sign in sign up
<<
Home , Corrin, Heme, Hydroxymethylbilane, Porphyrin

                                                                    ! "# $  #%## & '  ()  *   (  + ,-%

   .                           (                           /         %        /              0    1         2 (                    % 3                                   4           % 4 (       /                                        %  1                 /                    %3   4  / 56(    / /   78# (/       /   2                     %               /                  %3               ( /      /       / %             /     /     //   %   /                "$  %3              2        /    1   % 9          / /               /      /   %3         / 8":/              /    /   %    4          (           %   /     7#8         ;"%$!< "  /        4  %-                 "        8;%,!<          / =8 # %3 /                % >      /   /       78# 56                 %  /           (    /                     % 9/  (       /                /                % >     (     /                                      %   (         2       (                                %

                      

  "# ;  ?<< %% <  @ A ?  ? ??  7 7#8

>-B ;$ ;; ;#; $ >-B ;$ ;; ;#$#7

                

     ( #,   

PORPHYRINS AND IN MICROORGANISMS

Jonas Fyrestam

Porphyrins and heme in microorganisms

Porphyrin content and its relation to phototherapy and antimicrobial treatments in vivo and in vitro Jonas Fyrestam ©Jonas Fyrestam, Stockholm University 2018

ISBN print 978-91-7797-079-8 ISBN PDF 978-91-7797-080-4

Printed in Sweden by Universitetsservice US-AB, Stockholm 2017 Distributor: Department of Environmental Science and Analytical Chemistry List of papers This thesis is based on the following papers, which will be referred to by Roman numerals,

I. Fyrestam, J., Bjurshammar, N., Paulsson, E., Johannsen, A., & Östman, C. (2015). Determination of porphyrins in oral by liquid chromatography electrospray ionization tandem mass spectrometry. Analytical and bioanalytical chemistry, 407(23), 7013-7023.

II. Fyrestam, J., Bjurshammar, N., Paulsson, E., Mansouri, N., Johannsen, A., & Östman, C. (2017). Influence of culture conditions on production in Aggregatibacter actinomycetemcomitans and Porphyromonas gingivalis. Photodiagnosis and , 17, 115-123.

III. Fyrestam, J., & Östman, C. (2017). Determination of heme in microorganisms using HPLC-MS/MS and cobalt(III) protoporphyrin IX inhibition of heme acquisition in Escherichia coli. Analytical and Bioanalytical Chemistry, [Epub ahead of print].

IV. Fyrestam, J., & Östman, C. (2017). Escherichia coli is killed by 405 nm blue light due to its endogenous porphyrins induced by 5- aminolevulinic . Manuscript.

V. Bjurshammar, N., Malmqvist, S., Johannsen, G., Boström, EA., Fyrestam, J., Östman, C., & Johannsen, A. The Effect of Local Phototherapy on Gingival Inflammation - A Randomized Controlled Study. Manuscript. The contribution of Jonas Fyrestam to the papers in this thesis was as follows:

I. The respondent was responsible for design of the study, parts of the experimental work, all data analysis, and significant parts of the writing.

II. The respondent was responsible for generating the idea, design of the study, parts of the experimental work, all data analysis, and significant parts of the writing.

III. The respondent was responsible for generating the idea, design of the study, all of the experimental work, all data analysis and significant parts of the writing.

IV. The respondent was responsible for generating the idea, design of the study, all of the experimental work, all data analysis and all of the writing.

V. The respondent was involved in generating the idea, parts of the design of the study and minor parts of the writing. Contents

List of papers...... vii

Abbrebrevations ...... 11

Populärvetenskaplig sammanfattning ...... 13

Introduction and aims...... 16

Part 1. Background ...... 19 Phototherapy...... 19 Porphyrins and heme ...... 19 Porphyrins as photosensitizers in phototherapy ...... 22 of porphyrins and heme in bacteria ...... 25 Heme acquisition as a potential drug target ...... 30 Oral bacteria ...... 31 Aggregatibacter actinomycetemcomitans ...... 31 Porphyromonas gingivalis...... 32

Part 2. Determination of porphyrins and heme ...... 34 Samples and sampling...... 34 Cell lysis...... 35 Extraction and clean-up ...... 36 Solid phase extraction of porphyrins ...... 37 Optimization of heme extraction ...... 39 precipitation...... 42 Salting-out assisted liquid-liquid extraction ...... 45 Liquid chromatography of porphyrins ...... 45 HPLC columns...... 46 Mobile phases ...... 48 Mass spectrometry of porphyrins ...... 50 Stability of porphyrins and heme in solution ...... 53

Part 3. Applications ...... 56 Photo inactivation of bacteria ...... 56 Pilot study...... 56 In vitro experiments ...... 57 Porphyrin content in microorganisms ...... 59 Influence of culture conditions...... 60 Time of culturing ...... 61 Passage ...... 62 Cultivation medium ...... 63 Influence of different additives on E. coli ...... 64 5-aminolaevulinic acid ...... 64 ...... 67 Cobalt protoporphyrin IX...... 67 Singlet production of porphyrins ...... 68 Clinical study...... 71

Conclusions and future perspectives ...... 74

References...... 77

Acknowledgements...... 86 HapPhy Days? A cohort study of the happiness during PhD studies ...... 86 Abbrebrevations 5-ALA 5-aminolaevulinic acid 5-ALA-S 5-aminolaevulinic acid synthase aPT Antimicrobial phototherapy BOP Bleeding on probing CCF Central composite face design CFU Colony forming unit COP-O Coproporphyrinogen Co-PPIX Cobalt(III) protoporphyrin IX EDTA Ethylenediaminetetraacetic acid FECH GC Gas chromatography GI Gingival index HILIC Hydrophilic interaction chromatography His L- HMB HPLC High performance liquid chromatography IC Internal conversion ISC Inter system crossing LED Light emitting diode LLE Liquid-liquid extraction LOD Limit of detection logD Octanol-water partition coefficient at a specific pH logP Octanol-water partition coefficient LOQ Limit of quantification m/z Mass-to-charge ratio MS Mass spectrometry MS/MS Tandem mass spectrometry 11 1 O2

OD600 Optical density at 600 nm PBG PBG-D Porphobilinogen deaminase PDT Photodynamic therapy PFP Pentafluorophenyl PROT-O Protoporphyrinogen oxidase PT Photo therapy PI Plaque index pI Isoelectric point RNO N,N-Dimethyl-4-nitrosoaniline ROS Reactive oxygen species

S0 Ground state

*Sn Excited singlet state S/N Signal-to-noise ratio SIM Selected monitoring SPE Solid phase extraction SRM Selected reaction monitoring

*T1 Excited triplet state TPC Total porphyrin content URO III-S III synthase URO-D Uroporphyrinogen decarboxylase UV-Vis Ultraviolet-visible WHO World health organization ĭ Quantum yield

ĭǻ Quantum yield of singlet oxygen

12 Populärvetenskaplig sammanfattning

Ett av de främsta hoten mot människans hälsa är bakteriers ökande motståndskraft (resistens) mot antibiotika. Enligt Världshälsoorganisationen, WHO, är situationen nu så allvarlig att den hotar hela den globala hälsan och om resistensen fortsätter att öka så kan transplantationer och cancerbehandlingar inte längre genomföras utan stora infektionsrisker. Ett av de främsta sätten att begränsa spridningen av resistens är att minska användandet av antibiotika, där kan förebyggande åtgärder för en ökad hygien vara mycket effektivt. Om en infektion redan har brutit ut så kan alternativa behandlingsmetoder minska antibiotikaanvändningen. I slutet på 1800-talet fick fototerapi ett genombrott då dansken Niels Ryberg Finsen upptäckte att man kunde behandla hudtuberkulos med UV- ljus. För denna upptäckt fick han nobelpriset i medicin år 1903. I och med Alexander Flemings upptäckt av penicillinet år 1928, försvann intresset för för fototerapi helt, men har på senare år fått mer uppmärksamhet på grund av den ökande antibiotikaresistensen. För att en fototerapeuptisk behandling ska lyckas krävs tre huvudsakliga ingredienser: ljus, ett ämne som kan ta upp ljus (fotosensiterare) och syre. När en bakterie bygger en molekyl för att förvara järn inuti cellen så tillverkar den en rad olika ämnen för vilka den egentligen inte har någon biologisk användning av. Dessa kallas porfyriner och har en bra förmåga att ta upp blått ljus och en hypotes är att dessa porfyriner kan användas för att döda bakterier. Det övergripande målet med denna avhandling har varit att bestämma förekomsten av dessa porfyriner i olika typer av bakterier, studera hur väl de kan fungera som fotosensiterare och ifall mängden porfyriner kan kopplas ihop med hur känslig en bakterie är för fototerapi.

13 Att bestämma förekomst och halt av ett kemiskt ämne som finns i en bakterie är inte helt enkelt. Koncentrationen av porfyriner i bakterierna är mycket låga. Ofta handlar det om några hundra pikogram i en normal bakteriekultur, det vill säga miljarddelar av ett gram, eller 0,000000000001 gram. Man måste därför utveckla kemisk analytiska metoder för att kunna hitta och vara helt säker på att det är rätt ämne som man analyserar, samt mäta dessa låga halter och göra noggranna bestämningar av koncentrationen. I den första publikationen utvecklades en metod för att kunna mäta porfyriner i orala bakterier med hjälp av vätskekromatografi och masspektrometri. Denna metod användes sedan i den andra publikationen där vi lät bakterierna växa i olika miljöer och olika länge för att undersöka om detta påverkade innehållet av porfyriner. Vi kunde där visa att porfyrininnehållet inuti bakterierna varierar mycket beroende på hur bakterierna odlas. Detta skulle kunna vara en förklararing till varför resultaten från olika fototerapiexperiment beskrivna i litteraturen varierar. En standardisering av hur bakterierna kultiveras behövs därför för att kunna dra korrekta slutsatser från sina laborativa fototerapiexperiment. I den tredje publikationen utvecklades en metod för analys av hem, molekylen som förvarar järn inne i bakterien. Hem kan inte användas som fotosensiterare, men eftersom det är en viktig molekyl för bakterierna så kan ett lovande alternativ till antibiotika vara att man hindrar bakterien att få tillgång till denna. När ett kemiskt ämne som är väldigt likt hem gavs till bakterierna kunde man se att bakterierna tror att det är hem och tar upp detta ämne, likt en trojansk häst, vilket medförde att hemkoncentrationen minskade i bakterierna. Detta skulle därför kunna vara ett möjligt framtida alternativ till antibiotika. I det fjärde manuskriptet undersöktes hur bra bakterien E. coli kan dödas med blått ljus. I dessa experiment analyserades bakteriernas

14 porfyrininnehåll, samt undersöktes hur bra olika porfyriner fungerar som fotosensiterare. Resultaten visade att E. coli inte är känslig mot blått ljus och att den tillverkar väldigt små mängder av egna porfyriner. När ett ämne som bakterierna kan bygga porfyriner av tillsattes medförde detta till att E. coli började producera mycket porfyriner. Bakterierna belystes efter tillsatsen av detta ämne och det visade sig att relativt låga doser av blått ljus nu kunde döda upp till 99,9999 % av alla bakterier. Femte manuskriptet är en randomiserad klinisk studie där 48 personer delades in i två grupper, en kontrollgrupp och en behandlingsgrupp. Under 8 veckor fick behandlingsgruppen borsta tänderna med en tandborste med en inbyggd blå lampa i borsthuvudet. Kontrollgruppen använde samma tandborste fast utan blått ljus. Flera olika odontologiska parametrar undersöktes före, under och i slutet av studien. Man kunde notera en minskning av mängden plack och hur lätt tandköttet börjar blöda vilket tyder på en positiv effekt av det blå ljuset, men det gick inte att påvisa någon statistisk skillnad mellan kontroll- och behandlingsgruppen. En förklaring till detta kan vara att lysdioden i den tandborsten hade en våglängd på 450 nm och inte 405 nm vilket är den våglängd där porfyriner har sin maximala ljusabsorption.

15 Introduction and aims One of the greatest threats to the human health is increasing antimicrobial resistance among pathogens. According to the World Health Organization (WHO), this threatens the achievements of modern medicine [1]. Antibiotics are for example used in major surgery, organ transplantation, treatment of premature babies, and cancer chemotherapy. These are medical treatments which cannot be performed in a safe way if antibiotics not are effective enough [2]. Prevention of bacterial infections is one of the most efficient ways to reduce antibiotic use and this can be done by increased hygiene and sanitation improvements. However, preventive actions alone will not solve the problem. Thus finding alternatives for treatment of bacterial infections is of highest importance besides a more controlled use of antibiotics. Major research efforts have been launched in recent years to identify alternative antibacterial therapies. Phototherapy has been suggested as one of these alternatives, where visible light can be used to inactivate a number of different bacteria without the addition of an exogenous compound [3-5]. In order for a phototheraputical treatment to be successful, three components are necessary; light, oxygen and a photosensitizer. It is the photosensitizers that have been the subject for attention in this thesis. Porphyrins, intermediates of the biosynthesis of heme, are believed to work as endogenous photosensitizers [6,7]. Porphyrins absorb light in the blue part of the visible part of the electromagnetic spectrum and are capable to transfer this energy to oxygen that will create reactive oxygen species (ROS). However, the porphyrin content in bacteria has been sparingly investigated and it is difficult to compare these studies since the cultivation conditions among studies differ. Furthermore, the sensitivity for a specific microorganism to blue light has not been correlated to the porphyrin content,

16 1 as well as to the singlet oxygen ( O2) production from different porphyrins [8,9]. The primary aim of this thesis was to develop sensitive and selective analytical methods for determination of porphyrins and heme in microorganisms. These methods were then used to determine the heme and porphyrin content in bacteria in order to better interpret the results from phototherapy experiments. In Paper I, a method was developed for the extraction, clean-up and determination of porphyrins in oral bacteria using solid phase extraction with subsequent analysis by high performance liquid chromatography coupled to tandem mass spectrometry (HPLC-MS/MS). Several evaluation parameters were determined, such as linearity, limits of detection and quantification, accuracy, intra-day precision, recovery and matrix effects. In Paper II, the method from Paper I was used to analyze two different types of oral pathogens, Aggregatibacter actinomycetemcomitans and Porphyromonas gingivalis, for their porphyrin content. The aim was to investigate if the profile of endogenously produced porphyrins varies according to changes in cultivation parameters. In Paper III, a method for the determination of heme was developed by using experimental design for the optimization of a liquid-liquid extraction method for heme in Escherichia coli. The method was used to monitor the heme content when different additives related to heme biosynthesis were added to the cultivation broth. Further, a compound that is chemically similar to heme was added to investigate if such a compound could interfere with the heme acquisition mechanisms in E. coli. Heme acquisition mechanisms in bacteria could be a potential target for drugs. In Paper IV, E. coli was cultivated and treated with 405 nm light and analyzed for its porphyrin content. The precursor of porphyrins, 5-

17 , was used to bypass the negative feedback control of the heme biosynthesis and enhance the porphyrin production in E. coli. The photosensitivity and porphyrin content before and after administration of 5- aminolevulinic acid was compared, and the relative singlet oxygen production for three naturally occurring porphyrins was also determined. Paper V is a randomized clinical trial in which a toothbrush equipped with a 450 nm light-emitting diode (LED) was used for the subject’s daily oral hygiene during 8 weeks. The control group was compared to the treatment group for a set of inflammation markers and periodontal parameters.

18 Part 1. Background

Phototherapy In the late 19th century, phototherapy achieved a breakthrough when the Dane Niels Ryberg Finsen (1860-1904) found that tuberculosis of the skin could be treated with UV-light. These findings led to the Nobel Prize in medicine in 1903 [10]. When Sir Alexander Fleming (1881-1955) in 1928 discovered penicillin [11] the interest in phototherapy was lost, but has gained more attention in recent years due to increasing antibiotic resistance [1]. To succeed with a phototherapy treatment, three components are needed: light, a photosensitizer and molecular oxygen [12]. There are a number of publications where bacteria and other microorganisms have been shown to be inactivated after illumination of visible light [3-5,7,13-16]. There are generally two branches in phototherapy. In photodynamic therapy (PDT) a photosensitizer is added exogenously to the bacteria which then are illuminated with an appropriate wavelength, often in the red light spectra [17,18]. The other one uses endogenously produced photosensitizers which are naturally produced by the bacteria and microbes, usually called antimicrobial phototherapy or antimicrobial photodynamic therapy (aPT) [19].

Porphyrins and heme Porphyrins are macrocyclic compounds that occur naturally and have several important representatives, such as and heme [20,21]. All porphyrins share the same basic structure, a , where the four S\UUROHVDUHOLQNHGWRJHWKHUZLWKPHWKLQH &+í EULGJHV, Figure 1.

19 HOOC HOOC HOOC COOH COOH CH3 COOH COOH COOH H C HOOC HOOC 3 N NH N NH N NH

HN COOH N HN N HN COOH N HOOC COOH HOOC HOOC

H3C COOH COOH COOH COOH COOH

Uroporphyrin I Heptacarboxylic porphyrin I Hexacarboxylic porphyrin I

HOOC HOOC CH2 CH3 CH3 CH3 CH2 H3C COOH COOH H3C NH HOOC N NH N NH N N HN N N HN HN H3C CH HOOC HOOC 3 CH3 CH3

H3C H3C HOOC COOH COOH COOH Pentacarboxylic porphyrin I Coproporphyrin I Protoporphyrin IX

CH2 CH2 CH3 CH3 CH2 CH2 H3C H3C N N N Cl N Fe Fe N N N N

H3C CH3 H3C CH3

HOOC COOH HOOC COOH

Heme Hemin

Figure 1: Trivial names and structures for porphyrins, heme and hemin. Heme contains ferrous (Fe2+) while hemin is the oxidized form of heme containing ferric iron (Fe3+).

Naturally occurring porphyrins are characterized by their side chain substituents and are often named by the number of carboxylic acid groups. Other functional groups are methyl, ethyl and vinyl. Due to their common

20 basic structure, porphyrins have similar bands of absorbance in the region between 390-750 nm. They have an intense band of absorbance between 390-425 nm, the Soret band, and two to four weaker bands in the region between 480-700 nm, the Q bands [22]. In Figure 2 the absorption spectra of coproporphyrin III is shown.

1

0.5 Normalized absorbance

0 300 400 500 600 700 Wavelength [nm]

Figure 2: Absorbance spectra for coproporphyrin III showing the intense band of absorbance between 390-425 nm and four weaker bands between 480-700 nm.

The Soret band around 400 nm is often used for detection with UV-Vis due to the high absorption in this region [23]. Most porphyrins also exhibit characteristic fluorescence spectra in the region around 600 nm [22], and since fluorescence detection has 1-3 orders of magnitude higher sensitivity compared to UV-Vis it is the most frequently used detection method [24-27].

21 Porphyrins as photosensitizers in phototherapy In this work we have focused on the endogenous porphyrins role in aPT experiments. It has been shown that light with an emission wavelength in the Soret region of the absorption spectra of porphyrins is efficient for killing bacteria [14,28-31]. The illumination wavelengths used in such studies range between 405-460 nm [3,4,14-16,28-37]. Two studies have used different wavelength in order to determine if some wavelengths are better at killing bacteria. In those studies it was shown that illumination wavelengths > 420- 430 nm was not sufficient to inactivate certain bacteria [16,31]. However, when using light covering the Soret spectral band of porphyrins, i.e. wavelengths < 420-430 nm) the bacteria were inactivated. Only a few papers have determined the porphyrin content together with the photosensitivity in the selected microorganism. When this has been done, often nonspecific detection methods have been used such as UV-Vis or fluorescence and the analytical methods used often lack satisfactory validation [7,14,29,32,36]. Furthermore, the bacterial extracts are rarely subjected to separation with HPLC for detection of the individual porphyrin species, instead the extracts of the investigated bacteria are determined as a mixture [3,28,38]. The fluorescence spectra of different porphyrins are rather similar, and the porphyrin profile in these samples will be impossible to distinguish accurately. This also results in that no conclusion can be drawn regarding which porphyrin will have the major contribution to the photosensitivity of the bacteria. To be able to act as a photosensitizer, a compound must have a GHORFDOL]HG ʌ-electron system which can absorb a photon of a certain wavelength [39]. Upon absorption, a valence electron is raised from its ground state (S0) to an excited singlet state (*S1 and *S2, Figure 3). By internal conversion (IC) the photosensitizer will relax into the excited state 22 molecular oxygen [41]. is not particular reactive to biological systems, but can however react with itself to form more reactive species such as hydrogen peroxide [41]. In the type 2 reaction the ground state triplet oxygen accept the energy from the photosensitizer and become excited and go into its higher energy state; singlet oxygen. Singlet oxygen is highly reactive and will immediately react with its surroundings and e.g. oxidize cellular components leading to cytotoxicity. The singlet oxygen is believed to be the major damaging species in phototherapy [5,42-45]. Singlet oxygen has a very short lifetime in aqueous solutions, around 4 μs [46-48], due to waters high solvent deactivation properties via electronic-vibrational energy transfer [49]. In that timeframe the singlet oxygen diffuses about 100 nm which means that the photosensitizer needs to be close to a biologically important system to cause photo toxicity [46]. A compound that will generate singlet oxygen upon illumination is not a good photosensitizer per se. It is important that the photosensitizer is efficient in producing singlet oxygen relative to the number of absorbed photons. Einstein introduced the concept of quantum yield (ĭ) for a photochemical reaction, also known as the Stark-Einstein Law of Photochemical Equivalence [50]:

୬୳୫ୠୣ୰ ୭୤ ୣ୴ୣ୬୲ୱ ߔ(ɉ) = ୬୳୫ୠୣ୰ ୭୤ ୮୦୭୲୭୬ୱ ୟୠୱ୭୰ୠୣୢ

The quantum yield at a specific wavelength is the fraction between the number of events and the number of photons absorbed [51]. The event may be different kinds of photophysical processes (e.g., intersystem crossing, fluorescence and phosphorescence) or photochemical reactions (e.g., and singlet oxygen production). In an ideal situation for aPT 24 applications, one absorbed photon will be able to generate one singlet oxygen . Therefore the quantum yield has a value between 0 and 1. In Paper IV, the quantum yield of singlet oxygen formation for three different naturally occurring porphyrins was determined using a singlet oxygen probe.

Biosynthesis of porphyrins and heme in bacteria In this section the well-studied biosynthetic pathway of heme is explained briefly. For a more thorough review of heme biosynthesis, see [52]. The biosynthesis of heme has different biological pathways depending whether heme is synthesized in prokaryotic (C5-pathway) [53] or in eukaryotic (Shemin pathway) [54]. In both these pathways 5-aminolaevulinic acid (5-ALA) is produced, but in different ways. The compound 5- aminolaevulinic acid is the first building block of porphyrins and heme, Figure 4. SHEMIN PATWAY

O H2N + HOOC SCoA COOH Succinyl-CoA O H N 2 COOH C5 PATHWAY 5-ALA 5-ALA-S NH2 O Glutamyl-tRNA HOOC H Glutamate-1-semialdehyde

Figure 4: Formation of 5-ALA by the Shemin and C5 pathway. 25 In the C5 pathway, 5-ALA is formed from glutamate and glutametyl-tRNA, and in Shemin pathway from succinyl and glycine where it is catalyzed by the 5-aminolaevulinic acid synthase (5-ALA-S). The formation of 5-ALA is the rate-determining enzymatic step in heme biosynthesis and is tightly regulated by negative feedback control of heme [38,55]. Pure 5-ALA can be added to the bacterial cultures in order to bypass this negative feedback control [42,56]. If a bacterium has all the necessary for the continuous synthesis of heme, the bacteria will start an excessive production of porphyrin intermediates. Addition of 5-ALA to bacteria was performed in Paper IV in order to increase the intracellular content of porphyrin and investigate how the photosensitivity of E. coli was affected. Two 5-ALA are enzymatically catalyzed by 5-aminolaevulinic acid dehydratase to condense with each other and form porphobilinogen (PBG), a monopyrollic precursor, Figure 5.

HOOC COOH COOH COOH

+ O 5-ALA dehydratase O NH N 2 H H N H2N 2

5-ALA PBG

Figure 5: Formation of porphobilinogen from two 5-ALA molecules.

Porphobilinogen is the corner stone of porphyrins and heme and by condensation of four PBG molecules, the linear and symmetrical hydroxymethylbilane (HMB) is formed, Figure 6. Porphobilinogen deaminase is the enzyme that catalyzes the formation of HMB. 26 HOOC HOOC HOOC COOH COOH COOH NH HN PBG-D HO NH HN HOOC N H COOH H2N HOOC COOH

PBG HMB

Figure 6: Biosynthesis of hydroxymethylbilane from PBG.

Hydroxymethylbilane can react in two ways; enzymatically catalyzed by the enzyme uroporphyrinogen III synthase (URO III-S) to form the asymmetrical uroporphyrinogen III, or if URO III-S is absent, the unstable HMB will form the physiologically unimportant by spontaneous ring closure, Figure 7 [22]. At this stage of the heme biosynthesis, and also in the upcoming steps, are formed, not porphyrins. Porphyrinogens are non-fluorescent compounds which cannot act as photosensitizers themselves, but the oxidation products of porphyrinogens are the corresponding porphyrin [22]. Porphyrins are with one exception byproducts in the heme biosynthesis and they will not be further metabolized. Uroporphyrinogen III contains four groups which are subjected to stepwise decarboxylation into methyl groups by the enzyme uroporphyrinogen decarboxylase (URO-D), forming hepta-, hexa- and pentacarboxylic acid porphyrinogens of type III . However, the URO-D enzyme is not specific at which ring this decarboxylation starts, leading to a complex mix of different of these porphyrinogens. Nor are URO-D specific to uroporphyrinogen III, and the

27 same decarboxylation process can be seen for uroporphyrinogen I [22]. In the last decarboxylation step, catalyzed by URO-D, the coproporphyrinogen III is formed, containing four groups. Coproporphyrinogen oxidase (COP-O) converts two of the propionic acid groups into vinyl groups by oxidative decarboxylation, forming protoporphyrinogen IX. COP- O is highly specific to the coproporphyrinogen III isomer, so will not further be subject for enzymatic transformation [22]. By oxidation, protoporphyrinogen IX is converted to protoporphyrin IX, the only porphyrin that is actually part of the biosynthetic pathway of heme. This transformation is catalyzed by the enzyme protoporphyrinogen oxidase (PROT-O). The final step in the heme biosynthesis is the insertion of ferrous iron into the protoporphyrin IX ring structure by the enzyme ferrochelatase (FECH). Heme is rather toxic to cells so its synthesis is tightly regulated by a negative feedback control where heme inhibits the production of 5-ALA as mentioned above [38,55]. The killing efficiency in different aPT experiments varies and one possible explanation can be that the different culture conditions vary among studies [57]. If porphyrins make the major contribution to photosensitivity and are synthesized endogenously by the bacteria, the outcome of aPT experiments could be highly dependent on the bacterial intracellular concentration of porphyrins. In Paper II we investigated the intracellular concentrations of porphyrins when oral bacteria were cultivated in vitro using different culture conditions, such as, time of culturing, passage and different agar compositions.

28

Heme acquisition as a potential drug target Iron is an important in many biological systems and it is necessary for respiration in bacteria [58,59]. During the initial state of an infection, vertebrates have developed different iron withholding mechanisms that reduce the concentration of free iron that is available for the bacteria to negligible concentrations, < 10-18 M [60]. For example the two proteins lactoferrin and transferrin, are produced in excess [61]. Even though the concentration of free iron is low, a large reservoir of iron is present bound in heme molecules. Heme occurs in various in vertebrates, such as and . Pathogens have developed strategies to scavenge heme from these and then either transport heme into the cell or extract the iron from heme and transport it into the cell [60,61]. Since iron is essential for almost all bacteria, the heme acquiring mechanisms in pathogens have been suggested to be a potential drug target [62-67]. Limiting the access to iron could have an antimicrobial effect and one way could be to use a

N N compound that has high affinity for

Co the heme acquiring receptors on the

N N bacterial cell wall. Adding this compound to the site of treatment could block the bacterial uptake of heme and bacterial replication is

O reduced. Since the bacteria have a OH HO O high affinity for heme, a compound that will mimic heme could be useful. Figure 8: Structure of cobalt(III) protoporphyrin IX. In Paper III we investigated

30 cobalt(III) protoporphyrin IX (Co-PPIX), Figure 8, for the ability to mimic real heme and reduce the intracellular concentration of heme in E. coli. Cobalt protoporphyrin IX has previously been shown to reduce the growth of Porphyromonas gingivalis in both planktonic state and in biofilms [68]. The Co-PPIX molecule is chemically similar to heme with the difference that the essential iron (II) is replaced with the biologically insignificant cobalt (III). The hypothesis is that bacteria will take up Co-PPIX and this will potentially disrupt the heme in the bacteria.

Oral bacteria

Aggregatibacter actinomycetemcomitans A. actinomycetemcomitans, (prev. known as Actinobacillus actinomycetemcomitans) is a facultative anaerobic gram negative bacterium [69]. In 2006, A. actinomycetemcomitans was reclassified when it was shown that the bacterium is not dependent on exogenous sources of heme to be able to grow [70]. Boyce and coworkers [71] had previously categorized A. actinomycetemcomitans as an X-factor dependent bacterium (requires heme) since they could see that the bacterium was able to grow significantly better in the presence of heme. The few bacteria that were able to grow without the presence of heme were seen as minor variants of A. actinomycetemcomitans. However, A. actinomycetemcomitans has all the necessary enzymes for the de novo synthesis of heme, which is lacking in X- factor dependent bacteria [70]. Investigations of the iron source preferences for A. actinomycetemcomitans have shown that this specie has the possibility to utilize heme, hemin, FeCl3 and hemoglobin as external iron sources [72- 75]. Differences between strains are not uncommon, e.g., not all strains are

31 able to use hemoglobin as iron source [74]. Overall the uptake and utilization of iron, including the heme biosynthesis, has earlier not been well studied in A. actinomycetemcomitans. A. actinomycetemcomitans has been shown to be sensitive to blue light [3,76]. Cieplik et al. [3] illuminated A. actinomycetemcomitans with 460 nm light and observed a decrease of viable cells ZLWK•ORJ10 steps after a light dose of 150 J/cm2. In Paper I and II, A. actinomycetemcomitans were lysed and analyzed for porphyrin content using HPLC-MS/MS.

Porphyromonas gingivalis P. gingivalis is an anaerobic, gram negative bacterium that is X-factor dependent and requires external sources of iron for growth and virulence [73,77]. This specie can acquire iron from both organic sources, e.g., heme, hemoglobin, myoglobin, C, and transferrin) as well as from inorganic iron sources, e.g., FeCl3 and Fe(NO3)3 [78]. P. gingivalis lacks the enzyme 5-ALA-S which catalyzes the formation of 5-ALA. It also lacks most of the enzymes that are involved in the de novo synthesis of porphyrins and heme [79]. If 5-ALA is added to cultures of P. gingivalis, no excessive production of porphyrins can be seen [38]. P. gingivalis has been a subject for aPT experiments. Soukos and coworkers [9] investigated several black pigmented oral bacteria and their susceptibility to light in the region 380-520 nm. P. gingivalis was the least affected bacterium and also had the lowest concentration of porphyrins in the subsequent HPLC analysis. Only one porphyrin could be detected in P. gingivalis, coproporphyrin. In another study where P. gingivalis was grown with hemoglobin as an iron source instead of hemin, the sensitivity to irradiation decreased [8]. P. gingivalis is a black pigmented bacterium that becomes heavily blackened when grown in heme rather than hemin. The

32 EODFNSLJPHQWLVORFDWHGLQWKHFHOOZDOODQGFRQVLVWVRIȝ-oxo bisheme [80]. This compound has been proposed to act like an oxidative buffer, keeping an anaerobic environment inside the cell [80]. Since this compound also absorbs light in the blue region, a plausible assumption could be that the ȝ- oxo bisheme present in the cell walls protects the interior of the cell from blue light, and thus no singlet oxygen can be produced. P. gingivalis was analyzed for porphyrin content in Paper I and III.

33 Part 2. Determination of porphyrins and heme

Samples and sampling The samples that have been studied in this work are mainly bacterial cultures (Paper I-IV), but for method development purposes baker’s yeast was also used (Paper I and III). Sampling is one of the most crucial parts when analyzing porphyrins and heme in bacteria. If it is performed incorrectly, this will have severe impact on the accuracy in, as well as the precision of the reported concentrations. Even if a thorough sampling procedure is performed, there are some difficulties in sampling of bacteria, of which the estimation of sample weight is one. In most cases bacteria are cultivated on agar or in liquid broth, and there are differences in the sampling procedures. When grown in broth the cell density is often estimated by the optical density at 600 nm (OD600) using a spectrophotometer. This technique does not use the absorption of light that is normally the case when using a spectrophotometer. Instead it uses the scattering of light when a photon hits a particle or a cell. This scattering is proportional to the biomass concentration in the sample. Extra care should be addressed when dealing with samples with high density, where a multiple scattering effect may occur and the response is no longer linear to the biomass. In those cases the samples should EHGLOXWHGSUHIHUDEO\”2'7KH2'600 may then be calibrated to the dry weight of the investigated bacteria. Due to variations in size for different bacteria, this calibration should be performed for every species investigated. When growing bacteria in broth, sampling by centrifugation is the most commonly used method [6,32,77]. In Paper III and IV the sample weight of E. coli was determined by using OD600 and then centrifuged.

34 Problems arise when determinations of sample weight for bacteria grown on agar plates is necessary. In this case the bacteria form colonies which are large aggregates of many individual bacteria. The bacteria in those aggregates sticks together, therefore it is not possible to remove the bacteria from the agar plate and then place it in a liquid and determine the OD600, since this will give inaccurate estimations of the sample weight. A common way is to gravimetrically determine the sample weight by scraping off the bacteria from the agar plate and then put it in a pre-weighed test tube [81]. In this way the wet weight of the bacteria can be determined, but moisture differences may still give variations in the determined weight. In Paper I and II the sample sizes were determined gravimetrically from agar plates. The determination of sample weight is rarely mentioned in studies where the intracellular concentrations of porphyrins have been determined. Samples are instead normalized to the protein content or the optical density [9,32,82]. This complicates the comparison of the concentrations given in different studies. However, it can partly be solved by determining the porphyrin composition by using the percent distribution of the individual porphyrins [6,9,32,83], but gives no information about the absolute concentrations.

Cell lysis In order to release the intracellular content and make it available for chemical analysis, it is essential to lyse the cells. In Paper I-IV a method using a combination of a lysis buffer and ultrasonication was employed. The harvested microbial cells were first put in a Tris-EDTA buffer at pH 7.2 for one hour, protected from light. Tris/EDTA buffer has previously been shown to be effective in lysis of A. actinomycetemcomitans [84]. Tris has the

35 possibility to interact with lipopolysaccharides in the cell membrane, which causes a destabilization [85]. Divalent cations, such as Mg2+ and Ca2+, stabilize the membrane by crosslinking carboxylated and phosphorylated lipids. The chelating agent EDTA in the lysis buffer binds to these metal and destabilizes the cell membrane [86,87]. In many non-quantitative bioanalytical applications, e.g. protein purification and DNA isolation, a lysis efficiency of 100% is not of highest importance. When the aim is to perform quantification of intra cellular contents, it is important with a lysis efficiency as close to 100% as possible in order to have high accuracy in the reported concentrations A physical cell disruption method, using an ultrasonication probe, was added to the chemical lysis procedure to increase the efficiency. High frequency ultrasound energy (20-50 kHz) is applied to the sample medium through a metal probe. It creates cavities which implode and create shock waves causing the cells to rupture. To avoid excessive heating of the sample, the ultrasonication is performed in multiple short bursts and the sample should be placed in an ice-water bath. In Paper I the above mentioned method for cell lysis was investigated for its ability to lyse A. actinomycetemcomitans. After treatment with buffer and ultrasonication no viable bacteria could be seen on agar plates compared to XQWUHDWHG FRQWURO DQG WKHUHIRUH WKH O\VLV HIILFLHQF\ ZDV DVVXPHG WR EH • 99.98%.

Extraction and clean-up Two commonly used clean-up methods used for porphyrins and heme are solid phase extraction (SPE) and liquid-liquid extraction (LLE). The latter is the most commonly used method for the extraction of porphyrins from

36 plasma and red cells, as well as or studying bacterial extracts [6,22,32,44,88-90]. Solid phase extraction has mostly been used for samples and in a few studies of porphyrins in bacterial extracts, in all cases using C18 as the stationary phase [22,44].

Solid phase extraction of porphyrins Solid phase extraction was used for clean-up of bacterial extracts in Papers I and II. The major advantages of using SPE for porphyrin extraction is that it consumes low volumes of organic solvents, yields cleaner samples by removing interfering compounds, and the possibility to concentrate the analytes without excessive evaporation of extraction solvents [91]. Most microorganisms produce porphyrins in very low concentrations (ng/mg) [9]. In order to detect porphyrins in bacterial extracts it is necessary to concentrate the samples or use large sample sizes. Since many microorganisms are difficult to culture, it is often beneficial to concentrate the samples instead of using large sample sizes.

End-capped reversed phase C18 SPE column was used for clean-up of bacterial extracts in Papers I and II, due to the good retention for porphyrins. Naturally occurring porphyrins have a wide range of polarity due to differences in the number of carboxylic acid groups attached to the macrocyclic porphyrin structure. Uroporphyrin, with its eight carboxylic acid groups, is the most polar with a logP value of 3.51 at pH 1.7 (Chemspider, Chemicalize). Protoporphyrin IX, the most nonpolar porphyrin on the other hand has a logP of 6.43 at pH 4.5 which means a difference in mass transfer to octanol by a factor of around 1000. Since porphyrins are zwitterions, the partition coefficient changes dramatically at more physiologically pH. For example the logD values for uroporphyrin and protoporphyrin IX at pH 7.4 are -21.22 and 2.29 respectively (Chemspider,

37 Chemicalize). This difference makes it challenging to develop a clean-up method with recoveries close to 100% for all porphyrins. It is also difficult to find a suitable internal standard that will compensate for losses during the clean-up. To reduce the difference in the polarity of the compounds, the sample should be acidified prior to extraction and in Papers I-II 6M formic acid was used. Recoveries were investigated by passing a porphyrin standard through the clean-up steps. They ranged between 55-99%. With the lowest recoveries observed for uroporphyrin and protoporphyrin IX, 67 and 55 % respectively, Figure 9. Standard deviations were less than 10% for all investigated porphyrins, and the clean-up using SPE was considered to be acceptable for further studies.

120

100

80

60

40 Recovery [%] [%] Recovery

20

0 UP 7P 6P 5P CP MP (IS) PPIX

Figure 9: Recoveries for porphyrins using the SPE clean-up in Paper I. Error bars show the standard deviation for triplicate experiments (n=3). UP: uroporphyrin; 7P: 7-carboxylporphyrin; 6P: 6-carboxylporphyrin; 5P: 5-carboxylporphyrin; CP: coproporphyrin; MP: mesoporphyrin IX; PPIX: protoporphyrin IX.

38

Experimental design can be used to find optimal conditions by systematic alternation of the experimental settings [93]. By using experimental design the number of experiments can be reduced, and also account for variables that might be correlated. Each variable is given a minimum and maximum value, which are coded to -1 and +1 respectively. By varying each variable one at a time, a full factorial design is created, and this is illustrated by the red circles in Figure 10. Using only a full factorial design will not yield any information about the optimal settings of variables, it will only provide information about the significance of each variable [93]. There are a number of optimization designs that introduce a quadratic term into the model which will find an optimum within the experimental space. One of the most basic ones is central composite design which combines a full factorial design with a star design. The star design used in Paper III is a CCF design where the star points are at a coded distance of one from the center point and can be seen in Figure 10 as circles. The blue circle in the middle of the design shows the center point which was analyzed in triplicate to minimize the risk of missing non-linear relationships in the middle of the intervals and allowing for determination of confidence intervals [93]. In a first experiment the relative yield of heme using three different organic solvents; acetone, acetonitrile and methanol was evaluated. Acetone and HCl have been used in previous studies for the extraction of heme from different [94-97] and the extraction efficiency using this solvent was set to 1. Acetone and acetonitrile showed the same extraction properties, while methanol was only able to extract 66% of the heme relative to acetone. Pure acetonitrile and 1.7 M HCl were also evaluated and they could only H[WUDFW”RIDOOKHPH. Thus the optimization of the heme extraction was made with acetonitrile as the extractant.

40 compositions can be used for clean-up purposes without losing any heme from the sample. By using pure acetonitrile, and thus not extract any heme, lipophilic interfering substances in the lysate can be removed and provide cleaner samples. It is also possible to produce analyte free matrices that can be used for validation proposes e.g. evaluation of matrix effects.

Protein precipitation Many bioanalytical applications utilize protein precipitation as a fast sample clean-up to remove proteins from biological matrices [98]. By changing the surrounding environment, in which the proteins are stable, they can be precipitated. Protein solubility depends on ionic interaction with salts, polar interactions with aqueous solvent and electrostatic forces between charged molecules [98]. The different protein precipitation techniques are; use of organic solvents, changing the pH, metal ion and salting out [98]. By using organic solvents, the dielectric constant of the protein solution will decrease, which increases the attraction between charged molecules and the electrostatic interactions between proteins are enhanced. Organic solvents also remove the water surrounding the protein surface which decreases the hydrophobic interactions between proteins, thus enhancing protein aggregation by increasing the electrostatic interactions [99]. Protein precipitation can also be carried out by changing the pH of the protein solution. A protein has an isoelectric point (pI), i.e. the pH where the protein has a net charge of zero, and therefore it has poor solubility in aqueous solvents. At a pH less than the pI the proteins will have a positive net charge and acidic agents will form insoluble salts with the positively charged amino , leading to protein precipitation [98]. Salting out is also a common way to precipitate proteins in bioanalytical procedures. Only a few proteins are soluble in pure water and most require a

42 small concentration of ions (< 0.5 M) to remain stable. At higher concentrations of salts, the protein surface will become highly charged so proteins will interact with each other and precipitate [100]. In Paper III, methanol, acetonitrile and acetone were evaluated for their efficiency to precipitate proteins from S. cerevisiae lysate. The precipitating solvents were a mixture of organic solvents and HCl, in the same proportions that were optimized for heme extraction using acetonitrile as mentioned earlier. Pure acetonitrile as well as 1.7 M HCl were evaluated for their efficiency to precipitate proteins. For determination of protein concentration in solution, the simplest and most direct assay is to use the light absorption of peptide bonds below 210 nm [101]. One may also measure the absorbance at 280 nm, were amino acids containing aromatic side chains, i.e. tyrosine, tryptophan and phenylalanine, exhibit a strong UV-light absorption. A major disadvantage using this direct determination is that not all proteins contain these aromatic amino acids and will skew the determination of protein concentration [102]. Several reagent based protein assays have been developed instead. The method of choice for protein quantification in Paper III was the Bradford method [103]. The Bradford assay is a colorimetric protein assay based on the formation of a complex between the dye Brilliant Blue G, and proteins in the solution. When the dye is bound to the protein it will cause a shift in absorbance from 465 to 595 nm, and the amount of absorption is proportional to the protein concentration [103]. Bovine serum was used as a standard to check the linear response of the spectrophotometer. The Bradford method has many advantages. It is simple to perform, cheap, relative high sensitivity, and it will measure the absorbance in the visible light region. A major disadvantage in using the Bradford method is that it is not always compatible with some detergents, buffers and organic solvents.

43

Salting-out assisted liquid-liquid extraction Combining the results from optimized extraction of heme and the ability of different solvents to precipitate proteins, acetonitrile:HCl was chosen to be the solvent for extraction of heme and porphyrins in Paper III and IV.Due to its low extraction efficiency of heme and its good protein precipitation properties, pure acetonitrile was used for clean-up purposes. By first adding pure acetonitrile to the bacterial lysate, 100% precipitation of all proteins was performed with no extraction of heme. The sample was centrifuged to form a pellet of precipitated proteins and by discarding of the organic supernatant, unwanted hydrophobic compounds could be removed. By adding acetonitrile:HCl to the protein pellet, heme was extracted from the precipitated proteins. To remove hydrophilic interfering compounds, saturated MgSO4 solution and NaCl were added to the samples. By increasing the ionic strength of the solution, acetonitrile was no longer miscible with the aqueous solvent and a two phase liquid-liquid system was formed. Heme is hydrophobic and is partitioned into the upper acetonitrile layer. The evaluation of the matrix effects using this method for heme showed that the matrix effect was low, and the signal obtained for a spiked analyte free matrix was 106 ± 3 % compared to a standard dissolved in pure solvent.

Liquid chromatography of porphyrins Porphyrins are macrocyclic structures with inherent low volatilities and vapor pressures, which makes porphyrins unsuitable to analyze with gas chromatography (GC) [22], even though some studies have used GC to separate both metallo- and free base porphyrins [105-107]. However, the

45 recent development of high efficiency columns has made HPLC the technique of choice for porphyrin separation [22].

HPLC columns As mentioned above, porphyrins have varying arrangements and kinds of side chain substituent groups around the porphyrin’s macrocycle which gives these compounds varying degrees of hydrophobicity. These differences in hydrophobicity are suitable for the separation by reversed-phase chromatography [20]. The hydrophobicity of the side-chain substituents of porphyrins increases in the following order: CH=CH2 > CH2CH3 > CH3 >

CH2CH2COOH > CH2COOH [22]. Porphyrins have been able to be separated on reversed phase columns ranging from octadecyl carbon chain bonded silica (C18) to the least retentive of all alkyl group bonded phases, trimethylsilyl (C1) [22].

C1 columns allow fast separation of naturally occurring porphyrins, but the chromatographic resolution between porphyrins with 4-8 carboxylic acid groups is poor. C18-columns are more retentive than C1 columns and to longer analytical runs, but the resolution between porphyrins and some of their isomers will on the other hand be enhanced. The separation of a porphyrin mixture on a C18 column is shown in Figure 13. A number of different HPLC columns from a variety of manufacturers were evaluated for their ability to separate porphyrins, e.g. C18 (ACE), biphenyl (Restek), C18-PFP (ACE), Poroshell C18 (Agilent), Core-shell C18

(Phenomenex), ZORBAX HILIC Plus (Agilent), and C18-Phenyl (ACE). The biphenyl column was able to separate two isomers of 6-carboxyl porphyrin when using methanol as an organic modifier in the mobile phase which none of the other columns were able to do. When the biphenyl column was used with acetonitrile as the organic modifier in the mobile phase, these

46 two isomers co-eluted and the retention was similar to traditional C18 stationary phases. By using different organic solvents in the mobile phase, the biphenyl column can be switched between two different separation mechanisms, hydrophobic interactions using acetonitrile and ʌ-ʌ interactions using methanol. Acetonitrile will significantly weaken the ʌ-ʌ interactions between the porphyrins and the phenyl groups in the stationary phase, and when used with methanol an almost orthogonal elution compared to C18 can be achieved depending on the type of analytes [108].

PPIX

100

MPIX

CP

50 Hemin Rel. Int. [%] Int. Rel. 5P

6P 7P UP

0 0.00 1.00 2.00 3.00 4.00 Time [min]

Figure 13: Total ion chromatogram from the HPLC-MS/MS analysis of a standard mixture of porphyrins and hemin. UP: uroporphyrin; 7P: 7-carboxylporphyrin; 6P: 6-carboxylporphyrin; 5P: 5-carboxylporphyrin; CP: coproporphyrin; Hemin; MPIX: mesoporphyrin IX; PPIX: protoporphyrin IX. Flow rate: 1 mL/min (figure and caption from Paper III).

47 Another stationary phase that demonstrated better separation abilities than a regular C18, was the C18-PFP column. In this stationary phase, the octadecyl carbon chain is attached to the silica particles like a regular C18 column, but at the end a pentafluorophenyl (PFP) group is attached. This phase has several separation mechanisms; C18 hydrophobic separation mechanisms, ʌ-ʌ interactions with the aromatic ring, dipole-induced dipole and/or hydrogen bonding interaction with the electronegative fluorine atoms of the PFP- group. The porphyrin standard used in Paper I-IV contained small quantities of impurities of porphyrin isomers, in particularly for pentacarboxylic acid porphyrin. These isomers could be used to see how well the column separates naturally occurring porphyrin isomers from bacterial extracts. The

C18-PFP column exhibited the best isomer separation properties under the same chromatographic conditions, where three isomers were visible in the chromatogram. A 75 mm C18-PFP (75mm) was used for the HPLC porphyrin separation in Paper I and II, while in Paper III and IV a regular C18 column was used. Porphyrins were not well separated on HILIC columns, with extensive band broadening and co-elution as a result. Neither could the porphyrins be well separated on the Core shell C18 column, where the first four eluting compounds, uroporphyrin to pentacarboxylporphyrin, co-eluted with an extensive peak broadening and bad peak symmetry. Porphyrins are well separated on regular C18 stationary phases so this problem is most likely due to the core shell design of the particles in this type of column.

Mobile phases A commonly used mobile phase composition for porphyrin separation is the 1 M ammonium acetate-acetic acid buffer with a pH between 5.1 and 5.2, together with methanol/acetonitrile [14,25,109]. This mobile phase has

48 demonstrated good separation properties of different porphyrin isomers. Porphyrins are zwitterions, meaning that they may have both a positive and a negative charge depending on the pH. On reversed phase columns the ionization state of porphyrins will highly affect the chromatographic resolution, especially between isomers. At low pH, all carboxylic acids are fully protonated, but both the amine groups get protonated, leading to a positive net charge. At high pH the carboxylic acid groups will be deprotonated to a varying degree, leading to a negative net charge. A complete ionization of a porphyrin on a reversed phase column will result in poor retention, while a neutral porphyrin will have an excessive retention leading to long elution times. In the pH between 5.1 and 5.2 the optimal separation between porphyrin isomers can be achieved, especially for porphyrins with 4 to 8 carboxylic acid groups [22]. However, during experimental work for Paper I high signal suppression in the electrospray ion source could be observed using this buffer. It decreased the signal/noise- ratio (S/N) and increased the limit of detection (LOD) for all the analytes. HPLC-MS/MS analyses using a lower concentration of ammonium acetate (10 mM) did not reduce the ion suppression to acceptable levels. Bacteria contain porphyrins in low concentrations and sample sizes are often small. Generally a sample weight between 50 and 100 mg is possible to collect from five agar plates, therefore low LODs and LOQs are necessary rather than the possibility to separate different isomers. For these reasons the ammonium buffer was discarded as a mobile phase. Other mobile phases were evaluated such as ion pairing with triethylamine, acetonitrile and methanol as organic modifiers. Triethylamine resulted in faster elution but the chromatographic resolution was poor. No resolution between mesoporphyrin and protoporphyrin could be achieved. Acetonitrile was finally selected as the organic modifier due to higher

49 elution strength compared to methanol, which resulted in shorter run times (Paper I-IV).

Mass spectrometry of porphyrins The fragmentation of naturally occurring porphyrins mainly takes part at the carboxylic acid groups [22,110]. The macrocycle structure itself is very stable and does not fragment at reasonable collision energies [110]. The loss of m/z 59 Da corresponds to the benzylic cleavage of the propionic acid side • chain (loss of CH2COOH ), and the loss of 46 Da corresponds to the fragmentation of the acetic acid side chain (loss of HCOOH). The loss of 46 Da is specific to acetic acid substituents and can therefore be used for structural determinations [22]. Losses of 58 and 45 Da are results of the rearrangement of a proton between fragments and product ions [22]. Two other fragmentation patterns are often seen as well; the loss of water (m/z 18 Da) from the quasi-molecular ion and cleavage of methyl groups followed by rearrangement of a proton (loss of 14 Da) [22]. There is possible to see losses of 73 Da and the literature have contradictory explanations of the origin of these fragments. Fateen [111] explain those by loss of • CH2CH2COOH radical, while Lim [22] argues that these more likely are • • formed by losses of CH2COOH and CH3 followed by protonation of the resulting product ion. In Figure 14 the product ion spectrum is shown for pentacarboxylporphyrin I containing four propionic- and one acetic acid groups. These specific fragmentation patterns make porphyrins suitable to analyze with tandem mass spectrometry in selected reaction monitoring (SRM) mode. In general the LOD and LOQ are lower when using SRM, but the opposite was seen for most of the porphyrins. Calculating the S/N from

50 seven injections of a standard in both selected ion monitoring (SIM) and SRM mode showed that the average S/N was only higher in SRM for protoporphyrin IX and coproporphyrin (data not published). This is probably because of fewer carboxylic acid groups that can stabilize the molecule during fragmentation. Even though the S/N was lower in SRM this was selected as detection mode in Paper I-IV due to the higher selectivity.

51

Stability of porphyrins and heme in solution

One of the most crucial problems to overcome in porphyrin quantification is the tetrapyrrole’s tendency to aggregate in aqueous solutions [112]. Not accounting for this aggregation will most likely result in large quantification errors. There are several hypothesizes why porphyrin structures aggregate. The commonly accepted one is that porphyrins form aggregates by ʌ-ʌ stacking mechanisms [113]. The ʌ bonds in the porphyrin ring structure can be stacked in different ways and form large porphyrin complexes, stable enough to sustain a chromatographic run [114]. If using nonspecific detectors, such as UV-Vis and fluorescence, this will give multiple peaks when these dimers and higher aggregates have different retention times in the column. When porphyrins are analyzed with MS/MS, these aggregates are not visible as peaks (if not scanned for) but are manifested as a loss in peak area of the monitored porphyrin. Especially hydrophobic porphyrins with only two or less carboxylic acid groups such as heme and protoporphyrin have been shown to be extra sensitive to aggregation [38,112,114,115]. In this work a number of different experiments investigating the aggregation in different solvents have been performed (data not published). In an experiment heme and coproporphyrin were extracted from baker’s yeast and the content of porphyrins and heme in the extract was determined during 12 hours while stored in an autosampler tray protected from light. The peak area for coproporphyrin was stable during these 12 hours, while the peak area for heme drastically decreased. After 12 hours, 88% of the original peak area for heme was lost. In another experiment a porphyrin standard was dissolved in different solvents at equal concentration and the peak area from the HPLC-MS/MS analysis compared. All areas of the analytes were relative to areas given by a standard solved in 6 M formic

53 acid. Fairly concentrated hydrochloric acid (1.5 M) in general gave smaller peak areas compared to 6 M formic acid and more concentrated HCl (6 M) was even worse, where the response for uroporphyrin was below the limit of detection. When taking the same solvent and diluting it further with tris buffer, uroporphyrin could be detected once again. This demonstrates that HCl is not destructive to uroporphyrin; it only pushed the equilibrium towards aggregate formation. Once the solvent was changed, the equilibrium was shifted back in favor to uroporphyrin in free form. Lower concentration of formic acid gave lower relative peak areas due to be the porphyrins not being protonated to the same extent and therefore having reduced solubility in aqueous solvent. The solution with a concentration of 6 M formic acid was approximately 1:4 (v/v) formic acid:H2O, meaning that it has some characteristics of an organic solvent, which also increases the solubility, especially for the more hydrophobic meso- and protoporphyrin IX. In general, the area for more hydrophobic porphyrins increases with higher organic solvent composition, while it decreases for the more hydrophilic porphyrins. Due to tetrapyrrole’s tendency to aggregate in different solvents it is necessary to investigate if this artefact is present before doing any quantitative analysis. This is most easily done by preparing standards of the analytes in both organic and aqueous solvents, at an equal concentration, and to compare the peak areas. If the areas deviate from each other more than the instrumental precision, extra care should be addressed. If the method allows it, the use of ion pairing could be used to stabilize the porphyrins. A common ion pairing agent is tetrabutylammonium ion [114]. Another approach is to always keep the solvent composition in the sample the same as for standards. In this way the detector response factor will be the same for sample and standard and accurate quantification can be done. This latter

54

Part 3. Applications

Photo inactivation of bacteria

Pilot study Our research group has previously shown that single colonies of A. actinomycetemcomitans may emit red fluorescence when illuminated with blue light [116]. The intensity of this fluorescence increased during growth, but did not correlate to the size of the colony. This indicates that the bacterium accumulates fluorescent compounds probably due to the porphyrin content. During the work with Paper I intracellular porphyrins in the bacterium A. actinomycetemcomitans were identified. This, together with the support from the literature, led to cooperation with a Swedish company marketing toothbrushes with two incorporated 465 nm LEDs in the brush head. A set of brushes with 410 nm LEDs were manufactured in order to overlap the Soret band of porphyrins. A. actinomycetemcomitans was cultivated on blood agar plates and two toothbrushes were placed side by side and let to illuminate the center of an agar plate for 10 min in the morning and in the afternoon (total time of 20 min). Reference agar plates where no illumination was performed were treated the same way as the illuminated plates. In Figure 16 a reference plate and an illuminated plate are shown.

56 voltage power supply. By adjusting the voltage output the illumination energy reaching the bacterial cultures could be adjusted. Each column of LEDs in the matrix was used for illumination of bacteria in 10 mm cuvettes. In this way the illumination could be performed at relatively high cell number densities, serial diluted, and subsequently streaked on agar plates. Colony forming units (CFU) can be calculated on light treated bacteria and compared to reference samples where no illumination has been performed. In Paper IV E. coli was cultivated and illuminated. E. coli is a bacterium that has shown poor sensitivity for blue light, but the results from different studies vary [3,33]. In this study E. coli was illuminated with light doses ranging from 4.8–172.8 J/cm2 and no antimicrobial activity could be seen, Figure 17. Even a small increase in CFU during the illumination procedure was observed and E. coli was suspected to still grow during the illumination. To investigate if the porphyrin content could have an impact on the photo sensitivity of E. coli it was grown in broth supplemented with 5-ALA. As mentioned earlier, the addition of 5-ALA can to high concentrations of intracellular porphyrins if the bacteria have all necessary enzymes for the de novo synthesis of heme. E. coli sensitivity for blue light increased remarkably. By using a light dose of 4.8 J/cm2 the number of viable bacteria was reduced with more than 3 log10 steps (99.9% reduction). A light dose of 2 86.4 J/cm gave a reduction of viable bacteria > 6 log10 steps and thus the level to work as a disinfectant of E. coli was passed.

58 10

1 E. coli

0.1

0.01 E. Coli + ALA

0.001

Limit 0.0001 anitimicrobial activity Survival fraction 0.00001 Limit disinfectant 0.000001

0.0000001 0 30 60 90 120 150 180 Illumination light dose [J/cm2]

Figure 17: Comparison of the photosensitivity for E. coli grown in presence and absence of 5-ALA. A survival fraction of 1 corresponds to references where no illumination of bacterial culture has been performed. Values are means for three independent experiments and error bars represent the standard deviation. Red marks are single experiments using bacterial cultures with an OD600 > 0.1 (Figure and caption from Paper IV).

Porphyrin content in microorganisms When the porphyrin content was analyzed in three different microorganisms; S.cerevisiae, P. gingivalis and A. actinomycetemcomitans it was revealed that the porphyrin content differs between these microorganisms. Differences both in the total amount of porphyrins and their respective

59 concentrations were observed. The total amount of porphyrins was highest in A. actinomycetemcomitans, more than 4 times higher concentration than in S. cerevisiae. On the other hand, S. cerevisiae had the largest number of porphyrins detected and pentacarboxylic porphyrin was the only one of the naturally occurring porphyrins that not could be detected. In P. gingivalis only coproporphyrin and protoporphyrin IX were detected. As stated earlier P. gingivalis lacks crucial enzymes for the de novo synthesis of heme, among others all of the enzymes involved in the formation of 5-ALA is lacking so theoretically it should not be possible for P. gingivalis to synthesize any of the porphyrins. The findings were therefore not expected, but coproporphyrin and protoporphyrin IX have been detected in P. gingivalis previously [8,9], which is in agreement with our results. The presence of protoporphyrin IX could be due to P. gingivalis requiring heme to grow and has an uptake up heme from the growth medium. By enzymatic removal of iron from heme, protoporphyrin IX can be formed as been observed in Haemophilus influenzae [117]. Detection of coproporphyrin in P. gingivalis could not be explained by these mechanisms since there are currently no known enzymatic pathways that convert protoporphyrin IX to coproporphyrin. One possibility is that the growth medium contains coproporphyrin and P. gingivalis heme acquiring mechanisms also have an uptake of other such as coproporphyrin.

Influence of culture conditions Results from aPT experiments found in the literature differ greatly; one pathogen may in one study have an entirely different sensitivity for blue light than in another one. For example has E.coli been shown to possess low sensitivity for blue light in some studies [3,33] while in another E. coli could

60 be killed with blue light [30]. One explanation is that different light sources are used and they emit light of different wavelengths. Another observation was that the cultivation procedure for the same pathogen varied between studies. Since porphyrins are synthesized by the bacteria themselves, the way they are cultivated could affect the intracellular composition of porphyrins and hence their sensitivity to blue light. Three cultivation parameters were selected that often varies between studies and that could have a high impact on the biosynthesis of porphyrins; the cultivation time, passage number and the kind of cultivation media.

Time of culturing The two oral pathogens, A. actinomycetemcomitans and P. gingivalis were selected as model organisms for investigating the effect of culture time on the intracellular porphyrin profile in Paper II. A. actinomycetemcomitans is able to synthesize porphyrins endogenously while P. gingivalis is not. A. actinomycetemcomitans was grown and analyzed for porphyrin content at each day for 3 to 9 days, and P. gingivalis was grown and analyzed at 4, 7 and 11 days, Figure 18. P. gingivalis had almost the same porphyrin pattern during all the days of incubation; protoporphyrin IX being the major porphyrin species and coproporphyrin III at a relative low concentration level. For A. actinomycetemcomitans the pattern changed drastically during the incubation. At day 3 coproporphyrin III was the most abundant porphyrin making up 80% of the total porphyrin content (TPC), followed by protoporphyrin IX (16% TPC) and coproporphyrin I (4% TPC). At day 6 this pattern had reversed; the most abundant porphyrin was now protoporphyrin IX with 88% of TPC and coproporphyrin III was 9% of TPC.

61 actinomycetemcomitans was cultivated for 4 and 7 days with and without passage. At day 4 the porphyrin profile was more or less the same between passaged and non-passaged bacteria. However, at day 7 there was a large difference. Total porphyrin content increased 28 times upon passage. There was also a difference in the porphyrin profile, where except from coproporphyrin I, coproporphyrin III and protoporphyrin IX, also uroporphyrin and heptacarboxylic porphyrin could be detected in the passaged bacteria. This significant difference in TPC between passaged and non-passaged bacteria could have a high impact on the results from in vitro illumination experiments. In a real case scenario where bacteria in the oral cavity are illuminated, the bacteria will most likely have a high passage number. It is therefore recommended to use passaged bacteria in aPT experiments in vivo.

Cultivation medium There is a large variety of cultivation media. There are two main types, solid (agar) and liquid media (broth). Broth has traditionally been used in illumination experiments. This is due to the ease in performing dilution series and to calculate killing efficiency. Agar is a better choice when mimicking biofilms. P. gingivalis is an X-factor dependent bacterium, meaning that it requires heme to be able to grow. Heme can be added in pure chemical form, or by addition of blood to the agar. P. gingivalis were grown both on blood and blood free agar in Paper II and the bacteria was analyzed for porphyrin and heme content. The colonies were black pigmented and contained 5 times higher concentrations of heme when grown on blood compared to transparent when grown on blood free agar. Also the porphyrin profile was different. Coproporphyrin III could not be detected and the protoporphyrin

63 IX concentration was lower in bacteria grown on blood free medium. This further emphasizes the hypothesis that P. gingivalis has an uptake of coproporphyrin from the growth medium since blood contains coproporphyrin.

Influence of different additives on E. coli E. coli was used as a model when different additives related to heme biosynthesis were added to the growth medium. The primary focus was to analyze the bacterial extracts for heme content in order to see if these additives could change the heme concentration. The porphyrin content was also analyzed as they are intermediates in the heme biosynthesis.

5-aminolaevulinic acid 5-aminolaevulinic acid was added to the cultivation broth and after 48 h E. coli was analyzed for porphyrins and heme content, Figure 19 (Paper IV). The cultures became colored light purple after addition of 5-ALA. The chemical analyses revealed different porphyrin species in high concentrations compared to bacterial culture where no 5-ALA was added. In the control sample only coproporphyrin I and protoporphyrin IX could be detected, but with concentrations below LOQ. Addition of 5-ALA resulted in the production of uro-, hepta-, hexa-, penta-, copro-, and protoporphyrin IX. The total porphyrin content increased from 1.5 to 1882 nmol/g after administration of 5-ALA. The most abundant porphyrin was uroporphyrin followed by protoporphyrin IX. The broth, in which the bacteria were grown, was also analyzed and it contained up to 20 times higher concentrations of porphyrins than the bacterium itself. The only exception was protoporphyrin IX where only small amounts could be quantified in the

64 broth. This indicates that the bacterium excretes porphyrins when the intracellular concentration becomes too high. Extra caution should be taken if illumination experiments are carried out after administration of 5-ALA since the broth itself may produce singlet oxygen and have an antimicrobial effect. Bacteria should therefore be washed in new broth prior to illumination experiments. The reason why protoporphyrin IX not is excreted is unknown, but it is possible that if E. coli was allowed to grow for a longer period of time, it would start to excrete protoporphyrin IX as well.

65 Figure 19: Total ion chromatograms from the HPLC-MS/MS analysis of E. coli grown on LB broth (A) and after addition of 5-ALA (B). Chromatograms have been enlarged and normalized to CP I in the range 0.00-1.80 min. UP, uroporphyrin; 7P, 7-carboxylporphyrin; 6P I, 6-carboxylporphyrin I; 6P III*, 6-carboxylporphyrin III (tentatively identified), 5P, 5-carboxylporphyrin; CP I, coproporphyrin I; CP III, coproporphyrin III; PPIX, protoporphyrin IX (Figure and caption from Paper IV).

66 Hemin As mentioned earlier some bacteria are X-factor dependent and require heme to grow. These kinds of bacteria, such as P. gingivalis lack the enzymes for the de novo biosynthesis of heme. But how about bacteria that have all these necessary enzymes for heme biosynthesis? Can and will they acquire heme from their environment? E. coli was grown in broth spiked with hemin and after 24 h the bacteria were harvested and analyzed (Paper IV). The heme content in E. coli was found to increase 4.5 times in the treated culture compared to the control. E. coli is able to bind and transport heme through its cell walls. Whether E. coli synthesizes heme itself or takes it up from the environment is still unknown, but the higher concentrations of intra cellular heme are an indication that E. coli prefers to acquire heme from its environment when available in free form. The high affinity for heme could be a potential drug target for treatment of bacterial infections, utilizing structural analogs to heme.

Cobalt protoporphyrin IX Heme-acquiring systems in pathogens have been proposed to be candidates for a drug target [62]. Cobalt protoporphyrin IX (Co-PPIX) is a structural analog to heme where the iron (II) in heme has been replaced with cobalt (III). The macrocyclic ring structure and the side chain groups are identical. To investigate if E. coli could mistakenly recognize Co-PPIX as heme, the bacteria were analyzed after administration of Co-PPIX to the cultivation media in Paper IV. In one experiment Co-PPIX was added to the broth and E. coli was analyzed for heme content. If E. coli takes up Co-PPIX, the heme concentration should be lower than in a control where no Co-PPIX has been added. The result showed that heme concentration was reduced by 48%. In another experiment hemin and Co-PPIX were added in equimolar

67 concentrations. It was observed that the heme concentration once again was reduced by 52% compared to an experiment where only heme had been added to the broth. The lysate was additionally analyzed in fullscan and the m/z for [Co-PPIX+H]+ was extracted. It showed that only bacteria spiked with Co-PPIX contained this compound in high concentrations. These experiments show that Co-PPIX does mimic heme and that E. coli has an equal affinity for Co-PPIX as for heme. No antimicrobial effect could however be seen for Co-PPIX to E. coli cultures and the total cell number was equal after the administration. We believe that the capacity of E. coli to synthesize heme de novo is the reason that no antimicrobial effect was seen.

Singlet oxygen production of porphyrins 1 In Paper IV the O2 production for three naturally occurring porphyrins were determined by using a probe that reacts with oxidation products caused by reaction with singlet oxygen. Reaction with the probe will cause a shift in absorbance, and the loss of absorbance of the probe can be observed with UV-Vis. We chose one of the most commonly used protocol, which is highly 1 specific for O2 and also suitable for aqueous systems [46]. The probe used in this work was N,N-dimethyl-4-nitrosoaniline (RNO) and L-histidine (His) 1 1 was used for reaction with O2. L-histidine reacts with O2 and produces a trans- annular peroxide product [HisO2]. If RNO is present [HisO2] can react with RNO causing a loss in absorbance of RNO which can be monitored at 440 nm [118].

1 O2 + His Æ [HisO2](1)

[HisO2] + RNO Æ RNO2 + products (2)

68 The porphyrins were dissolved individually in 0.01 M solutions of histidine and then 50 μM of RNO was added. The illumination was performed in 10 mm path length cuvettes and analyzed with UV-Vis after illumination with varying light doses. It is important to perform experiments where no photosensitizer has been added to the solution in order to rule out that the probe is not sensitive for the light. The slope of the plot of bleached RNO vs the irradiation time is proportional to the rate of production of singlet oxygen produced by the photosensitizer [46]. Figure 20 shows the photo bleaching of RNO using protoporphyrin IX as a photosensitizer and a reference sample where no porphyrin has been added.

1,4 A RNO + PPIX B RNO 0 - 60 min 1,2 0 - 120 min

1,0

0,8

0,6 Absorbance 0,4

0,2

0,0 300 350 400 450 500 550 300 350 400 450 500 550 nm nm

Figure 20: UV-Vis spectra for the photo bleaching of RNO, with (A) and without (B) protoporphyrin IX as a photosensitizer. The cuvette was illuminated for a total time of 60 min in (A), and in (B) for 120 min (figure from Paper IV).

69 To determine the quantum yield of singlet oxygen (ĭǻ) generated by a porphyrin it is necessary to determine how many singlet oxygen molecules that are produced as a function of absorbed photons by the porphyrin molecule. Doing this in absolute terms is very complex. A more common, alternative way to perform these kinds of determinations is by using a compound with a well determined ĭǻ. By matching the absorbance (at a specific wavelength) of both the reference compound and the photosensitizer of interest, one can assure that the same numbers of photons are absorbed in each case. The ĭǻfor the porphyrin of interest can be calculated by the following equation:

݇ = ௎௡௞௡௢௪௡ ߔ௱௎௡௞௡௢௪௡ ߔ௱ௌ௧ௗ ൬ ൰ ݇ௌ௧ௗ

Where Unknown and Std denote the quantum yield of singlet oxygen produced by the investigated photosensitizer and the standard, respectively, and k is the slope of the loss in absorbance of RNO vs the irradiation time.

Determining the ĭǻ of photosensitizers using this method will only explain in what ratio an absorbed photon will generate one singlet oxygen molecule. It does not tell how well a photosensitizer absorbs photons, i.e. the compounds extinction coefficient. By taking the concentration needed for each photosensitizer to reach the same optical density in consideration, a relative ĭǻ of each photosensitizer may be determined. This was done in Paper IV for uroporphyrin, coproporphyrin and protoporphyrin IX and shown in Figure 21.

70 1 Uroporphyrin III

0.8 Coproporphyrin III

Protoporphyrin IX 0.6

/A)/[C] 0.4 0

ln (A 0.2

0 0 102030 Irradiation time [min]

Figure 21: Relative photo bleaching of RNO by 405 nm irradiation for three different porphyrins. A0 and A is the absorbance of RNO measured at 440 nm at time 0 and the specific irradiation time respectively (figure from Paper IV).

Uroporphyrin III was the porphyrin that was most efficient in producing singlet oxygen, followed by coproporphyrin III and protoporphyrin IX. It should be stressed that the ĭǻ of the different porphyrins is not the only parameter that determines if a porphyrin is efficient at killing bacteria. Depending on the different physical properties of porphyrins, they may concentrate at different subcellular locations in bacteria. The location of a photosensitizer is of highest importance since the lifetime of singlet oxygen is only 4 μs in aqueous solutions and in that time it will only diffuse about 100 nm [46].

Clinical study A clinical study was performed at Karolinska Institutet (Stockholm, Sweden). Forty eight first year students completed an 8-week study, where 71 toothbrushes with and without incorporated 450 nm blue LED were compared, Figure 22. The participants in the study were divided into two control groups and two intervention groups. Persons that were smokers had used antibiotics less than three month prior the study and/or having periodontitis were excluded from the study. Three clinical parameters were chosen to be monitored during the study: bleeding upon probing (BOP), Figure 22: Toothbrush, with the incorporated 450 nm LED turned gingival index (GI) and plaque index on, used in the clinical study. (PI). Several inflammation were also analyzed such as: Interleukin- ȕ ,/ ,/-6, IL -8, and tumor necrosis factor alpha in saliva and gingival crevicular fluid. After 8 weeks there were no statistical differences at level p=0.05 in BOP, GI and PI, between the group using the 450 nm blue light toothbrush and the group without blue light, Figure 23. The overall trend is that the clinical parameters for the group using blue light have decreased more from baseline than the control group. The decrease in PI was significantly different from the control group at a significance level of p=0.06. When inflammation markers were analyzed in saliva and/or gingival crevicular fluid, the concentrations were close to or below LOD for several markers. Markers that could be detected and quantified were IL-1B and IL-8 in saliva, and MMP-8, and TIMP-1in gingival crevicular fluid. There was no statistically significant difference in any of these markers between the control and intervention group (p•0.44). However, within the intervention group, inflammation markers were significant reduced after 8 weeks

72

Conclusions and future perspectives Porphyrins have been shown in a number of different studies to be a promising class of compounds to use for inactivation of bacteria, either by their photosensitizing abilities or by their biological importance. This field of science, however, suffers from insufficient information about the intracellular porphyrin content in bacteria in order to draw accurate conclusions. Most of the research in aPT carried out today is focused on in vitro illumination of bacteria and the results from the studies differ. This thesis presents new analytical methods for accurate determinations of porphyrins and heme, together with novel insights in the porphyrin content in bacteria and also the connection between porphyrin content and the bacterial sensitivity for blue light. A method for determination of porphyrins in oral bacteria has been developed and used to detect porphyrins in two oral pathogens, P. gingivalis and A. actinomycetemcomitans. It was shown that both these pathogens contained porphyrins in varying amounts and the porphyrin profile is highly dependent on the culture conditions, such as time of culturing, passage, and type of agar. At most the total porphyrin content increased 28 times, only by performing a single passage of the bacteria. For obvious reasons cultivation conditions could have an impact on aPT experiments if the photosensitizer concentration varies to such a high degree. This could explain why variations are observed in the results from aPT experiments and standardized cultivation procedures are required to be able to draw accurate conclusions from aPT experiments. However, there is a need for further investigations in how the porphyrin profile in different pathogens differs depending on the culture conditions. In the aPT field there is not yet any consensus if any porphyrin species is better as a photosensitizer than another. In fact, the porphyrin content has not 74 been able to be correlated to the photosensitivity for blue light. The relative quantum yield for three naturally occurring porphyrins were determined and it could be seen that there is quite a large difference in singlet oxygen production when using a 405 nm light source. By using 5-ALA a change in the porphyrin content of E. coli could be seen. E. coli is normally non- sensitive to blue light, but after administration to 5-ALA the photosensitivity increased considerably together with the porphyrin content. This shows that porphyrins clearly affect the photosensitivity of bacteria. However, due to the design of these experiments we were not able to draw any definite conclusions of the contribution of the individual porphyrins on the photosensitivity. It is not only the quantum yield of singlet oxygen that defines how well a porphyrin acts as a photosensitizer inside bacteria. The hydrophobicity will most likely determine where it is located. If a photosensitizer is positioned close to biologically important systems it can be an efficient photosensitizer even if it has a low quantum yield of singlet oxygen. Interesting experiments can be carried out in the future where 5- ALA is administered to bacteria at different periods of time. This will cause different porphyrin profiles and aPT experiments can be performed to determine the photosensitivity at these times. In this way it should be possible to alter the porphyrin content in the same microorganism and the photosensitivity could be correlated to a specific porphyrin. It is of course important to perform in vitro studies to understand the basic mechanisms of aPT experiments, but it is essential that aPT also work in vivo. In the clinical study we could see a weak trend that the 450 nm LED toothbrush possessed a phototherapeutic effect for three clinical indices. All indices were decreased in the intervention group, but there were no statistically significant difference compared to the control group. However, four inflammation markers were significantly decreased in the intervention

75 group while only one decreased significantly in the control group. In future studies a toothbrush with a LED emitting wavelengths close to 405 nm should be used for better targeting the porphyrin absorbance. It could also be possible to use high powered LEDs emitting light with higher intensity to achieve a phototherapeutic effect. Despite the biological importance of heme, there are no evaluated analytical methods for the determination of heme in microorganisms. We developed and evaluated an extraction and clean up method for heme using a fast and simple salting-out LLE, solving one of the most crucial problems with the determination of heme, aggregation. The method was developed for the simultaneous determination of porphyrins and all analytes were baseline separated within a runtime of 4 min using HPLC-MS/MS system. Addition of a heme lookalike compound had high affinity for the heme acquiring mechanisms of E. coli and decreased the intracellular content of heme. Since heme is the major iron source for pathogens during an infection in vertebrates, limiting the access for this compound could be promising for future antimicrobial treatments. If one is allowed to speculate freely; the heme acquiring mechanisms in pathogens may be used in future applications to transport antimicrobial compounds across the bacterial envelope. By using a compound that chemically looks like heme, “Trojan horse” compounds, it could be possible to exploit the heme acquisition pathways to selectively introduce antimicrobial compounds.

76 References

[1] World Health Organization. Antimicrobial resistance: global report on surveillance. : World Health Organization; 2014. [2] Laxminarayan R, Duse A, Wattal C, Zaidi AK, Wertheim HF, Sumpradit N, et al. Antibiotic resistance—the need for global solutions. The Lancet infectious diseases 2013;13(12):1057-1098. [3] Cieplik F, Späth A, Leibl C, Gollmer A, Regensburger J, Tabenski L, et al. Blue light kills Aggregatibacter actinomycetemcomitans due to its endogenous photosensitizers. Clin Oral Investig 2014;18(7):1763-1769. [4] De Sousa DL, Lima RA, Zanin IC, Klein MI, Janal MN, Duarte S. Effect of twice-daily blue light treatment on matrix-rich biofilm development. PloS one 2015;10(7):e0131941. [5] Hamblin MR, Viveiros J, Yang C, Ahmadi A, Ganz RA, Tolkoff MJ. Helicobacter pylori accumulates photoactive porphyrins and is killed by visible light. Antimicrob Agents Chemother 2005 Jul;49(7):2822-2827. [6] Battisti A, Morici P, Signore G, Ghetti F, Sgarbossa A. Compositional analysis of endogenous porphyrins from Helicobacter pylori. Biophys Chem 2017;229:25-30. [7] Kumar A, Ghate V, Kim M, Zhou W, Khoo GH, Yuk H. Kinetics of bacterial inactivation by 405nm and 520nm light emitting diodes and the role of endogenous coproporphyrin on bacterial susceptibility. Journal of and Photobiology B: 2015;149:37-44. [8] Henry CA, Judy M, Dyer B, Wagner M, Matthews JL. Sensitivity of Porphyromonas and Prevotella species in liquid media to argon laser. Photochem Photobiol 1995;61(4):410-413. [9] Soukos NS, Som S, Abernethy AD, Ruggiero K, Dunham J, Lee C, et al. Phototargeting oral black-pigmented bacteria. Antimicrob Agents Chemother 2005 Apr;49(4):1391-1396. [10] Grzybowski A, Pietrzak K. From patient to discoverer—Niels Ryberg Finsen (1860–1904)—the founder of phototherapy in dermatology. Clin Dermatol 2012;30(4):451-455. [11] Fleming A. On the antibacterial action of cultures of a penicillium, with special reference to their use in the isolation of B. influenzae. Br J Exp Pathol 1929;10(3):226. [12] Grabow WW. Method For Determination Of Singlet Oxygen Quantum Yields For New Fluorene-based Photosensitizers In Aqueous Media For The Advancement Of Photodynamic Therapy. 2004.

77 [13] Imada K, Tanaka S, Ibaraki Y, Yoshimura K, Ito S. Antifungal effect of 405- nm light on Botrytis cinerea. Lett Appl Microbiol 2014;59(6):670-676. [14] Ashkenazi H, Malik Z, Harth Y, Nitzan Y. Eradication of Propionibacterium acnes by its endogenic porphyrins after illumination with high intensity blue light. FEMS Immunology & Medical Microbiology 2003;35(1):17-24. [15] Maclean M, MacGregor SJ, Anderson JG, Woolsey G. Inactivation of bacterial pathogens following exposure to light from a 405-nanometer light- emitting diode array. Appl Environ Microbiol 2009 Apr;75(7):1932-1937. [16] Maclean M, MacGregor SJ, Anderson JG, Woolsey G. High-intensity narrow- spectrum light inactivation and wavelength sensitivity of Staphylococcus aureus. FEMS Microbiol Lett 2008;285(2):227-232. [17] Khurana D, Sharnamma B, Rathore PK, Tyagi P. Photodynamic Therapy-A Ray towards Periodontics. IOSR Journal of Dental and Medical Sciences 2014;13(3):64-71. [18] Garcez AS, Núnez SC, Azambuja Jr N, Fregnani ER, Rodriguez HM, Hamblin MR, et al. Effects of photodynamic therapy on Gram-positive and Gram-negative bacterial biofilms by bioluminescence imaging and scanning electron microscopic analysis. Photomedicine and laser surgery 2013;31(11):519-525. [19] Prasanth CS, Karunakaran SC, Paul AK, Kussovski V, Mantareva V, Ramaiah D, et al. Antimicrobial Photodynamic Efficiency of Novel Cationic Porphyrins towards Periodontal Gram-positive and Gram-negative Pathogenic Bacteria. Photochem Photobiol 2014;90(3):628-640. [20] Bu W, Myers N, McCarty JD, O’Neill T, Hollar S, Stetson PL, et al. Simultaneous determination of six urinary porphyrins using liquid chromatography–tandem mass spectrometry. Journal of Chromatography B 2003;783(2):411-423. [21] %LHVDJD03\U]\ĔVND.7URMDQRZLF]03RUSK\ULQVLQDQDO\WLFDOFKHPLVWU\ A review. Talanta 2000;51(2):209-224. [22] Lim CK. High-performance liquid chromatography and mass spectrometry of porphyrins, and bilins. : World Scientific; 2010. [23] Rimington C. Spectral-absorption coefficients of some porphyrins in the Soret-band region. Biochem J 1960 Jun;75(3):620-623. [24] Hindmarsh JT, Oliveras L, Greenway DC. Plasma porphyrins in the . Clin Chem 1999 Jul;45(7):1070-1076. [25] Božek P, Hutta M, Hrivnáková B. Rapid analysis of porphyrins at low ng/l DQGȝJOOHYHOVLQKXPDQXULQHE\DJUDGLHQWOLTXLGFKURPDWRJUDSK\Pethod using octadecylsilica monolithic columns. Journal of Chromatography A 2005;1084(1):24-32.

78 [26] Melanson JE, Lucy CA. Enhanced detection of porphyrins by capillary electrophoresis-laser induced fluorescence. Electrophoresis 2002;23(11):1689-1694. [27] Penner MH. Ultraviolet, visible, and fluorescence spectroscopy. Food analysis: Springer; 2010. p. 387-405. [28] Zhang Y, Zhu Y, Chen J, Wang Y, Sherwood ME, Murray CK, et al. Antimicrobial blue light inactivation of Candida albicans: In vitro and in vivo studies. Virulence 2016;7(5):536-545. [29] Wang Y, Wu X, Chen J, Amin R, Lu M, Bhayana B, et al. Antimicrobial blue light inactivation of gram-negative pathogens in biofilms: in vitro and in vivo studies. J Infect Dis 2016;213(9):1380-1387. [30] Murdoch LE, Maclean M, Endarko E, MacGregor SJ, Anderson JG. Bactericidal effects of 405 nm light exposure demonstrated by inactivation of Escherichia, Salmonella, Shigella, Listeria, and Mycobacterium species in liquid suspensions and on exposed surfaces. The Scientific World Journal 2012;2012. [31] Fukui M, Yoshioka M, Satomura K, Nakanishi H, Nagayama M. Specific- wavelength visible light irradiation inhibits bacterial growth of Porphyromonas gingivalis. J Periodont Res 2008;43(2):174-178. [32] Nitzan Y, Salmon-Divon M, Shporen E, Malik Z. ALA induced photodynamic effects on Gram positive and negative bacteria. Photochemical & Photobiological Sciences 2004;3(5):430-435. [33] Abana CM, Brannon JR, Ebbott RA, Dunigan TL, Guckes KR, Fuseini H, et al. Characterization of blue light irradiation effects on pathogenic and nonpathogenic Escherichia coli. MicrobiologyOpen 2017. [34] Ganz RA, Viveiros J, Ahmad A, Ahmadi A, Khalil A, Tolkoff MJ, et al. Helicobacter pylori in patients can be killed by visible light. Lasers Surg Med 2005;36(4):260-265. [35] Feuerstein O, Persman N, Weiss EI. Phototoxic Effect of Visible Light on Porphyromonas gingivalis and Fusobacterium nucleatum: An In Vitro Study¶†. Photochem Photobiol 2004;80(3):412-415. [36] Soukos NS, Stultz J, Abernethy AD, Goodson JM. Phototargeting human periodontal pathogens in vivo. Lasers in medical science 2015;30(3):943-952. [37] Fontana CR, Song X, Polymeri A, Goodson JM, Wang X, Soukos NS. The effect of blue light on periodontal biofilm growth in vitro. Lasers in medical science 2015;30(8):2077-2086. [38] Dietel W, Pottier R, Pfister W, Schleier P, Zinner K. 5-Aminolaevulinic acid (ALA) induced formation of different fluorescent porphyrins: a study of the biosynthesis of porphyrins by bacteria of the human digestive tract. Journal of Photochemistry and Photobiology B: Biology 2007;86(1):77-86. 79 [39] Heuck G. New strategies for the induction of endogenous porphyrins in eurcaryotic and procaryotic cells 2012. [40] Cieplik F, Tabenski LK, Buchalla W, Maisch T. Antimicrobial photodynamic therapy for inactivation of biofilms formed by oral key pathogens. Frontiers in microbiology 2014;2014(5):1-17. [41] Castano AP, Demidova TN, Hamblin MR. Mechanisms in photodynamic therapy: part one—photosensitizers, photochemistry and cellular localization. Photodiagnosis and photodynamic therapy 2004;1(4):279-293. [42] Strakhovskaya M, Shumarina A, Fraikin GY, Rubin A. Endogenous porphyrin accumulation and photosensitization in the yeast Saccharomyces cerevisiae in the prHVHQFHRIƍ-dipyridyl. Journal of Photochemistry and Photobiology B: Biology 1999;49(1):18-22. [43] Wakeman CA, Stauff DL, Zhang Y, Skaar EP. Differential activation of Staphylococcus aureus heme detoxification machinery by heme analogues. J Bacteriol 2014 Apr;196(7):1335-1342. [44] Romiti R, Schaller M, Jacob K, Plewig G. High-performance liquid chromatography analysis of porphyrins in Propionibacterium acnes. Arch Dermatol Res 2000;292(6):320-322. [45] Wainwright M. Photodynamic antimicrobial chemotherapy (PACT). J Antimicrob Chemother 1998 Jul;42(1):13-28. [46] Kochevar IE, Redmond RW. [2] Photosensitized production of singlet oxygen. Meth Enzymol 2000;319:20-28. [47] Wu H, Song Q, Ran G, Lu X, Xu B. Recent developments in the detection of singlet oxygen with molecular spectroscopic methods. TrAC Trends in Analytical Chemistry 2011;30(1):133-141. [48] Baier J, Fuß T, Pöllmann C, Wiesmann C, Pindl K, Engl R, et al. Theoretical and experimental analysis of the luminescence signal of singlet oxygen for different photosensitizers. Journal of Photochemistry and Photobiology B: Biology 2007;87(3):163-173. [49] Lindig BA, Rodgers MA, Schaap AP. Determination of the lifetime of singlet oxygen in water-d2 using 9, 10-anthracenedipropionic acid, a water-soluble probe. J Am Chem Soc 1980;102(17):5590-5593. [50] Bolton JR, Mayor-Smith I, Linden KG. Rethinking the Concepts of Fluence (UV Dose) and Fluence Rate: The Importance of Photon-based Units–A Systemic Review. Photochem Photobiol 2015;91(6):1252-1262. [51] Braslavsky SE. Glossary of terms used in photochemistry, (IUPAC Recommendations 2006). Pure and Applied Chemistry 2007;79(3):293-465. [52] Heinemann IU, Jahn M, Jahn D. The of heme biosynthesis. Arch Biochem Biophys 2008;474(2):238-251.

80 [53] Beale SI, Castelfranco PA. 14C incorporation from exogenous compounds LQWRį-aminolevulinic acid by greening cucumber cotyledons. Biochem Biophys Res Commun 1973;52(1):143-149. [54] Shemin D, Rittenberg D. The utilization of glycine for the synthesis of a porphyrin. J Biol Chem 1945;159(2):567-568. [55] Kennedy J, Pottier R, Pross D. Photodynamic therapy with endogenous protoporphyrin: IX: basic principles and present clinical experience. Journal of Photochemistry and Photobiology B: Biology 1990;6(1-2):143-148. [56] Hamblin MR, Hasan T. Photodynamic therapy: a new antimicrobial approach to infectious disease? Photochemical & Photobiological Sciences 2004;3(5):436-450. [57] Dai T, Gupta A, Murray CK, Vrahas MS, Tegos GP, Hamblin MR. Blue light for infectious diseases: Propionibacterium acnes, Helicobacter pylori, and beyond? Drug Resistance Updates 2012;15(4):223-236. [58] Cassat JE, Skaar EP. Iron in infection and immunity. Cell host & microbe 2013;13(5):509-519. [59] Nairz M, Haschka D, Demetz E, Weiss G. Iron at the interface of immunity and infection. Front Pharmacol 2014 Jul 16;5:152. [60] Skaar EP. The battle for iron between bacterial pathogens and their vertebrate hosts. PLoS pathogens 2010;6(8). [61] Braun V, Hantke K, Koester W. Bacterial iron transport: mechanisms, genetics, and regulation. Met Ions Biol Syst 1998;35:67-146. [62] McLean KJ, Munro AW. Drug targeting of heme proteins in Mycobacterium tuberculosis. Drug Discov Today 2017;22(3):566-575. [63] Owens CP, Chim N, Goulding CW. Insights on how the Mycobacterium tuberculosis heme uptake pathway can be used as a drug target. Future medicinal chemistry 2013;5(12):1391-1403. [64] Lobo SA, Scott A, Videira MA, Winpenny D, Gardner M, Palmer MJ, et al. Staphylococcus aureus haem biosynthesis: characterisation of the enzymes involved in final steps of the pathway. Mol Microbiol 2015;97(3):472-487. [65] Mike LA, Dutter BF, Stauff DL, Moore JL, Vitko NP, Aranmolate O, et al. Activation of heme biosynthesis by a small molecule that is toxic to fermenting Staphylococcus aureus. Proc Natl Acad Sci U S A 2013 May 14;110(20):8206-8211. [66] Dutter BF, Mike LA, Reid PR, Chong KM, Ramos-Hunter SJ, Skaar EP, et al. Decoupling Activation of Heme Biosynthesis from Anaerobic Toxicity in a Molecule Active in Staphylococcus aureus. ACS Chem Biol 2016 May 20;11(5):1354-1361.

81 [67] Choby JE, Skaar EP. Heme synthesis and acquisition in bacterial pathogens. J Mol Biol 2016;428(17):3408-3428. [68] Olczak T, Maszczak-Seneczko D, Smalley JW, Olczak M. Gallium (III), cobalt (III) and copper (II) protoporphyrin IX exhibit antimicrobial activity against Porphyromonas gingivalis by reducing planktonic and biofilm growth and invasion of host epithelial cells. Arch Microbiol 2012;194(8):719-724. [69] Henderson B, Ward JM, Ready D. Aggregatibacter (Actinobacillus) actinomycetemcomitans: a triple A* periodontopathogen? Periodontol 2000 2010;54(1):78-105. [70] Norskov-Lauritsen N, Kilian M. Reclassification of Actinobacillus actinomycetemcomitans, Haemophilus aphrophilus, Haemophilus paraphrophilus and Haemophilus segnis as Aggregatibacter actinomycetemcomitans gen. nov., comb. nov., Aggregatibacter aphrophilus comb. nov. and Aggregatibacter segnis comb. nov., and emended description of Aggregatibacter aphrophilus to include V factor-dependent and V factor- independent isolates. Int J Syst Evol Microbiol 2006 Sep;56(Pt 9):2135-2146. [71] Boyce JM, Frazur J, Zinnemann K. The growth requirements of Haemophillus aphrophillus. J Med Microbiol 1969 Feb;2(1):55-62. [72] Nagata H, Ikawa Y, Kuboniwa M, Maeda K, Shizukuishi S. Characterization of Hemoglobin Binding to< i> Actinobacillus actinomycetemcomitans. Anaerobe 2002;8(3):109-114. [73] Grenier D, Leduc A, Mayrand D. Interaction between Actinobacillus actinomycetemcomitans lipopolysaccharides and human hemoglobin. FEMS Microbiol Lett 1997;151(1):77-81. [74] Hayashida H, Poulsen K, Kilian M. Differences in iron acquisition from human haemoglobin among strains of Actinobacillus actinomycetemcomitans. Microbiology 2002 Dec;148(Pt 12):3993-4001. [75] Graber KR, Smoot LM, Actis LA. Expression of iron binding proteins and hemin binding activity in the dental pathogen Actinobacillus actinomycetemcomitans. FEMS Microbiol Lett 1998;163(2):135-142. [76] AbdulAzeez AR, Mahmood MS, Ali WM. Phototoxic Effect of Visible Blue Light on Aggregatibacter Actinomycetemcomitans and Porphyromonas Gingivalis in Patients with Chronic Periodontitis: An in-Vitro Study. Journal of Baghdad College of Dentistry 2015;27(1):144-150. [77] Genco CA, Odusanya BM, Brown G. Binding and accumulation of hemin in Porphyromonas gingivalis are induced by hemin. Infect Immun 1994 Jul;62(7):2885-2892. [78] Bramanti TE, Holt SC. Roles of porphyrins and host iron transport proteins in regulation of growth of Porphyromonas gingivalis W50. J Bacteriol 1991 Nov;173(22):7330-7339.

82 [79] Roper JM, Raux E, Brindley AA, Schubert HL, Gharbia SE, Shah HN, et al. The enigma of cobalamin ( B12) biosynthesis in Porphyromonas gingivalis. Identification and characterization of a functional pathway. J Biol Chem 2000 Dec 22;275(51):40316-40323. [80] Smalley JW, Silver J, Marsh PJ, Birss AJ. The periodontopathogen Porphyromonas gingivalis binds iron protoporphyrin IX in the mu-oxo dimeric form: an oxidative buffer and possible pathogenic mechanism. Biochem J 1998 May 1;331 ( Pt 3)(Pt 3):681-685. [81] Whiteaker JR, Fenselau CC, Fetterolf D, Steele D, Wilson D. Quantitative determination of heme for forensic characterization of Bacillus spores using matrix-assisted laser desorption/ionization time-of-flight mass spectrometry. Anal Chem 2004;76(10):2836-2841. [82] Kwon SJ, de Boer AL, Petri R, Schmidt-Dannert C. High-level production of porphyrins in metabolically engineered Escherichia coli: systematic extension of a pathway assembled from overexpressed genes involved in heme biosynthesis. Appl Environ Microbiol 2003 Aug;69(8):4875-4883. [83] 1LW]DQ<.DXIIPDQ0(QGRJHQRXVSRUSK\ULQSURGXFWLRQLQEDFWHULDE\į- aminolaevulinic acid and subsequent bacterial photoeradication. Lasers in Medical Science 1999;14(4):269-277. [84] Olsen I, Brondz I. Differentiation among closely related organisms of the Actinobacillus-Haemophilus-Pasteurella group by means of lysozyme and EDTA. J Clin Microbiol 1985 Oct;22(4):629-636. [85] Voss J. Effects of organic cations on the gram-negative cell wall and their bactericidal activity with ethylenediaminetetra-acetate and surface active agents. Microbiology 1967;48(3):391-400. [86] Goldschmidt MC, Wyss O. The role of tris in EDTA toxicity and lysozyme lysis. Microbiology 1967;47(3):421-431. [87] Repaske R. Lysis of gram-negative organisms and the role of versene. Biochim Biophys Acta 1958;30(2):225-232. [88] Ishida T, Yu L, Akutsu H, Ozawa K, Kawanishi S, Seto A, et al. A primitive pathway of porphyrin biosynthesis and enzymology in Desulfovibrio vulgaris. Proc Natl Acad Sci U S A 1998 Apr 28;95(9):4853-4858. [89] Letoffe S, Heuck G, Delepelaire P, Lange N, Wandersman C. Bacteria capture iron from heme by keeping tetrapyrrol skeleton intact. Proc Natl Acad Sci U S A 2009 Jul 14;106(28):11719-11724. [90] Kjeldstad B, Johnsson A, Sandberg S. Influence of pH on porphyrin production in Propionibacterium acnes. Arch Dermatol Res 1984;276(6):396- 400.

83 [91] Cai Y, Jiang G, Liu J, Zhou Q. Multiwalled carbon nanotubes as a solid-phase extraction adsorbent for the determination of bisphenol A, 4-n-nonylphenol, and 4-tert-octylphenol. Anal Chem 2003;75(10):2517-2521. [92] Mayfield JA, Dehner CA, DuBois JL. Recent advances in bacterial heme protein biochemistry. Curr Opin Chem Biol 2011;15(2):260-266. [93] Lundstedt T, Seifert E, Abramo L, Thelin B, Nyström Å, Pettersen J, et al. Experimental design and optimization. Chemometrics Intellig Lab Syst 1998;42(1):3-40. [94] Espinas NA, Kobayashi K, Takahashi S, Mochizuki N, Masuda T. Evaluation of unbound free heme in cells by differential acetone extraction. Plant Cell Physiol 2012 Jul;53(7):1344-1354. [95] Stillman LC, Gassman ML. Protoheme extraction from plant . Anal Biochem 1978;91(1):166-172. [96] Schneegurt MA, Beale SI. Biosynthesis of protoheme and from glutamate in maize. Plant Physiol 1986 Aug;81(4):965-971. [97] Thomas J, Weinstein JD. Measurement of heme efflux and heme content in isolated developing . Plant Physiol 1990 Nov;94(3):1414-1423. [98] Polson C, Sarkar P, Incledon B, Raguvaran V, Grant R. Optimization of protein precipitation based upon effectiveness of protein removal and ionization effect in liquid chromatography–tandem mass spectrometry. Journal of Chromatography B 2003;785(2):263-275. [99] Saleh A. Analysis of drugs and proteins in biological matrices, using different types of sample preparations and mass spectrometry. 2014. [100] Duong-Ly KC, Gabelli SB. Chapter Seven-Salting out of Proteins Using Ammonium Sulfate Precipitation. Meth Enzymol 2014;541:85-94. [101] Stoscheck CM. [6] Quantitation of protein. Meth Enzymol 1990;182:50-68. [102] Aitken A, Learmonth MP. Protein determination by UV absorption. The protein protocols handbook 2009:3-6. [103] Bradford MM. A rapid and sensitive method for the quantitation of microgram quantities of protein utilizing the principle of protein-dye binding. Anal Biochem 1976;72(1-2):248-254. [104] Sedmak JJ, Grossberg SE. A rapid, sensitive, and versatile assay for protein using Coomassie brilliant blue G250. Anal Biochem 1977;79(1-2):544-552. [105] Klesper E, Corwin A, Turner D. High pressure gas chromatography above critical temperatures. J Org Chem 1962;27(2):700-&. [106] Karayannis NM, Corwin AH. Hyperpressure gas chromatography, III: Gas chromatography of porphyrins and metalloporphyrins. Anal Biochem 1968;26(1):34-50.

84 [107] Marriott P, Gill J, Evershed R, Eglinton G, Maxwell J. Capillary GC and GC- MS of porphyrins and metalloporphyrins. Chromatographia 1982;16(1):304- 308. [108] Yang M, Fazio S, Munch D, Drumm P. Impact of methanol and acetonitrile on separations based on p–p interactions with a reversed-phase phenyl column. Journal of Chromatography A 2005;1097(1):124-129. [109] Lim C, Rideout J, Peters T. High-performance liquid chromatography of dicarboxylic porphyrins and metalloporphyrins: retention behaviour and biomedical applications. Journal of Chromatography A 1984;317:333-341. [110] Benton CM, Lim CK, Moniz C, Jones DJ. Travelling wave ion mobility mass spectrometry of 5-aminolaevulinic acid, porphobilinogen and porphyrins. Rapid Communications in Mass Spectrometry 2012;26(4):480-486. [111] Fateen E, Abd-Elfattah A, Gouda A, Ragab L, Nazim W. Porphyrins profile by high performance liquid chromatography/electrospray ionization tandem mass spectrometry for the diagnosis of . Egyptian Journal of Medical Human Genetics 2011;12(1):49-58. [112] Brown SB, Shillcock M, Jones P. Equilibrium and kinetic studies of the aggregation of porphyrins in aqueous solution. Biochem J 1976;153:279-285. [113] Hunter CA, Sanders JK. The nature of. pi.-. pi. interactions. J Am Chem Soc 1990;112(14):5525-5534. [114] Bommer JC, Hambright P. General laboratory methods for tetrapyrroles. Heme, Chlorophyll, and Bilins: Methods and Protocols 2002:39-67. [115] Gledhill M. The determination of in marine phyto-and bacterioplankton. Mar Chem 2007;103(3):393-403. [116] Bjurshammar N, Johannsen A, Buhlin K, Tranæus S, Östman C. On the red fluorescence emission of Aggregatibacter actinomycetemcomitans. Open Journal of Stomatology 2012;2(4). [117] Loeb MR. Ferrochelatase activity and protoporphyrin IX utilization in Haemophilus influenzae. J Bacteriol 1995 Jun;177(12):3613-3615. [118] .UDOMLü,0RKVQL6($QHZPHWKRGIRUWKHGHWHFWLRQRIVLQJOHWR[\JHQLQ aqueous solutions. Photochem Photobiol 1978;28(4-5):577-581.

85

HapPhy Days? A cohort study of the happiness during PhD studies

Jonas Fyrestam*

A B S T R A C T

Objectives: This work aims to test the hypothesis that PhD students are more at risk of large fluctuations in the normalized happiness index.

Methods: A retrospective cohort study of one male PhD student born in 1981 was con- ducted. The normalized happiness index of the PhD student was monitored during five years and compared to a control group where none of the objects had been exposed to PhD studies.

Results: It could be seen that the normalized happiness index fluctuated significant (p=0.0001) for the PhD student when compared to the control group. However, area un- der a curve was well above the area over a curve.

Conclusions: These data suggest that suffering a PhD program will most likely cause high fluctuations in normalized happiness index. Mental stress during publication pro- cedures, failed experiments and thesis writing may well explain low normalized happi- ness index values while awesome colleagues and a magnificent working environment correlates well with high ones.

1. Introduction were enrolled. In order to limit the num- ber of confounders only Swedish PhD stu- It is commonly accepted that PhD stu- dents born in May 1981 were selected. To dents from the unit of analytical chemistry maintain the full integrity of the PhD stu- (ACESk) at Stockholm University are ex- dent, he was anonymized with the code: traordinary. However, rarely is the happi- JF_81. As a control group all other males ness during the PhD studies mentioned in Sweden, not finished a PhD, were se- after the successful dissertation proce- lected (n= 5 043 523). dure. Normalized happiness index (NHI) is a novel tool to estimate the happiness 2.2 Study design over time and was used in the present study. JF_81 was first assigned a main supervi- sor with excellent skills in chromatog- 2. Methods raphy and analytical chemistry in gen- This was a retrospective cohort study eral1. With a free hand and encouraging with no ethical approval at all. words he let JF_81 evolve as a researcher and as a person, something that he will 2.1 Participants forever be thankful. Subsequently, an assistant supervisor All males (n=1) working at ACESk, during was allotted to JF_812. Fruitful discus- 2012-2017, whose body weight is most eas- ily measured in metric tons (0.1 ± 0.01) 1 C. Östman 2 G. Thorsén

86 sions about research and teaching took ty > 99.9% was used the first half of the place and JF_81 was always amazed by PhD studies). his broad knowledge in so many fields of science. 2.4 Normalized happiness index JF_81 was placed in an office together JF_81 was instructed to rank every work- with several other PhD students3. Due to ing day with a number between -100 to them the working days for JF_81 became +100 where -100 correspond to low happi- a real joy and he is grateful for their ness (sadness) and +100 correspond to friendship and support. Additionally, all high happiness. By doing super advanced past and present PhD students and staff statistical treatment these numbers were at ACESk made the working days so much converted into NHI values ranging from -1 easier to go through. to +1. JF_81 was also instructed to take Between the years 2012-2017 numer- notes on specific events that happened ous questions and problems not related to during the PhD studies. The control research turned up for JF_81. Thanks to groups NHI values for each day were nor- the administrative and technical staff4 malized to 0. these problems were easily solved. These 5 persons, together with some others , made 3. Results and discussion the nutrition intake at 11:00 a.m. one of the highlights of the day. In Figure 1 the NHI values during JF_81 Regarding the research JF_81 had the PhD study is plotted together with some pleasure to work with several highly specific events that took place. Area under skilled scientists and diploma workers6. a curve (AUC) is highlighted in green, Without them the scientific work would while area over a curve (AOC) is high- have been much more challenging. lighted in red. AUC is 63% and AOC is At the end of 2017, JF_81 had to 37%, so most of the time JF_81 was happy summarize the past five years by writing compared to the control group. However, it the thesis. Due to the kindness and inval- is clearly visible that there are high fluc- uable comments from several people the tuations in NHI values. Most of the time manuscript was highly improved7. JF_81 is very happy or very sad (NHI val- ues > 0.3 and < -0.3 respectively), rarely in 2.3 Chemicals between. These peaks and valleys corre- spond well with specific events that took Ethanol (purity between 2.1 to 95%) was place during the PhD study. used occasionally. Nicotine and caffeine was used on a daily basis and obtained 4. Conclusion from various manufacturers. Garlic (puri- JF_81 wants to thank everybody at ACESk for making this a fantastic time!

3 F. Mashayekhy, F. Idaresta, R. Avagyan and A. Acknowledgement Saleh 4 L. Elfver and J. Rutberg The Swedish research council and Elin 5 R. Westerholm, U. Nilsson, C. Östman and K. Paulsson are acknowledged for financial Lidén 6 N. Bjurshammar, E. Paulsson, N. Mansouri and J. support. Carlsson. 7 C. Östman, U. Nilsson, R. Westerholm, C. de Wit, L. Ilag, G. Thorsén, J. Aasa, F. Idaresta, R. Avagyan, E. Paulsson and H. Lim 87