectrometer (TOF MS)and an oscilloscope,the screen of which was photographed with a Polaroid camera.This limited it to quite small molecules eluting at relatively large time intervals.Ryhage(25)interfaced the GC with a conventional mass ometer,the slow eed of which limited the system to the rec of data in the single-ion monitoring mod de.Using the phooplate.high-res mass spectra of 16 alkaloids eluting from a single injection of the crude extract from the bark o Aspidasperma quebracbo blanco.Three of these were known alkaloids;the molecular structure of the remaining 13 we determined from their high-resolution mass spectra (26).Within two decades we elucidated the structures of more than 40 indole alkaloids,culminating-in collaboration with chemists from the lilly research laboratories ofIndianapolis.in-in the structures ofthe dimeric indole alkaloids vinblast tine and vincristine (27),the first anticancer drugs,which are still being ed today in chemotherapy. By the mid-1980s the search fora natural product for medical use subsided.Thanks,at least in part,to mass spectrometry,most major plant alkaloids were known and had been pharmacologically tested.Rather than laboriously isolating biologically active compounds from natural sources it became general practice to generate complex mixtures-libraries-of molecules by chemical reactions of a mixture of related molecules with one or a few reactants,followed by the isolation of biologically active products,if any.Mass spectrometry also plays an important role in this area of biopharmaceutical research 5.LARGE PROTEINS In the meantime,our work on the primary structure of proteins had continued,but so had other methodologies.The by far most useful technique was the stepwise removal and identification of the N-terminal amino acid developed by Pehr Edman(28),soon automated (29)and commercialized. It had s me drawbacks:It did t acylated at theNt nus,which nany mamm pr ns are,an d it had difficulties if the C-t erm nal end of nonpolar,hydrophobic amino acids.Because these two situations do not interfere -in fact were even an advantage with our mass spectrometric approach-it was logical to combine the two. This was done in collaboration with Gobind H.Khorana,professor of biology at MIT,who was working on the structure and activity of bacteriorhodopsin from Halobacterinm balobium,a protein that loops seven times through the cell wall and thus has long stretches of hydrophobic amino t to be 248 amino acids long (30). By the lat :1970s,DNAs ed tothe point that it became feasible to determi the nucleotide sequence of the DNA coding for a protein and translate it using th genetic code into the corresponding amino acid sequence of the protein.Paul Schimmel,professor of biology at MIT at the time,had isolated from Escbericbia coli alanyl-tRNA synthetase,a very large protein,and was interested in its structure and active site.He was planning to use the method of Maxam and Gilbert(31)to determine the DNA sequence of the strand coding for the protein He realized the potential problems that could lead to rrors:The electrophoresis strips had to be read manually and peatedly;a single nissed or e insertednuccotidcewodreih an entirely wrong amino acid yond that point;two compen ating erro would lead put pp四3冲平p3sauo3ue甲1pRp0北1oh0o3ua6as00eo1 ww.annualrecieusorg.Strucure Determination of Natural Prodncts 9

AC08CH01-Biemann ARI 10 June 2015 13:25 Industries of Manchester, UK, had commercialized as the MS-9. Recording onto magnetic tapes was attempted, but interfacing the spectrometer with a computer was accomplished only years later. The ability to record a complete mass spectrum within a minute or less enabled us also to interface the gas chromatograph to the mass spectrometer in an efficient and practical way (15). Previous efforts were hampered by the lack of a suitable way of recording the spectra. Gohlke (24) had used a time-of-flight mass spectrometer (TOF MS) and an oscilloscope, the screen of which was photographed with a PolaroidR camera. This limited it to quite small molecules eluting at relatively large time intervals. Ryhage (25) interfaced the GC with a conventional mass spectrometer, the slow scan speed of which limited the system to the recording of data in the single-ion monitoring mode. Using the photoplate, we were able to record the high-resolution mass spectra of 16 alkaloids eluting from a single injection of the crude extract from the bark of Aspidosperma quebracho blanco. Three of these were known alkaloids; the molecular structure of the remaining 13 we determined from their high-resolution mass spectra (26). Within two decades, we elucidated the structures of more than 40 indole alkaloids, culminating—in collaboration with chemists from the Lilly Research Laboratories of Indianapolis, IN—in the structures of the dimeric indole alkaloids vinblastine and vincristine (27), the first anticancer drugs, which are still being used today in chemotherapy. By the mid-1980s the search for a natural product for medical use subsided. Thanks, at least in part, to mass spectrometry, most major plant alkaloids were known and had been pharmacologically tested. Rather than laboriously isolating biologically active compounds from natural sources, it became general practice to generate complex mixtures—libraries—of molecules by chemical reactions of a mixture of related molecules with one or a few reactants, followed by the isolation of biologically active products, if any. Mass spectrometry also plays an important role in this area of biopharmaceutical research. 5. LARGE PROTEINS In the meantime, our work on the primary structure of proteins had continued, but so had other methodologies. The by far most useful technique was the stepwise removal and identification of the N-terminal amino acid developed by Pehr Edman (28), soon automated (29) and commercialized. It had some drawbacks: It did not work with proteins that are acylated at the N-terminus, which many mammalian proteins are, and it had difficulties if the C-terminal end consisted of a number of nonpolar, hydrophobic amino acids. Because these two situations do not interfere—in fact were even an advantage with our mass spectrometric approach—it was logical to combine the two. This was done in collaboration with Gobind H. Khorana, professor of biology at MIT, who was working on the structure and activity of bacteriorhodopsin from Halobacterium halobium, a protein that loops seven times through the cell wall and thus has long stretches of hydrophobic amino acids. This protein turned out to be 248 amino acids long (30). By the late 1970s, DNA sequencing methods had progressed to the point that it became feasible to determine the nucleotide sequence of the DNA coding for a protein and translate it using the genetic code into the corresponding amino acid sequence of the protein. Paul Schimmel, professor of biology at MIT at the time, had isolated from Escherichia coli alanyl-tRNA synthetase, a very large protein, and was interested in its structure and active site. He was planning to use the method of Maxam and Gilbert (31) to determine the DNA sequence of the strand coding for the protein. He realized the potential problems that could lead to errors: The electrophoresis strips had to be read manually and repeatedly; a single missed or erroneously inserted nucleotide would result in an entirely wrong amino acid sequence beyond that point; two compensating errors would lead to a protein sequence correct at both ends, but with an incorrect stretch in the middle; and even www.annualreviews.org • Structure Determination of Natural Products 9 Annual Rev. Anal. Chem. 2015.8:1-19. Downloaded from www.annualreviews.org Access provided by 45.58.110.168 on 08/27/16. For personal use only.

if the nucleotide sequence is entirely correct,it can be read in three different reading frames,of which only one represents the correct protein.When Paul discussed these problems with me,it occurred to me that all these errors can be easily detected and corrected by determining the amino acid sequence of a number of relatively small peptides scattered across the entire protein.For this. our GCMS technique was very well suited.Using a simple computer program,we generated the hypothetical an tched thes ence corresp onding to each of the three reading fra es and then h the di-tonhee it of the coor some peptides that region.In this way,the primary structure of the alanyl-tRNA synthetase was determined.It turned out to be 875 amino acids long (32). Word about this novel approach got around quickly,and even before this work was published lete ary at Y RNA synthe e from E.coli.While we by Michael Barber (33)appeared tha drast ally change mas spectrome approach to peptide and protein sequencing.Bombardingalarge peptide lved in glycerol with a beam of argon atoms at keV energy produced protonated molecules(M+H)+.Thus,fast atom bombardment(FAB)made it possible to measure the molecular weight of sizeable peptides,as demonstrated with met-lys-bradykinin(mol.wt.1318).We immediately implemented thisioniza- tion method on our mass spectrometers and continued the work on glutaminyl-tRNA synthetase vith the ise of a tryptie digest of the protein and matched the molecuar eights of the result- ing pepti to th tryptic peptide ding to the frames of the nucleotide sequence smaller GCMS peptides,except that we now could cover larger stretches of the DNA sequences This was the reverse of the approach now termed peptide mass fingerprinting in proteomics (34) for the identification of a protein using the now known human or other genomes.The primary structure of Gln-tRNA synthetase was quickly determined by this approach;it turned out to be 550 amino acids long (35).Four m inoacyl-tRNA synthetases,ranging in size from 324 to 990 amino acids,followed soon thereafter(36-39) 6.TANDEM MASS SPECTROMETRY Although FAB mass spectra exhibited very abundant(M+H)signals,they showed little fragmen tation,unless a large sample of pure material was used.This was an advantage,because one could me ure the molec ular weights of peptides in simple mixtures without being confused by fragn ions,but on ould not dedu mino acid nces.To inducefr ion tocause it to fragment.Thi onthan inesuchshelm,intandm mass pectromeer The ion to be fragmented is isolated by the first mass spectrometer(MS-1),which passes it through a cell filled with the collision gas and mass-analvzes the resulting fragment ions in the second mass spectrometer(MS-2).To achieve good resolution and sensitivity,two high-resolution spectrom- eters need to be used.Fred MeLafferty at Comnell University of Ithaca,NY,had constructed such an inst ment by 1980 (40),and in 1983 a similar MS/MS (ZAB 4F)fro m VG Analytical of manche d at he National In utes of Health and Environmental S (Research Triangle P k,NC)(41).I had seen that instrument in operatio on and de tha it would be very useful for our ongoing work on proteins.I also knew that JEOL of Akishima Japan,had just come out with a new design for a double-focusing mass spectrometer(HX 110) (42)and made arrangements with that company to build an MS/MS,using these new ion optics, 10

AC08CH01-Biemann ARI 10 June 2015 13:25 if the nucleotide sequence is entirely correct, it can be read in three different reading frames, of which only one represents the correct protein. When Paul discussed these problems with me, it occurred to me that all these errors can be easily detected and corrected by determining the amino acid sequence of a number of relatively small peptides scattered across the entire protein. For this, our GCMS technique was very well suited. Using a simple computer program, we generated the hypothetical amino acid sequence corresponding to each of the three reading frames and then matched these with the di- to pentapeptides we had identified in a partial acid hydrolysate of the protein. When some peptides matched in one reading frame, and some in another, the error could be identified and corrected by reinspection of the electrophoresis strip corresponding to that region. In this way, the primary structure of the alanyl-tRNA synthetase was determined. It turned out to be 875 amino acids long (32). Word about this novel approach got around quickly, and even before this work was published we started a collaboration with Peter Soll, professor of biology at Yale University, on glutaminyl- ¨ tRNA synthetase from E. coli. While we were working on this protein with the GCMS methodology, a paper by Michael Barber (33) appeared that drastically changed the mass spectrometric approach to peptide and protein sequencing. Bombarding a large peptide dissolved in glycerol with a beam of argon atoms at keV energy produced protonated molecules (M+H)+. Thus, fast atom bombardment (FAB) made it possible to measure the molecular weight of sizeable peptides, as demonstrated with met-lys-bradykinin (mol. wt. 1318). We immediately implemented this ionization method on our mass spectrometers and continued the work on glutaminyl-tRNA synthetase with the use of a tryptic digest of the protein and matched the molecular weights of the resulting peptides to the tryptic peptides expected for the proteins corresponding to the three reading frames of the nucleotide sequence. Errors were detected and corrected as done earlier with the smaller GCMS peptides, except that we now could cover larger stretches of the DNA sequences. This was the reverse of the approach now termed peptide mass fingerprinting in proteomics (34) for the identification of a protein using the now known human or other genomes. The primary structure of Gln-tRNA synthetase was quickly determined by this approach; it turned out to be 550 amino acids long (35). Four more aminoacyl-tRNA synthetases, ranging in size from 324 to 990 amino acids, followed soon thereafter (36–39). 6. TANDEM MASS SPECTROMETRY Although FAB mass spectra exhibited very abundant (M+H)+ signals, they showed little fragmentation, unless a large sample of pure material was used. This was an advantage, because one could measure the molecular weights of peptides in simple mixtures without being confused by fragment ions, but one could not deduce their amino acid sequences. To induce fragmentation, additional energy had to be imparted onto the protonated molecular ion to cause it to fragment. This is accomplished by collision with an inert gas, such as helium, in a tandem mass spectrometer (MS/MS). The ion to be fragmented is isolated by the first mass spectrometer (MS-1), which passes it through a cell filled with the collision gas and mass-analyzes the resulting fragment ions in the second mass spectrometer (MS-2). To achieve good resolution and sensitivity, two high-resolution spectrometers need to be used. Fred McLafferty at Cornell University of Ithaca, NY, had constructed such an instrument by 1980 (40), and in 1983 a similar MS/MS (ZAB 4F) from VG Analytical of Manchester, UK, was installed at the National Institutes of Health and Environmental Sciences (Research Triangle Park, NC) (41). I had seen that instrument in operation and decided that it would be very useful for our ongoing work on proteins. I also knew that JEOL of Akishima, Japan, had just come out with a new design for a double-focusing mass spectrometer (HX 110) (42) and made arrangements with that company to build an MS/MS, using these new ion optics, 10 Biemann Annual Rev. Anal. Chem. 2015.8:1-19. Downloaded from www.annualreviews.org Access provided by 45.58.110.168 on 08/27/16. For personal use only

点击进入文档下载页（PDF格式）

共21页，试读已结束，阅读完整版请下载

点击下载（PDF格式）

浏览记录