Regulation of nitrogen fixation in Rhodospirillum rubrum Through proteomics and beyond

Tiago Toscano Selão

©Tiago Toscano Selão, Stockholm 2010

ISBN 978-91-7447-125-0, pp 1 – 71

Printed in Sweden by US-AB, Stockholm 2010

Distributor: Department of Biochemistry and Biophysics Cover image: Three-dimensional rendering of the Gaussian peaks for a 2D-PAGE gel, generated

using PDQuest 7.3.0.

Dedicado aos meus pais e

à memória dos meus avós.

Publication list

The work presented on this thesis is based on the following publications, referred to in the text by the corresponding Roman numerals: I – Selão, T. T., Nordlund, S. and Norén, A. ”Comparative proteomic studies in Rhodospirillum rubrum grown under different nitrogen conditions”, J Prot Res, 2008, 7: p. 3267-75

II – Teixeira, P. F.*, Selão, T. T.*, Henriksson, V., Wang, H., Norén, A. and Nordlund, S. ”Diazotrophic growth of Rhodospirillum rubrum with 2-oxoglutarate as sole carbon source affects the regulation of nitrogen metabolism as well as the soluble proteome”, Res. Microb., 2010, in press

III – Selão, T. T., Branca, R., Lehtiö, J., Chae, P. S., Gellman, S. H., Rasmussen, S., Nordlund, S. and Norén, A. ”Identification of the chromatophore membrane complexes formed under different nitrogen conditions in Rhodospirillum rubrum”, submitted

IV – Selão, T. T.*, Teixeira, P. F.* and Nordlund, S. ”The activity of dinitrogenase reductase ADP-ribosyltransferase of Rhodospirillum rubrum is controlled through reversible complex formation with the PII protein GlnB”, manuscript

V – Selão, T. T., Edgren, T., Wang, H., Norén, A. and Nordlund, S. ”The effect of pyruvate on the metabolic regulation of nitrogenase activity in Rhodospirillum rubrum with darkness as switch-off effector”, submitted * Authors contributed equally

Additional publications Vintila, S., Selão, T. T., Norén, A., Bergman, B., El-Shehawy, R., “Characterization of nifH gene expression, modification and rearrangement in Nodularia spumigena strain AV1”, submitted

Contents

I. Introduction....................................................................... 9

1. Bacteria, the largest population of Earth ...................................... 9 2. One of the elements of life - nitrogen........................................... 9 3. Nitrogen fixation and assimilation ............................................. 11

3.1. The diazotrophs .......................................................................................11 3.2. Nitrogenase and nitrogen fixation............................................................13 3.3. Nitrogen assimilation...............................................................................17 3.4. Regulation of nitrogen fixation and assimilation .....................................19

4. An overview of the field of proteomics...................................... 34 4.1. Gel-based proteomics ..............................................................................34 4.2. Gel-free proteomics .................................................................................37 4.3. Mass spectrometry ...................................................................................37

II. Adaptation to different nitrogen conditions .................. 40

1. The interconnection between nitrogen fixation and carbon fixation pathways ........................................................................... 40 2. Keeping the (redox) balance....................................................... 42

III. Post-translational regulation of nitrogenase activity – guarding the guards ............................................................ 44

1. Regulation of ammonium switch-off ......................................... 44 2. Regulation of energy switch-off................................................. 46

IV. Future prospects ........................................................... 50

V. Concluding remarks....................................................... 51

VI. Sumário em português ................................................. 53

VII. Acknowledgements..................................................... 55

VIII. References ................................................................. 58

Abbreviations AAA+ ATPase associated with diverse cellular activities Arc Aerobic respiratory control FNR Fumarate and nitrate reduction regulator GAF cGMP phosphodyesterase; adenylyl cyclase; FhlA Km Michaelis-Menten constant SDS-PAGE Sodium dodecylsulphate – polyacrylamide gel

electrophoresis σ Sigma subunit of the RNA polymerase

8

“Cantando espalharei por toda parte, Se a tanto me ajudar o engenho e arte. [...]

E vós, Tágides minhas, pois criado

Tendes em mi um novo engenho ardente Se sempre, em verso, humilde, celebrando

Foi de mi vosso rio alegremente, Dai-me agora um som alto e sublimado,

Um estilo grandíloco e corrente Por que de vossas águas Febo ordene

Que não tenham enveja às de Hipocrene.

Dai-me hũa fúria grande e sonorosa, E não de agreste avena ou frauta ruda,

Mas de tuba canora e belicosa, Que o peito acende e a cor ao gesto muda.”

Luís Vaz de Camões, in “Os Lusíadas”, Canto I

9

I. Introduction

1. Bacteria, the largest population of Earth

The multitude of forms of life present all around us is estimated to have arisen from a single common ancestor over 3,5 billion years ago 1. Of all the different organisms that science has now described, bacteria are undoubtedly the most widely spread and adaptable of all. They have been found in all different kinds of habitats on Earth, from deep within the planet’s crust to the surface of the seas, living within mammalian intestines or in acid mine drainage. Deinococcus

radiodurans is perhaps the best example of the extreme hardiness of bacteria, being able to survive exposure to cold, large pH variations, dehydration and, for what it is most known, high doses of ionizing radiation 2. The extreme adaptability of bacteria has allowed them to become ubiquitous on our planet, to the extent that they are estimated to be the largest contributors to the world’s biomass 3. It was even recently suggested that complex animals, including humans, should actually be seen as “super-organisms”, in view of the fact that they host about ten times as many bacterial cells as their own and that these bacteria actually interact and affect the animals’ own metabolism 4.

2. One of the elements of life - nitrogen

In 1772 Daniel Rutherford, a Scottish chemist and physician, described the presence of a gas or “air”, as it was then called, in the atmosphere. By removing what Joseph Black had named in 1754 as “fixed air” (carbon dioxide) from the atmosphere, generated by burning coal, he isolated what he entitled “noxious air”. It was later named “azote” (“lifeless”) by Antoine Lavoisier 5 as animals would suffocate in it and even the flame of a candle would quickly die. The

10

work of Lavoisier and, almost simultaneously of Joseph Priestley and Carl Wilhelm Scheele made clear that “azote”, or nitrogen, was indeed an elementary constituent of not only air but also of many other substances. The name “nitrogen”, most commonly used in the English language, derives from the Greek word “νιτρον” (“nitron”), nitre or saltpetre (potassium nitrate), a compound used in the manufacture of gunpowder and soap and also used as a fertilizer. The root of the word denotes the fact that even though it was not until the 18th century that nitrogen was recognized as an element, manufactured substances containing it had been in use for centuries. Nitrogen is one of the most common elements in biomolecules and can very frequently become a growth limiting factor for living beings, due to the fact that it is not always available in a biologically usable form. The triple bond in molecular nitrogen (N2) is the strongest among naturally occurring substances (bond dissociation enthalpy of 946 kJ.mol-1) 6, giving rise to the unreactiveness of atmospheric nitrogen. This high energy threshold makes atmospheric dinitrogen biologically unavailable to most organisms. However, a group of prokaryotes produces an enzyme complex, nitrogenase, which catalyzes reduction of atmospheric dinitrogen to ammonium, at room temperature and pressure, thus replenishing the biological nitrogen pool (Figure 1). This amazing feat of evolution becomes even more astonishing if we bear in mind that ammonium production at an industrial scale by the Haber-Bosch process requires a very high energy input, with temperatures in excess of 300 ºC, pressures of 15-20 MPa and specific catalysts.

Figure 1 – The global nitrogen cycle

11

3. Nitrogen fixation and assimilation

3.1. The diazotrophs

Diazotrophic bacteria are responsible for an important step in the global nitrogen cycle, the reduction of atmospheric dinitrogen to ammonium. The capacity to perform this reaction is exclusive of prokaryotic organisms. Several symbiotic relations between plants and diazotrophs are also well established, thus allowing eukaryotic organisms to directly harness the power of nitrogen reduction. Symbiotic diazotrophs interact with plants using three different strategies. Endosymbiotic diazotrophs are perhaps the most widely known type of symbiotic nitrogen fixing bacteria, infecting the roots of plants and establishing an elaborate type of intracellular symbiosis, in which extensive signalling pathways exist between host and symbiont. A well-studied example is the relation between Rhizobiaceae and leguminous plants, such as pea or soybean, with the infection by these bacteria inducing formation of root nodules in the plant and resulting in morphological and metabolic changes in the bacteria themselves, ultimately leading to their differentiation into nitrogen fixing “organelles” named "bacteroids" 7. These symbiotic organisms are the basis for the agricultural practice of crop rotation, used by farmers around the world for generations to increase yields. Other symbiotic diazotrophs, termed endophytic, establish themselves within the vascular tissue of the host plants, as is the case for Gluconoacetobacter diazotrophicus and Herbaspirillum

seropedicae and their host, sugar cane 8, 9 or Azoarcus sp. and its symbiosis with rice and Kallar grass 10. Finally, some of these bacteria can interact in a rather non-invasive and to some extent less elaborate form with their hosts, by colonizing the surface or specific cavities of plants. These associative symbionts are exemplified by the relations established between Azospirillum spp. and different kinds of grasses 11. On the other hand, most nitrogen fixing bacteria can be found in a free-living form. Already in 1893 Winogradsky demonstrated that Clostridium pasteurianum could use atmospheric dinitrogen 12 and, since then, many other organisms were shown to be able to perform nitrogen fixation. Some of the most widely studied examples include the γ-proteobacteria Klebsiella pneumoniae and Azotobacter

12

vinelandii, with the first crystal structure of the nitrogenase complex obtained using the proteins from the latter organism 13. Free-living diazotrophic organisms are found in a variety of environments and are usually classified as aerobic, anaerobic or photosynthetic. This last group is of special interest, not only for the connection between these two very important processes (photosynthesis and nitrogen fixation) but also for some of the particular aspects regarding regulation of nitrogen fixation that were first elucidated in purple non-sulphur phototropic bacteria of the genus Rhodospirillaceae. The work described in this thesis focuses on nitrogen fixation regulation in Rhodospirillum rubrum, a member of the Rhodospirillaceae family.

3.1.1. The main character, Rhodospirillum rubrum

In 1887 Erwin von Esmarch described a micro organism that he had found living in a sample of Berlin’s tap water in which a dead mouse had been incubated. He could successfully isolate this micro organism and observe its “beautiful wine-red colour”, the spiral shape of the cells and the fact that stab cultures would lose their characteristic pigmentation in the part exposed to oxygen 14. He named this organism, due to its colour and shape, “Spirillum rubrum”. Later on, the same organism was shown to live in freshwater sediments (perhaps from there contaminating the water sample used by Esmarch) all around the world and to be a purple non-sulphur photosynthetic bacterium by Hans Molich, who then renamed it “Rhodospirillum

rubrum” 14.

R. rubrum is an α-proteobacterium, able to grow on a very wide selection of conditions and substrates. It is a Gram-negative bacterium and its inner membrane forms invaginations that upon cell breakage result mostly in inside-out vesicles. These vesicles, called chromatophores, were found ideal for bioenergetics studies, as they contain the photosynthetic apparatus and the ATP synthase complex, among other components 15-17. R. rubrum was the model organism used to study, among other phenomena, the connection between cell illumination and oxidation of cytochromes 18, as well as photophosphorylation and light-induced redox reactions, using chromatophore preparations 19-21. In 1949 Kamen and Gest described the light-dependent hydrogen production and nitrogen fixation by R. rubrum 22, 23. It was the first

13

time such activity was demonstrated in photosynthetic bacteria and later it was also in R. rubrum that the metabolic regulation of nitrogenase by ADP-ribosylation was demonstrated for the first time 24-26. R. rubrum is a very versatile organism and can grow both aerobically in the dark and anaerobically in the light, using several carbon and nitrogen sources 27. Its genome is publicly available and annotated at the US Department of Energy Joint Genome Institute (http://www.jgi.doe.gov), a tool that greatly facilitates genomic and proteomic studies.

3.2. Nitrogenase and nitrogen fixation

The enzyme complex responsible for reduction of dinitrogen is, as mentioned, nitrogenase. Even though the presence of nitrogenase had already been known at the time it was not until 1960 that Carnahan and collaborators succeeded in preparing cell-free extracts with nitrogenase activity from C. pasteurianum

28. They also described that its activity would be lost within 30 seconds when extracts were exposed to oxygen. This extreme sensitivity to oxygen is due to particular details of the structure of the complex, as we shall discuss further on, and hints at the possibility that this enzyme system could have evolved prior to the existence of an oxidizing atmosphere 29. Three main types of nitrogenase have been described so far. The most common and most widely studied form is the molybdenum-containing type 30, 31. A second type of enzyme is usually referred to as “alternative nitrogenase”, as it contains either vanadium or iron, in alternative to molybdenum 32. The different nitrogenases share many characteristics and some organisms have the capacity to express two (as R. rubrum) or even three types of nitrogenase, although not simultaneously 33, 34. The alternative nitrogenases are usually only expressed when Mo concentrations in the environment are limiting and it has been suggested that these systems could be primitive lineages of the nitrogenase enzyme that have been kept throughout evolution 35. A third kind of nitrogenase has been found only in Streptomyces thermoautotrophicus. This is a very dissimilar type of nitrogenase, depending on superoxide oxidation for activity 36 and several of the details regarding its reaction mechanism still await further study.

14

The molybdenum-containing nitrogenase (or simply “nitrogenase”, as it shall henceforth be referred to) is a complex of two metalloproteins, dinitrogenase reductase or iron protein (Fe protein) and dinitrogenase or molybdenum-iron protein (MoFe protein) 37, 38. The atomic structure of the nitrogenase complex (from A. vinelandii) is shown in Figure 2A. As can be seen, dinitrogenase is an α2β2 heterotetramer, with two NifD and two NifK subunits whereas dinitrogenase reductase is a homodimer of the NifH protein 37, 39-41. Nitrogen reduction takes place deep within the NifD subunit of dinitrogenase, where the iron-molybdenum cofactor (FeMo-co) resides 42, 43. At the interface between NifD and NifK subunits in the αβ dimer lies an [8Fe:7S] cluster, the P-cluster 43, 44. The NifH dimer possesses a [4Fe:4S] cluster, at the interface between monomers, as well as MgATP binding and hydrolysis sites 45. It is through this metallocluster that electrons are transferred from the physiological electron donors to the P-cluster in dinitrogenase and from this to the FeMo-co and dinitrogen (Figure 2A). All of these metalloclusters are extremely sensitive to oxidation, thus explaining the sensitivity of nitrogenase to oxygen 46. The synthesis, maturation and insertion of the FeMo-co into the core of dinitrogenase require an elaborate network of several gene products, part of the nif (nitrogen fixation) gene cluster and, in a series of recently published experiments, this process could be reconstituted in vitro 41, 47. As mentioned above, electrons are transferred from the iron-sulphur cluster within the Fe protein to the P-cluster and FeMo-co in a NifDK dimer. The Fe protein physically docks to the MoFe protein and, after electron transfer, hydrolysis of the MgATP molecule bound to the Fe protein will result in dissociation of the nitrogenase complex. Each ATP hydrolysis cycle (2 ATP per cycle) results in one electron transferred to the P-cluster (Figure 2B), making nitrogen reduction a highly demanding reaction 48-51. A minimum of eight electrons and, correspondingly, 16 ATP molecules are required for reduction of one dinitrogen molecule to ammonium. As part of the nitrogen reduction reaction, protons are always reduced in at least a 2:1 molar ratio to N2, even if the latter is present at very high concentrations 52, 53 (Equation 1).

N2 + 8 H+ + 8 e- + 16 ATP � 2 NH3 + H2 + 16 ADP + 16 Pi (Equation 1)

15

A

B Figure 2 – Three-dimensional structure and turnover cycle of the 2:1 nitrogenase complex, stabilized by ADP-AIF4

–. A – Atomic structure generated using the coordinates for the A. vinelandii enzyme (PDBID: 1N2C) and PyMol 1.0. NifDK subunits are coloured in shades of blue and NifH in green. The bound ATP and metal centers are only highlighted in one half of the complex, for simplicity. B – Nitrogenase turnover cycle. Nitrogenase subunits are coloured as in Figure 2A. Colour of the [4Fe-4S] cluster of NifH denotes oxidation state - red, reduced; green, oxidised. Adapted from 54. The components are not to scale. Nitrogenase can also catalyze the reduction of a variety of other substrates including acetylene 53, 55. The acetylene-reducing capacity of this complex plays a major role in nitrogen fixation research as the product of reduction, ethylene, can be measured using gas chromatography 53. This method is easier to use, faster, less expensive, and more adaptable to field work than the previously used manometric methodologies and is still the method of choice employed in nitrogen fixation studies in laboratories throughout the world.

16

3.2.1. Electron transport to nitrogenase in R. rubrum

The electron transport chain to nitrogenase is in many diazotrophs poorly characterized and can have different unique aspects depending on the organism. Nevertheless, all nitrogenase enzymes require low potential reducing equivalents. Work in our laboratory in recent years has shed some light on the electron transport to nitrogenase in R.

rubrum. Two pathways were found to be operating in R. rubrum nitrogen fixing cells, with the major pathway operating under photoheterotrophic conditions and using the FixABCX proteins 56, 57. The Fix system is similar to the electron transfer flavoprotein (ETF) and electron transfer flavoprotein ubiquinone-oxidoreductase (ETF-QO) involved in electron transfer reactions in both mitochondria and bacteria 58, even though in R. rubrum both fixABCX and “typical” ETF/ETF-QO genes can be found, just as in other diazotrophic bacteria, such as Bradyrhizobium japonicum 59. R. rubrum FixC was shown to associate to the membrane and to bind FAD 56. FixX is a ferredoxin-like protein, with an iron-sulphur cluster binding motif, showing similarity to ETF-QO. R. rubrum FixA and FixB are similar to ETF and were shown to have NAD(P)H dehydrogenase activity and bind one FAD molecule per dimer 60. It was proposed that FixAB could be reduced by NADH and the electrons then transferred to the FAD moiety in FixC. FixC would deliver one of the electrons up-gradient to the iron-sulphur cluster in FixX and another electron down-gradient to ubiquinone in the membrane, generating reduced ubiquinol that would be reoxidized by reversed electron flow, generating NADH 60. This reversed electron flow would be supported by the proton motive force generated by photosynthetic activity (see Figure 3) and some data exists supporting the importance of an energized chromatophore membrane for the process of nitrogen reduction in R. rubrum 61 as well as in other diazotrophic bacteria 62, 63. A similar process has also been proposed to occur in photoheterotrophically grown Rb. capsulatus, as a means to maintain redox balance in the cell 64. The second electron pathway to nitrogenase in R. rubrum includes NifJ, a pyruvate:ferredoxin oxidoreductase, which can directly link the oxidation of pyruvate to reduction of ferredoxin 65, 66. This pathway was shown not to be required for nitrogen fixation under photoheterotrophic conditions 57, although it could play an important role under anaerobic conditions in the dark. In both cases the major

17

direct electron donor to dinitrogenase reductase was shown to be the ferredoxin FdN 67.

Figure 3- Proposed mechanism for electron transfer to nitrogenase through the FixABCX system, using reversed electron flow. The proton gradient generated by photosynthesis is used by NADH dehydrogenase (I) to reoxidise the quinone pool, generating NADH. Succinate dehydrogenase (II) activity and electron transfer from FixC reduce the quinone pool. The different components are not to scale.

3.3. Nitrogen assimilation

Ammonium, once produced by the nitrogenase reaction or obtained from the surrounding environment, is assimilated in the form of amino acids. There are two major pathways for nitrogen assimilation, one through glutamine synthetase (GS) and glutamate:2-oxoglutarate aminotransferase (GOGAT), also named the GS/GOGAT cycle

68, and another using glutamate dehydrogenase (GDH). In both of these cases, the end-products are glutamate and glutamine. We shall now examine these two pathways in more detail.

3.3.1. The GS/GOGAT cycle

Glutamine synthetase is a key enzyme in bacterial metabolism as it provides the only reaction for glutamine biosynthesis. Due to its vital role, GS deletions will result in glutamine auxotrophy 69. The GS/GOGAT cycle is the preferred means for nitrogen assimilation in energetically favourable, nitrogen-poor conditions 70,

18

71. In it, ammonium is condensed with glutamate to yield glutamine, a reaction that depends on hydrolysis of ATP 72 (Equation 2). Following glutamine synthesis, GOGAT will use it to transaminate one molecule of 2-oxoglutarate (2OG) to glutamate, thus completing the cycle and releasing one molecule of glutamate for metabolic purposes 69 – see Equation 3.

NH4+ + glutamate + ATP � glutamine + ADP + Pi

(Equation 2)

Glutamine + 2OG + NADPH � 2 glutamate + NADP+ (Equation 3)

The net result of the action of these two enzymes is thus the biosynthesis of a biological amino group donor (glutamate), at the expense of ATP and reducing power. This pathway is preferred in conditions of nitrogen deficiency over the glutamate dehydrogenase pathway owing to the low Km for ammonium (0.1 mM) and glutamine (0.25 mM) of GS and GOGAT, even if it requires an “extra” input of energy in the form of ATP 72, 73. Both GS and GOGAT have been purified from R. rubrum and partially characterized 74, 75. GS, as we shall discuss later on, is tightly regulated, a fact that can be related to the energetic demands of the overall reaction.

3.3.2. Glutamate dehydrogenase

The second pathway for nitrogen assimilation utilizes glutamate dehydrogenase (GDH). This enzyme, unlike the GS/GOGAT cycle, does not require ATP hydrolysis, condensing ammonium with the carbon skeleton of 2OG, directly producing glutamate, at the expense of NAD(P)H 69 – Equation 4.

NH4+ + 2OG + NAD(P)H � glutamate + NAD(P)+

(Equation 4) Escherichia coli GDH has a higher Km for ammonium than GS and is mostly used in environments where ammonium supplies are plentiful. Also, as it does not require ATP hydrolysis, GDH activity is usually correlated with lower energy conditions 70, 71.

19

3.4. Regulation of nitrogen fixation and assimilation

As can be easily deduced from all the equations presented above and from the complexity of the synthesis, maturation and insertion of the FeMo-co into nitrogenase, all of these processes require a very large input of energy and cellular resources. We shall now scrutinize how the processes of nitrogen fixation and assimilation are regulated, with special focus on the case of R. rubrum. Given that the PII proteins play a major role in the regulation of both processes, we begin by turning our attention to them.

3.4.1. The role of the PII proteins in nitrogen fixation and assimilation

It was through the pioneering work of Bennett Shapiro that the PII proteins were first discovered. While trying to isolate the GS-modifying enzyme, Shapiro separated two protein constituents as gel filtration peaks, which he named PI and PII 76. His work established that adenylylation of GS could only be obtained if the two fractions (PI and PII) were present, along with ATP and 2OG. It took more than ten years before the PII fraction could be identified in Klebsiella

aerogenes fractions as the protein GlnB 77. Since then, the importance of PII proteins and their widespread presence throughout all domains of life have been well established 78. Presently, the designation “PII proteins” encompasses what is known as a “superfamily” of proteins. These are usually small proteins, of roughly 12-13 kDa, forming (homo)trimeric complexes. PII proteins have a few loop regions, one of which, a protruding, flexible loop named T-loop is of key importance to signalling and protein-protein interactions 79, 80. PII proteins are usually divided into 3 subgroups, each named after an archetype protein, regarding their different functions in cellular processes, gene organization and protein sequence. These are named GlnB-like, GlnK-like or NifI-like 78. Proteins of the first subgroup, including all GlnB-like proteins and named after the homonymous protein of E. coli, are found in monocistronic operons or co-transcribed with either glnA (the gene encoding GS) or nadE (encoding an ammonium-dependent NAD synthetase) 78. The genes encoding the proteins in the second subgroup, named after the second PII paralog found in E. coli, GlnK 81, are usually found in an operon with the gene encoding the ammonium transport protein

20

AmtB 82. However, it has become clear in recent years that this is not always the case, as in Azospirillum brasilense, where the glnZ gene encodes a GlnK-like protein, despite the fact that it is not linked to an amtB gene. These proteins are then classified as GlnK-like based also on sequence similarity 78. While these two first groups can sometimes have overlapping functions in the cell, the third group (NifI-like) is rather distinct from the first two, containing the PII-like proteins found in association with the nifH gene in Euryarchaeota, clostridia, chlorobia and δ-proteobacterial diazotrophs 83. These usually occur in pairs (nifI1 and nifI2), also known as nifH-proximal and nifH-distal genes. The functions of NifI-like proteins are also distinct from that of other PII proteins, being involved in post-translational regulation of nitrogenase activity by direct interaction with dinitrogenase (see section 3.4.3) 84. PII proteins are signal integrator molecules known for their roles in signalling and regulation of metabolic processes, through protein-protein interactions. These interactions are affected by binding of small effector molecules (ATP, ADP, 2OG) 85, 86 and/or by post-translational modification 78, 83, 87, although this is not a universally conserved regulatory strategy. As demonstrated by work in the laboratory of Alexander Ninfa using the GlnB and GlnK proteins from E. coli, PII trimers can bind up to three molecules of 2OG and ADP/ATP per trimer 85, 86. The binding of ATP exhibits negative cooperativity and, upon ATP binding, up to three molecules of 2OG are also able to bind. 2OG can only bind to the PII protein trimer when ATP is present and, as ADP will compete with ATP for the same binding site, higher ADP:ATP ratios will displace 2OG, due to conformational changes in the proposed 2OG binding site 88, 89. Recently a high resolution structure of Az. brasilense

GlnZ showed that 2OG is bound at the interface between the T-loops of two adjacent PII protein subunits in the trimer. This binding site is in close proximity to the ADP/ATP binding site, thus explaining the effect of ADP:ATP ratios on 2OG binding 90. The characteristics of ATP and ADP binding allow the PII trimer to act as an adenylate energy charge sensor 85 and have actually been explored to create a novel in vivo ADP:ATP ratio sensor 91. At the same time, 2OG can be seen as a cellular carbon/nitrogen status indicator, with 2OG levels usually inversely correlated with nitrogen availability 78, 92, 93. By the binding of these different molecules PII proteins integrate diverse metabolic signals and, by altering the three-dimensional structure of

21

the trimer and consequently the interactions with other proteins, they transduce different signals. PII proteins can also in many cases be post-translationally modified. The most common form of modification in proteobacteria is the uridylylation of a conserved tyrosine at position 51, as is the case for E. coli GlnB 94, 95. Other types of modification, phosphorylation of a conserved serine in cyanobacteria 96 and adenylylation in actinobacteria (such as Corynebacterium glutamicum) 97, have also been reported, usually as a reflection of nitrogen availability status 78,

87. The additional complexity added with the possibility of one, two or three modified subunits in the PII trimer leads to the (theoretical) availability of some 56 different conformational and, consequently, signalling states 98. In R. rubrum three different PII paralogs have been described and named GlnB, GlnJ and GlnK 99. The gene for the first paralog (glnB) is co-transcribed with glnA, as is characteristic of GlnB-like PII proteins. The transcription of the glnBglnA operon was shown to be under the control of two different promoters, a nitrogen-responsive σ

54-dependent promoter in addition to the constitutive σ70 promoter

(see section 3.4.3. for more details). The final transcript is also processed, yielding more glnA than glnB 100. glnJ and glnK are both found in association and co-transcribed with amtB genes, amtB1 and amtB2 respectively. Both GlnJ and GlnK belong to the GlnK-like family. While the glnJamtB1 operon was shown to be regulated in response to nitrogen availability, depending on the NtrBC system, the glnKamtB2 operon does not respond to different nitrogen levels. The R. rubrum glnK transcript is, in fact, the least abundant of all those encoding PII proteins 101.

3.4.2. PII protein uridylylation

Modification of PII proteins adds an extra dimension of complexity to regulation through protein-protein interactions. In proteobacteria PII proteins are uridylylated by a bifunctional enzyme named uridylyltransferase (UTase) / uridylyl-removing (UR) enzyme or GlnD, encoded by glnD. The uridylylation activity of R. rubrum UTase/UR requires the binding of ATP and 2OG by the PII proteins, which will induce a conformation favouring binding and uridylylation by UTase/UR. Unlike the case for the E. coli enzyme, high glutamine concentrations

22

do not inhibit PII uridylylation. The deuridylylation reaction, on the other hand, is highly stimulated by glutamine binding to UTase/UR, with ATP and 2OG concentrations having little or no effect on activity (Figure 4). The presence of equal amounts of unmodified PII protein will inhibit the deuridylylation of modified PII in vitro 102. It was also recently shown that varying ADP:ATP ratios can influence the uridylylation activity, with higher ratios translating into less GlnJ uridylylation 103.

Figure 4 – Regulation of R. rubrum PII uridylylation status by UTase/UR

3.4.3. Genetic regulation of nitrogen fixation

Nitrogen fixation is a demanding process for the cell. As nitrogenase has a low turnover number, diazotrophic organisms produce large amounts of this enzyme (as much as 20% of the cellular protein content in some cases) in order to survive using dinitrogen as a nitrogen source. In addition, owing to the oxygen-sensitivity of the metal clusters, nitrogenase must be protected so as to avoid exposure and inactivation by oxygen. These responses include growth restriction to intracellular anaerobic conditions, metabolic and/or physiological adaptations and also strict genetic regulation in response to oxygen 54, 104. Genetic control of nitrogen fixation in response to nitrogen availability is a common trait in diazotrophs, with NifA as the transcription factor responsible for regulation of nif gene expression in Proteobacteria 54. The N-terminal of NifA proteins harbours a GAF domain, which has been shown to be of great importance to signalling, in some cases being able to bind small molecules 105. In A. vinelandii, 2OG binding by NifA modulates the activity of a second domain, the AAA+ domain, in response to interaction with another regulatory

23

protein, NifL 106. However, this has not been demonstrated for other nitrogen fixing bacteria, such as K. pneumoniae or Az. brasilense 54. Even though strategies may differ, the outcome is always to regulate expression and/or activation of NifA. In many organisms, such as K.

pneumoniae, NifA expression is regulated by the activation of the two-component sensor-transducer NtrB/C system. Under nitrogen limitation, the histidine kinase activity of NtrB catalyzes the phosphorylation of NtrC and thus activation of transcription of the nifLA genes. Expression and activation of NifA induces transcription of other nitrogen fixation (nif) genes. In nitrogen rich conditions, the phosphatase activity of NtrB catalyzes the dephosphorylation of NtrC, inactivating nif gene transcription. The activity of NtrB is in turn regulated by interaction with GlnB 107. In R. rubrum strains expressing the unmodifiable variant Y51F-GlnB, i.e. with a substitution at the conserved uridylylation site, activation of nif gene expression is severely impaired and nitrogenase activity is therefore greatly reduced. A strain in which the glnB gene is deleted shows no activity and only very minor expression of nitrogenase 108. The presence of GlnB and its uridylylation state are, thus, important for nitrogenase expression in R. rubrum. However, unlike the cases previously mentioned, the NtrB/C system is not required for NifA expression, with R. rubrum ntrBC deletion strains having 80% of the nitrogenase activity of the wild type cells 109. The activity of NifA is directly regulated by interaction with GlnB and it is the uridylylated form of the PII protein that will activate NifA and, consequently, nif gene transcription 110. Therefore, the uridylylation status of GlnB and the activity of UTase/UR are crucial for nitrogenase expression. In some organisms, such as Rb. capsulatus and Rb. sphaeroides, there is a higher order of regulation, through the RegB/RegA two-component system. This system provides an additional layer of complexity by responding to changes in the redox and energy status of the cells. In Rb. capsulatus RegB/RegA have been implied in the regulation of such diverse processes as photosynthesis, the transition from aerobic to anaerobic respiration, carbon fixation and nitrogen fixation 111. Also, in Rb. sphaeroides, mutations in the regB gene not only block transcription of the cbb regulon (containing the genes for RuBisCO and other enzymes of the Calvin-Benson-Bassham [CBB] cycle for carbon fixation) but also of the nif operon. Furthermore, strains carrying a non-functional cbbM gene (the gene for RuBisCO) are able to derepress nitrogenase synthesis, even in the presence of

24

what would usually be inhibitory ammonium concentrations 112. The same effect has also been observed in an R. rubrum RuBisCO-deficient strain, which hints at the existence of a global regulator similar to RegB/RegA 112. However, no homologues for this particular system have so far been found in R. rubrum thus still leaving this question unanswered.

3.4.4. Metabolic regulation of nitrogen fixation

In addition to the control of nitrogenase expression, some organisms also employ post-translational modification of dinitrogenase reductase in order to control nitrogenase activity in response to extracellular stimuli. Already in their first studies on nitrogen fixation in R. rubrum

Kamen and Gest had noticed that addition of ammonium would temporarily inhibit hydrogen evolution in nitrogen fixing cultures 113. Later studies showed that the period of inhibition has a direct correlation with the amount of ammonium added 114. R. rubrum

nitrogenase can also be inactivated in response to the addition of glutamine 115 as well as in response to light withdrawal 116 and, as in the case of ammonium, this inactivation is reversible and due to post-translational modification 115, 116, an effect known as nitrogenase switch-off. Through the work of Neilson and Nordlund it was also found that post-translational regulation of nitrogenase was lost upon cell breakage as none of the conditions inducing switch-off in vivo had the same effect on extracts 115. The nature of the modifying group and the location of the modification has been the subject of study of several groups 117-119. In 1985 the modification was shown to be an ADP-ribosylation in an exposed and conserved arginine residue of NifH (Arg101), within an equally conserved site in all NifH homologues 119. The modification hinders the formation of the Fe protein – MoFe protein active nitrogenase complex 120, 121, as it is located in the area that participates in the docking to the MoFe protein 13 thus inhibiting electron transfer and consequently nitrogen reduction. Modified Fe protein does not compete with the unmodified form for the MoFe protein, when both are present in nitrogenase activity reaction mixtures 120. Only one of the two NifH subunits in the dinitrogenase reductase complex is modified, giving modified Fe protein a characteristic double band pattern in SDS-PAGE gels, a tool that allows researchers to easily follow NifH modification in vivo as well as in vitro 116. The double

25

band pattern of post-translationally modified Fe protein has been used extensively as a marker for identification of other diazotrophic bacteria having a similar regulatory system.

The activating enzyme – DRAG

By the time the nature of the modifying group was discovered it had already been known for some time that a protein component was needed for reactivation of inactive nitrogenase. This “activating factor” was specific for modified NifH, required MgATP and Mn2+ or Fe2+ for activity and could be released from R. rubrum chromatophore membranes using 0.5 M NaCl 122, 123. It was later purified and characterized and shown to be a 32 kDa monomeric protein, which was then named Dinitrogenase Reductase Activating Glycohydrolase (or DRAG) 124. DRAG is constitutively expressed 125 from the draTGB operon, an operon containing also the gene for the inactivating enzyme (DRAT) as well as a protein of so far unknown function (encoded by draB) 126. The constitutive expression of both DRAG and DRAT is slightly surprising as under nitrogen rich conditions the substrate for both of these enzymes (Fe protein) will not be expressed. So far, no other activities are known for DRAG or DRAT in addition to the regulation of nitrogenase activity. DRAG was purified and characterised and, despite initial reports that claimed DRAG to be an oxygen-sensitive enzyme, this was later shown to be merely an artefact of the preparation 127. As the name indicates, DRAG catalyzes the removal of the ADP-ribose group from dinitrogenase reductase, with a Km of 74 µM 124. Recently, a high-resolution structure of R. rubrum DRAG was obtained. A reaction intermediate using ADP-ribose could also be resolved to high resolution and provided excellent insights into the reaction mechanism of this enzyme 128.

The inactivating enzyme – DRAT

Soon after the identification of the modifying group Paul Ludden and co-workers purified the enzyme responsible for the inactivation activity to near homogeneity from R. rubrum cells and proposed the name Dinitrogenase Reductase ADP-ribosylTransferase (DRAT) 129. The enzyme was shown to be a monomeric protein of about 30 kDa and, like DRAG, it is present in very scarce amounts in the cell 129.

26

DRAT was shown to be highly specific for the Fe protein 129 and to prefer the oxidised form. However, and unlike DRAG, air-oxidized Fe protein is not a substrate for DRAT and all activity assays must, therefore, be performed under anoxic atmosphere 130. DRAT modification of R. rubrum Fe protein requires an ADP-ribose donor, with (β-)NAD+ being the most physiological supplier, with a Km of 2 mM 129, even though other molecules, such as nicotinamide hypoxanthine dinucleotide or etheno-NAD can substitute for NAD+ in

vitro 130. NAD+ is required for DRAT-Fe protein interaction 131, as well as ADP and it has been speculated that ADP binding to Fe protein regulates modification by DRAT 131, although, as we shall see later on, this may not be the case.

Regulation of the switch-off effect

As DRAT activity requires NAD+ and DRAG activity will cleave the ADP-ribose group from the Fe protein, the two enzymes should not be active at the same time, in order to avoid depleting the cellular NAD pool. Indeed, both DRAG and DRAT are subject to post-translational control. However, the nature of the signals regulating their respective activities is still, even after more than 30 years, the subject of ongoing research. The signal transduction pathway seems to be kept throughout a variety of organisms as was shown by switching the draTGB operon between R. rubrum and Az. brasilense 132. In fact, even organisms that do not possess such a post-translational regulation of nitrogenase were able to modify the Fe protein in response to switch-off stimuli once the dra operon was transferred to them 133, 134, thus demonstrating the seemingly widespread nature of the signal transduction pathway. In R. rubrum the PII proteins GlnB and GlnJ were shown to play a vital role in the regulation of this activity. Either one of the two proteins can substitute for the other in the regulation of post-translational modification of nitrogenase but none can be substituted by GlnK in that role 99. The relevance of PII proteins for post-translational regulation of nitrogenase has also been demonstrated in other organisms, such as Rb. capsulatus 135 and Az. brasilense 136, 137.

DRAG regulation by membrane sequestration

As previously mentioned DRAG is found in association with the membrane in both ammonium and in darkness switched-off R. rubrum

27

cells. Membrane sequestration is thought to be the mechanism by which DRAG is inactivated during switch-off, as it will not allow interaction with its subtrate (ADP-ribosylated-Fe protein). Recently several advances have furthered our understanding of this mechanism and shown that the ammonium/methylammonium transporter AmtB, in complex with GlnK or GlnK-like proteins, is crucial for correct regulation in ammonium switch-off. The widespread genetic linkage observed for the amtB and glnK genes 82 suggested the existence of a physiological link between the two proteins and, as observed in E. coli and A. vinelandii, GlnK is localized to the inner membrane in response to a variation in the nitrogen status in an AmtB-dependent manner. In nitrogen-limited cells GlnK is uridylylated and, even though both AmtB and GlnK expression is highly induced, the two proteins do not interact until an influx of nitrogen brings about the deuridylylation of the PII protein.138. Thus, the UTase/UR control of the PII uridylylation status allows efficient signalling of the nitrogen status, transmitted through a differential localization of GlnK. Structures of this AmtB-GlnK complex from E. coli are already available and have provided valuable information on how GlnK regulates transport through AmtB 139, 140. An interesting detail of these structures is that ADP was found in all nucleotide-binding sites of GlnK. It was shown that AmtB-GlnK complex formation in E. coli requires the presence of ADP and that GlnK can be specifically released from AmtB-containing membranes by washing with ATP and 2OG 141. In R. rubrum GlnJ was also found to require AmtB1 for membrane association 142 and, as in the case of E. coli, the complex can be dissociated in the presence of Mn2+, 2OG and low ADP:ATP ratios 103,

143. AmtB2, the other AmtB homologue, has no influence on either GlnJ association to the membrane or nitrogenase switch-off 144. It should at this point be noted that it is the unuridylylated form of GlnJ that binds to AmtB1 and that the conditions required for complex dissociation (MnATP and 2OG) are also the conditions stimulating uridylylation of PII proteins, as previously discussed. Upon ammonium switch-off in R. rubrum, DRAG is localized to the chromatophore membrane and this association was shown to be dependent on the presence of AmtB and GlnJ. As a consequence, AmtB1 mutants have impaired ammonium switch-off 142. Membrane sequestration effectively separates DRAG from ADP-ribosylated-Fe protein, allowing nitrogenase switch-off. AmtB-dependent membrane

28

sequestration of DRAG was also seen in Az. brasilense 145. However, the specific membrane partner for DRAG association is still not yet clearly identified. DRAG could bind “on the side” of the AmtB1-GlnJ complex, interacting with the interface between the two proteins or only with AmtB1. A third possibility is that there is a still unknown membrane partner in the vicinity of the AmtB1 protein with which DRAG could also interact. Nitrogenase is also modified in response to energy signals, namely darkness in R. rubrum and Rb. capsulatus

146 and anaerobiosis in Az.

brasilense 147. However AmtB1-deficient R. rubrum strains respond in a different manner to darkness, depending on the nitrogen source used. Indeed, when using glutamate-grown nif-derepressed R. rubrum strains lacking AmtB1, the darkness switch-off response is perturbed, with nitrogenase activity being reduced to ~50% 144. However, when grown diazotrophically, darkness switch-off is only marginally affected in R. rubrum amtB1 mutant strains 101. As recently described, PII proteins are differentially uridylylated upon darkness-induced switch-off, depending on which nitrogen source is used. In glutamate-grown cultures PII proteins will be deuridylylated in response to darkness whereas diazotrophically-grown cultures will maintain PII proteins uridylylated 148. This difference could eventually play a significant role in the control of nitrogenase darkness switch-off in these two conditions as an uridylylated GlnJ will not associate to AmtB1 and, thus, DRAG binding to the membrane could be partially hindered. However, as DRAG can still be found to a large extent by the membrane upon darkness switch-off in diazotrophic conditions, even in a strain lacking AmtB1 – in which (uridylylated) GlnJ is only found in the cytosolic fraction 101 – the darkness switch-off response in R. rubrum under these conditions may effectively include an as yet unknown membrane partner. A similar effect is observed in AmtB-deficient Rb. capsulatus strains, with switch-off being affected only in response to ammonium but not to darkness 146. It could be speculated that another (presumably integral or membrane-associated) protein is involved in DRAG membrane sequestration. This putative protein could be involved in sensing energy changes in the cell and, by interaction with DRAG and/or AmtB, regulate the darkness switch-off response.

29

DRAT regulation by protein-protein interaction

Whereas some of the details regarding DRAG regulation in nitrogenase switch-off are already known, the mechanisms controlling DRAT are much less understood. DRAT fractions obtained from R.

rubrum cultures were shown to be able to modify Fe protein in vitro 129. However, this “purified fraction”, as well as those used in later studies based on the same purification procedure, contains a low-mass contaminant, which was at the time considered unimportant for the regulation of the ADP-ribosylation system. On the other hand, attempts to remove this contaminant fraction lead to “unacceptable losses” of DRAT activity 129. It was demonstrated by yeast-two hybrid studies that DRAT and GlnB from both R. rubrum and Rb. capsulatus can form a complex 149,

150. This putative complex was further studied using purified GlnB and DRAT proteins from Az. brasilense. DRAT is a notoriously unstable protein and Az. brasilense DRAT over-expression in E. coli usually results in the formation of inclusion bodies. However, co-expression of GlnB with DRAT was found to decrease the formation of these inclusion bodies and stabilize DRAT in solution 151. In vitro studies using the purified proteins from Az. brasilense showed that DRAT interaction with GlnB is stabilized by the presence of ADP and, conversely, destabilized by addition of ATP and 2OG 151. It should be noted once more that the same conditions that result in Az. brasilense

DRAT-GlnB complex dissociation in vitro allow release of GlnJ from R. rubrum chromatophore membranes and uridylylation of PII proteins. Modulation of the interaction is most probably due to binding of the small effector molecules by GlnB with the consequent changes in three-dimensional structure of the PII trimer regulating its interaction with DRAT. Huergo and co-workers have speculated that this interaction with GlnB could regulate DRAT activity in vivo in response to fluctuations in the ADP:ATP ratios and 2OG concentration, which would also explain the afore-mentioned need for ADP in the DRAT activity assay 129. We shall discuss the influence of GlnB on DRAT activity in a later section. The redox state of the pyridine nucleotide pool was also previously implicated in the regulation of the switch-off effect, with NAD+ being able to induce Fe protein ADP-ribosylation and consequent switch-off 152, 153. As NAD+ is the preferred substrate for DRAT it could be that induction of switch-off is a reflex of the higher abundance of one of the substrates (NAD+). Also, a change in the NAD+:NADH ratio (and,

30

thus, the cellular redox state) could be sensed by other protein(s) influencing DRAT or DRAG activity. Indeed, a decrease in the intracellular NAD(P)H concentration was observed when R. rubrum

cells were subjected to different switch-off effectors 152.

Other strategies for nitrogenase switch-off

So far we have only discussed post-translational regulation of nitrogenase by ADP-ribosylation. However, nitrogenase has also been shown to be regulated by direct interaction with proteins of the PII family in Methanococcus maripaludis. In this methanogenic archaea the PII protein homologues NifI1 and NifI2 form a complex that directly interacts with the NifDK nitrogenase subunits and effectively interrupts electron flow in response to ammonium, causing switch-off 84, 154, 155. As these PII proteins are highly spread among archaea and anaerobic bacteria it has been suggested that this may be a more common regulatory strategy than initially though 84. Some studies in cyanobacteria have also hinted at the possibility that at least in some cyanobacteria nitrogenase may be regulated be post-translational modification 156, 157. However, further studies are necessary as substantial data is still missing to validate those claims.

3.4.4. Genetic and metabolic regulation of nitrogen assimilation

Nitrogen assimilation, just as nitrogen fixation, is the subject of tight regulation, both at the transcriptional and post-translational level. Unlike nitrogen fixation, the assimilation process must be present at all times, to ensure the cells are supplied with the preferred biologically usable forms of nitrogen – amino acids, usually glutamate and glutamine. The process of nitrogen assimilation through the GS/GOGAT pathway is so fundamental that E. coli cells devote a large percentage (ca. 15%) of their overall ATP supply to this process alone 69.

Genetic control of glutamine synthetase

Under nitrogen rich conditions, the gene for GS, glnA, is transcribed from σ70-dependent promoters, the “house-keeping” RNA polymerase form. On the other hand GS expression is induced in nitrogen

31

deficiency 78. The Ntr system has been implicated in the regulation of expression of the glnA gene in several organisms, including E. coli 69, R. rubrum 100 and K. pneumoniae 158. In R. rubrum, unlike in enteric bacteria 159, the glnB and glnA genes are together in a functional glnBA operon, which was also shown to be under the control of both σ

70- and σ54-dependent promoters, thus

responding to nitrogen availability 100, 160.

Post-translational control of glutamine synthetase

Glutamine synthetase is a known target for feedback regulation by nitrogen-containing molecules, such as alanine, serine, histidine, tryptophan and others. In the case of the E. coli the different inhibitors can only induce partial reduction of activity, in a process known as cumulative inhibition 161. The R. rubrum enzyme is also sensitive, albeit with some differences in the extent of inhibition, to some of the same feedback inhibitors as its E. coli counterpart 162. GS can also be regulated by post-translational modification. Adenylylation of a conserved tyrosine residue (Tyr397 in the E. coli

enzyme) in each of its twelve subunits decreases the activity of GS sequentially, with increasing levels of modification decreasing GS activity 163. In fact, GS adenylylation in E. coli was the first described example of the use of nucleotydylation as a means to control enzyme activity 164 and was since then found to occur in a variety of organisms, including R. rubrum 162, 165-168. Both the addition and removal of the AMP group are catalysed by the same bifunctional enzyme, adenylyltransferase (ATase). This enzyme is regulated by interaction with the PII proteins and, as the uridylylation state of the PII proteins is also controlled by the concentrations of 2OG and glutamine, one can easily correlate GS activity with that of UTase/UR. In R. rubrum any of the three (non-uridylylated) PII proteins can stimulate the adenylylation activity 99, 108 but this reaction, unlike the case for the E. coli enzyme, does not require the presence of glutamine 93. PII-UMP and 2OG inhibit the adenylylation reaction although 2OG inhibition can be relieved by high ADP:ATP ratios in vitro 103. Regulation of the deadenylylation reaction, however, does not require neither PII-UMP nor 2OG, with the only requirement being Pi (Figure 5). This makes deadenylylation the standard activity for R. rubrum

ATase, with 2OG concentration and ADP:ATP ratios controlling the

32

shift between both activities, through binding to the PII proteins (and, indirectly, by changing their uridylylation status) 93, 103. Figure 5 – Regulation of R. rubrum GS post-translational modification by ATase

GS activity is usually measured by the colorimetric γ-glutamyltransferase assay 169. This assay can be used to estimate the number of modified E. coli GS subunits by using either Mn2+ (in which case both modified and unmodified forms are active) or Mg2+ (when only unmodified GS will be active) in the reaction and determining the relation between the two 170. In the case of the R.

rubrum enzyme such estimation is not possible as the behaviour of the modified and unmodified forms of GS in respect to Mn2+ or Mg2+ differs from that of E. coli 168, 171. However, a decrease in GS activity using this assay can nevertheless be correlated with adenylylation and vice-versa 162. Why is the post-translational control of GS activity so important that it would require such an elaborate control system? One could argue that regulation of GS expression should suffice to adapt to different nitrogen supplies. However, control of expression provides a very crude means for regulation and does not allow fine-tuning of enzymatic activities in response to rapid changes in the environment. In E. coli a sudden influx of ammonium causes rapid consumption of 90% of the ATP pool and a 20-fold increase in glutamine concentration before GS activity is regulated 172 and this would be an unsustainable condition for the cells if it was maintained for very long. Likewise, it was recently shown that deleting the gene for ATase (glnE) results in severe growth defects in R. rubrum and accumulation of secondary mutations that ultimately lead to low GS activity 173. GS adenylylation could then have evolved in organisms in which GS has

33

a high specific activity as a means to protect the intracellular pools of critical metabolites.

3.5. 2-oxoglutarate, a key metabolite

Whether discussing nitrogenase expression or regulation, PII protein modification or GS adenylylation in R. rubrum, 2OG is a ubiquitous metabolite, directly or indirectly related to all of these processes. Its concentration is usually inversely proportional to the nitrogen status of the cell and usually this is perceived through the action of the PII proteins. 2OG is an intermediate of the Krebs cycle but, due to its role in the process of nitrogen assimilation and, in some cases, carbon fixation 174, it occupies a very central role in the cellular metabolic routes. In vitro, high 2OG concentrations stimulate UTase/UR to uridylylate PII proteins and this not only activates NifA and nif gene expression but it also stimulates DRAG release from the chromatophore membranes, at the same time that GS adenylylation by ATase is inhibited. The concentration of 2OG was shown to be higher in nitrogen fixing R. rubrum cells than in cultures grown in nitrogen rich conditions 93, correlating the in vitro data with physiological effects in vivo. Addition of ammonium to R. rubrum nitrogen fixing cultures was shown to cause a temporary increase in the intracellular concentration of glutamine 116, due to the increased flux through the GS/GOGAT pathway, at the expense of 2OG. The combined effect of the “glutamine surge” and 2OG pool depletion will activate deuridylylation of PII proteins, resulting in inhibition of nif gene expression, GS adenylylation and GlnJ (as well as DRAG) association to the membrane. Even though this is an over-simplified view of the regulatory schemes controlling nitrogen fixation and assimilation in R.

rubrum, as other factors such as the redox balance and ADP:ATP ratios must also be taken into account, it allows us to have a global view of the interplay between the different processes in the cell.

34

4. An overview of the field of proteomics

After the explosion of data deriving from the many gene and genome sequencing projects in the late 1990’s, the term “proteomics”, a contraction of “protein” and “genomics”, arose as an analogy to the study of the genome 175. However, the proteome is a much more fluid concept, as it refers to the entire set of proteins present in an organism or specific part of an organism, at a given time, under a certain set of conditions. As it can be easily understood, unlike the genome, which is usually a more static entity, the proteome will vary according to the environmental and intracellular conditions and, in multicellular organisms, different cells will have different proteomes, even though their genetic material should be the same. Also, if one adds the extra dimensions of post-translational modification and protein-protein interactions, the study of the proteome assumes an even more considerable breadth 176. In practical terms, proteomic analysis is usually synonymous with techniques designed to facilitate the identification of proteins in a given sample and quantitate differences in their relative abundance. It should however be noted that this is a rather reductive view of the field of proteomics as, for example, studies regarding in vitro characterization of the interaction between different proteins in a complex can also be encompassed by the definition of “proteomics”. Nevertheless, and taking the first description of practical proteomics, two major branches are today distinguished, based on the platform used to perform the study – namely, gel-based and gel-free proteomics.

4.1. Gel-based proteomics

One of the most widely spread techniques in any life sciences laboratory today is that of the polyacrylamide gel electrophoresis, with its many variants. The most common technique is that employing an anionic denaturing detergent, sodium dodecylsulphate or SDS, which allows migration of the target proteins through the polyacrylamide gel matrix upon application of an electric field and their separation solely based on the relative mass of each protein – a technique known as (one-dimensional) SDS-PAGE. In this way, complex mixtures of proteins can be easily separated, allowing the relative quantification of each band, usually by staining the protein bands with a dye 177. This is

35

a relatively fast and inexpensive technique but it also has its drawbacks. Proteins of similar mass or the same protein with different post-translational modifications, for example, will migrate closely together, sometimes in the same band and, thus, this technique has a relatively low separation capacity as well as low throughput. In order to increase resolution for many polypeptides in the same sample, several techniques emerged, making use of two different intrinsic properties of proteins. In this way, proteins could be separated according not only to their (unitary) mass but also with respect to their isoelectric point (pI) or the overall mass of the complex in which they could be found, for instance.

Two-dimensional isoelectric focusing (IEF) / SDS-PAGE (2D-PAGE)

The first attempts at establishing a two-dimensional acrylamide gel electrophoresis separation based on the proteins’ pI and mass were performed using acrylamide gels cast inside thin cylindrical tubes, containing zwitterionic molecules, termed ampholites, used to generate a pH gradient throughout the gel, upon application of an electric field. Protein samples loaded onto the gel will migrate to the zone where they have no net charge, where they will have reached their pI. The first dimension gel was then removed from the tube and loaded on a regular SDS-PAGE gel for the second dimension separation 178-180. However, this first technique had some reproducibility problems related to the pH gradient in the first dimension 181. This posed a major hurdle until the development of the immobilized pH gradient (IPG) technique. Instead of “free” ampholytes, the molecules used to create the IPG – later termed “Immobilines” – have acrylamido groups that, upon IEF gel casting, are covalently cross-linked to the gel matrix, rendering the pH gradient stable and reproducible 182. The mass production of these first-dimension gels reduced variability even further and allowed increased output on the technique, even though it is still rather labour-intensive. Already in the first attempts, 2D-PAGE was shown to be a powerful technique, allowing the separation of hundreds or even, in some cases, a few thousands of proteins from a complex sample into independent, discrete spots in a single gel 183. With the development of gel image analysis algorithms, it was now possible to instantly quantify changes in sample sets containing several hundreds of proteins. As post-translational modifications induce a shift in the pI, this technique can also monitor modifications as well as discriminate between different

36

protein isoforms. Nevertheless, 2D-PAGE also has its disadvantages and flaws, the most notorious being the under-representation of membrane proteins, very high and very low mass proteins or those with alkaline pI as well as the “shadowing” of low-abundance proteins by other more abundant proteins in the sample 184. Despite considerable advance in this area, many membrane proteins are insoluble in the detergent/chaotrope mixture usually used to prepare protein samples for the first dimension of 2D-PAGE and, of those that can be solubilised, many precipitate at or near their pI during the IEF run 185. Other specialized gel systems have been devised to address this problem, using different combinations of denaturing detergents 186 or, in order to analyse the formation of protein complexes in the membrane, a combination of gentle detergents, extracting proteins and complexes in near-native conditions, followed by a denaturing SDS-PAGE second dimension, a technique that will be described in more detail below. Despite its shortcomings, 2D-PAGE is still one of the most widely used quantitative proteomic techniques in laboratories worldwide and, in combination with mass spectrometry, is still an efficient, reliable and relatively economical tool.

Two-dimensional Blue Native (BN) / SDS-PAGE (2D BN/SDS-PAGE)

One of the most common systems used for analysis of membrane protein complexes is the one developed by Schägger and von Jagow in 1991, termed Blue Native (BN) electrophoresis. This system allows the separation of membrane proteins in (near-)native complexes, extracted from the membrane using mild, non-ionic detergents, such as dodecylmaltoside (DDM), Triton X-100 or digitonin 187. The name of the technique comes from the usage of the Coomassie Brilliant Blue (CBB) G-250 dye in the sample and cathode buffers. The binding of the dye to the protein complexes results in an overall negative surface charge, allowing them to migrate in the gel according to mass only. The individual components of each complex can be analysed using a second dimension denaturing gel run, as for the 2D-PAGE system. The overall result is a two-dimensional map that not only allows quantification of changes in the membrane and membrane-associated proteome but also allows the study of protein-protein interactions – all in the same system. As can be perceived, the system depends on the efficiency of extraction of the membrane protein complexes by the chosen

37

detergent. Even if a mild detergent is used, the addition of the CBB G-250 dye, though vital for the technique, may also cause disaggregation of some more loosely bound protein subunits 188. Despite its very informative characteristics 2D BN/SDS-PAGE is a time-consuming, labour-intense and relatively low-throughput system, just as other gel-based proteomics techniques.

4.2. Gel-free proteomics

The fastest-growing trend in proteomic studies is the use of gel-free systems, based on on- or off-line coupling of liquid chromatography and mass spectrometer systems. These systems offer several advantages when compared with the “traditional” gel-based methods: identification and relative quantification of (mostly) all the proteins present in very complex samples, higher sensitivity, reproducibility and, due to automation, higher throughput 189. These techniques usually involve fractionation using different chromatographic platforms of the peptide mixture resulting from the digestion of all the proteins in the sample and subsequent identification (and, frequently, relative quantification) of the proteins present in the sample using mass spectrometry. By analogy to the technique used to sequence large DNA sequences, this kind of approach is usually called “shotgun” proteomics 190. Frequently, in order to allow relative quantification of the samples, the peptides are tagged using, for example, different radioactive isotope-containing molecules. Even though these techniques present several improvements in the limit of detection and analysis throughput, the implementation cost is usually much higher than that of the gel-based techniques, as they require a much more sophisticated setup, both at the hardware as well as at the software level 189. As such, they are still too costly to be as widespread as the more common gel-based techniques.

4.3. Mass spectrometry

If it is true that the improvements in separation techniques allowed proteomics to grow in “size”, i.e., in resolution and in number of proteins individually separated in a sample, it was only with the implementation of mass spectrometry, in its different varieties, that quantitative proteomic studies could finally become a robust tool for research. Prior to the advent of protein mass spectrometry, protein identification was achieved by the use of either specific antibodies or

38

Edman degradation. However, even though it is a highly reliable method, Edman degradation has some severe limitations – it is very time-consuming and expensive and N-terminal modifications will limit its applicability 191. Mass spectrometers usually have three main components: the ion source (with ion accelerators and optics), the mass analyser modules and the integrated data-processing software. The adaptation of mass spectrometry ionization techniques to the ionization of peptides and proteins resulted in the introduction of “softer” ionization methods such as “matrix-assisted laser desorption/ionization” (MALDI) or “electrospray ionization” (ESI). These can ionize peptides and even whole proteins without extensive degradation. MALDI uses a laser to ionize and desorb dried samples mixed with a matrix, creating a cloud of usually singly charged ions. ESI, and its successor, nanospray ionisation, ionize peptides and proteins directly from solution, using a thin capillary to which high voltage is applied, resulting in the formation of a conic cloud of vaporised peptide ions. ESI and nanospray allow analysis of multiply charged ions and, coupled to low mass limit spectrometers, facilitate detection of high mass molecules 191. The mass analyser is the “core” of the mass spectrometer and many different types exist. One of the most common and widely spread is the “time-of-flight” (ToF), based on the time required for the different peptides to “fly” from one end to the other of the so-called “flight tube”. Other mass analysers detect mass:charge ratio (m/z) of the peptides, either by the usage of the intrinsic resonance frequency (as ion traps, Orbitraps and ion cyclotron resonance analysers) or detection of m/z stability, as in the case of ion quadrupoles (Q). Many mass spectrometers have hybrids of two or more of these analysers – an example is the ESI-Q-ToF mass spectrometer 192.

MALDI-ToF

One of the most basic and widely used setups is that where a MALDI ionization source is coupled to a ToF mass analyser – the MALDI-ToF mass spectrometer. Although it can also be coupled to a nano-liquid chromatography (LC) system to enhance sensitivity and separation of different peptides in a sample, this kind of spectrometer is commonly used off-line, in combination with gel-based proteomics. Gel spots containing the protein of interest are digested using a specific protease (most commonly trypsin) and the peptide digest is

39

analysed off-line in the MALDI-ToF. The resulting spectrum contains a combination of the different monoisotopic peptide masses resulting from the protein digestion and is compared with a database containing all the theoretical mass lists for all the proteins in a specific set, derived from genome sequences. This identification process, as it generates a characteristic mass list for each protein, is termed “protein mass fingerprinting” 192. Even though it is one of the most widely spread and used mass spectrometers around the world, the MALDI-ToF has many limitations. Its relatively low sensitivity and resolution, requiring the collection of a large number of shots for a good signal-to-noise ratio, hinder identification of low-abundance peptides and, as it uses laser pulses to ionize peptides, there is a high degree of variability between each shot.

Orbitrap

Unlike the “pulsed analysis” of the MALDI-ToF, other mass spectrometers physically “trap” the peptides in the sample to analyse, allowing a faster workflow, higher throughput, versatility and sensitivity. One of the most recent developments in this field is the Orbitrap, an ion trap whose main feature is a spindle-shaped central electrode. Commonly, a linear ion quadrupole trap is placed in front of the Orbitrap (LTQ-Orbitrap) and samples, injected from on-line LC, are ionised using ESI or nanospray. The inclusion of collision cells allows peptide sequencing, resulting in an instrument with very high sensitivity and almost unrivalled accuracy (0.2 ppm at signal-to-noise ratio of >10.000) 193. The Orbitrap is a very powerful new tool in the field of proteomics albeit still at a relatively high cost, which hinders its more widespread usage. In the first part of this thesis we have approached some of the basic concepts regarding nitrogen metabolism in diazotrophic bacteria, as well as some of the technical aspects regarding proteomic studies. We shall now discuss some of the most recent findings regarding the adaptation of R. rubrum to different nitrogen availability conditions as well as some more detailed aspects of nitrogen metabolism regulation.

40

II. Adaptation to different nitrogen conditions

1. The interconnection between nitrogen fixation and carbon fixation pathways

How does R. rubrum respond to a shift in nitrogen availability – from nitrogen rich to nitrogen fixing conditions? We have already discussed how nitrogenase is usually only expressed when nitrogen is limiting and this is, naturally, the hallmark of nitrogen fixation. The GS/GOGAT cycle enzymes as well as GDH are also differentially regulated. The expression of these enzymes is known to respond to the cellular nitrogen status, as previously discussed, and the effects seen on R. rubrum in nitrogen fixing conditions (I) correlate well with these previous observations. However, there are other effects that occur upon a change in the nitrogen status in the cell. One of the most striking is the down regulation of several enzymes involved in different carbon fixation and utilization pathways in nitrogen fixing conditions. These include ribulose–1,5–bisphosphate carboxylase / oxigenase (RuBisCO), some enzymes of the reductive tricarboxylic acid cycle and others proteins, involved in the synthesis of the storage compound poly-hydroxybutyrate (PHB) (I).

RuBisCO

The regulation of the CBB cycle enzymes has previously been shown to be linked to nitrogenase synthesis, with R. rubrum RuBisCO mutants derepressing nitrogenase synthesis even in the presence of ammonium 112. Conversely, it was also demonstrated that mutations in the FixABCX pathway would induce up regulation of RuBisCO in diazotrophically grown R. rubrum cultures 56. Thus, RuBisCO can be seen not only as an enzyme catalysing carbon fixation but also as an “electron sink”, diverting the excessive reducing power that can no

41

longer be used by nitrogenase under those conditions, in order to maintain the redox balance in the cell.

Such a correlation between RuBisCO and nitrogenase has been previously shown not only in R. rubrum but also in other purple non-sulphur nitrogen fixing bacteria, such as Rb. capsulatus, Rb. sphaeroides and Rhodopseudomonas palustris 112, 194, 195. It has been proposed that the RegA/RegB system could sense changes in the cellular redox state and regulate the activity of CbbR and NifA accordingly. The transmembrane region of RegB is thought to sense redox changes, although the exact nature of the signal and its sensing mechanism is still not understood 196. Even though there is no homologous system in R. rubrum it is clear that a similar “redox sensor/transducer” mechanism must be present, in order to regulate not only the expression of the cbb and nif operons but also of other processes, such as the PHB synthesis or regulation of the TCA cycle enzymes. Interestingly, artificially lowering the intracellular 2OG pool in nitrogen fixing conditions can also influence nitrogenase and RuBisCO expression levels. As the PII modification status is perturbed in these conditions, efficient NifA activation is also hindered, resulting in lower levels of nitrogenase. It could be that the lower levels of nitrogenase under such conditions lead to an imbalance in the redox state and that this, in turn, induces the expression of alternative electron sinks, e.g. RuBisCO (II).

The Reductive TCA cycle

Described for the first time in 1966 in Chlorobium thiosulfatophilum, the reductive tricarboxylic acid (RTCA) cycle faced severe criticism before it was finally accepted as an independent means to fix carbon in anaerobic bacteria 174. This pathway uses specific enzymes, 2-oxoglutarate synthase and pyruvate synthase, both requiring the reducing power of reduced ferredoxins for catalysis, effectively reverting the “common” TCA cycle. Coupled to this cycle, another specific enzyme, ATP-dependent citrate lyase, couples CO2 fixation by the RTCA to acetyl-CoA production

197. This system was also found to exist in R. rubrum, with activity measurements demonstrating the presence of both 2OG synthase, pyruvate synthase and citrate lyase in photolithoautotrophically-grown R. rubrum cultures. However, no citrate lyase activity could at the time be demonstrated in photoheterotrophic conditions and it could be that the RTCA cycle is incomplete under such conditions 198, perhaps being present only as a means to regulate redox poise. In fact, 2OG synthase is down regulated in nitrogen fixing conditions, mirroring the

42

behaviour of RuBisCO (I), and this could indicate that this cycle is also subjected to redox regulation.

2. Keeping the (redox) balance

Differential regulation of both carbon and nitrogen fixation and

utilization pathways is necessary in order to maintain the carbon-to-nitrogen balance in the cell but, moreover, it also allows control over the cellular redox balance. Both nitrogen fixation and carbon fixation require reducing equivalents and, therefore, it is important for the cell to be able to regulate flow through each of these pathways, so as not to generate an imbalance in the reductant pools. We have already seen how different nitrogen availability levels affect the soluble proteome, namely the level of expression of nitrogenase and RuBisCO (I and II). However, such changes also influence the chromatophore membrane components. FixC, for instance, was shown to associate to the chromatophore membrane and the fixABCX operon is subjected to regulation by nitrogen availability 56. Although the system developed for membrane complex analysis did not allow the identification of any membrane protein components associating to FixC, the higher level of FixC in the chromatophore membrane samples of nitrogen fixing cells (III) is in accordance with those previous studies.

The succinate dehydrogenase complex (Sdh) is up regulated under nitrogen fixing conditions (III). It was previously shown that under photoheterotrophic conditions the TCA cycle is essential for the maintenance of normal levels of nitrogenase activity in R. rubrum, perhaps generating the reducing equivalents needed for the nitrogen reduction activity. Specific inhibition of the TCA cycle using fluoroacetate results in decreased nitrogenase activity, an effect that can be reverted by addition of NAD(P)H 199. In E. coli the expression of Sdh and other TCA cycle enzymes is regulated by the ArcA/B and FNR sensor/transducer systems 200, 201. These systems respond to changes in carbon source and oxygen availability and it was shown that the ArcA/B system can directly sense variations in the quinone pool redox state 202, 203. Under reducing conditions ArcB, an integral membrane protein sensor kinase, is activated and phosphorylates ArcA, the cytosolic effector. This enhances binding affinity of ArcA for the promoter regions of genes encoding TCA cycle and respiratory chain components. Oxidising conditions revert this phosphorelay system 204 and, as ArcA has an inhibitory effect on the expression of the TCA cycle enzymes under

43

reducing conditions, the redox poise will in this way regulate their expression level 205. It was recently shown in R. rubrum that chromatophore membrane biosynthesis was regulated by the redox state of the quinone pool. Transfer to microaerophilic conditions and reduction of the quinone pool induces chromatophore biogenesis and it was speculated that sensing of the quinone pool redox state could be a more general signal than previously thought 206. Searches on the R. rubrum genome for homologues of either ArcA/B or FNR were unsuccessful. However, as changes in the cellular redox state are able to influence the proteome and, given that the quinone pool is able to reflect these fluctuations, it is possible that a system with similar characteristics to ArcA/B, FNR or RegA/RegB is also present in R. rubrum. As nitrogenase activity will require a great part of the cellular reducing power, this could induce an oxidation of the quinone pool and perhaps, through either direct or indirect sensing of these changes, bring about the activation of a signal relay system, ultimately regulating RuBisCO and TCA cycle enzymes, among other targets (Figure 6). Such a system could be considered a global regulator, regulating transcription/activation of NifA, CbbR and other regulators. In this way changes in the nitrogen availability would be reflected in the cellular redox balance and, to maintain redox homeostasis, in both nitrogen and carbon fixation and utilization pathways, by regulation of the proteome.

Figure 6 – Proposed mechanism for proteome regulation in different nitrogen conditions, through sensing of changes in the redox state of the quinone pool.

44

III. Post-translational regulation of nitrogenase activity – guarding the guards

1. Regulation of ammonium switch-off

Addition of ammonium to nitrogen fixing R. rubrum cultures induces, as we have discussed, DRAG and GlnJ membrane association in an AmtB1-dependent manner. In this way DRAG is sequestered and, as it is physically separated from its substrate, it is no longer able to remove the ADP-ribose moieties from the modified Fe protein. But how is DRAT activity regulated upon ammonium switch-off? It had previously been proposed, based on yeast two-hybrid and mutagenesis studies, that R. rubrum DRAT and GlnB could form a complex 149. Co-expression of N-terminally tagged R. rubrum DRAT and GlnB in E. coli showed that the two proteins can, in fact, form a stable complex in vivo. This complex, once isolated, is stabilized by the presence of ADP and destabilized by low ADP:ATP ratios and the presence of 2OG (IV). These results are in agreement with those previously reported for the Az. brasilense DRAT-GlnB complex 151. Intracellular 2OG concentrations are lower in nitrogen rich than in nitrogen fixing cells 93 and, as such, the sensitivity of the DRAT-GlnB complex to 2OG concentration could be reflected on the regulation of DRAT activity. Indeed, the DRAT-GlnB complex was found to be the physiologically active form, being able to modify R. rubrum Fe protein in the presence of ADP and β-NAD+. Lower ADP:ATP ratios and the presence of 2OG destabilize complex formation and, consequently, inhibit DRAT activity (IV). Considering the difference in 2OG concentrations in vivo between nitrogen fixing and nitrogen rich conditions it could be extrapolated that addition of ammonium will induce a higher flux through GS/GOGAT (and GDH), decreasing the internal pool of 2OG. This elevated flux through GS could also lead to a momentary increase in the ADP:ATP ratios, until GS is modified. The lower concentration of 2OG and the possible increase

45

in ADP concentration, through binding to GlnB, will stabilize the DRAT-GlnB complex and, thus, lead to Fe protein modification. It should be kept in mind that these conditions also lead to PII protein demodification and the association of GlnJ and DRAG to the chromatophore membrane. As the ammonium is consumed, the 2OG pool will once again increase and destabilize the DRAT-GlnB complex, leading to DRAT inactivation. At the same time, the elevated concentrations of 2OG will also stimulate PII modification and both GlnJ and DRAG are released from the membrane, reactivating modified Fe protein (Figure 7).

Figure 7 – Proposed mechanism for regulation of post-translational modification of nitrogenase in R. rubrum in response to ammonium. E. coli AmtB1 and GlnK atomic structures (PDBID: 2NUU) as well as R. rubrum DRAG (PDBID: 2WOE) were generated using PyMol 1.0 and the respective PDB coordinates.

In this way, intracellular 2OG concentrations will have a big impact on the regulation during ammonium switch-off, controlling both DRAT-GlnB complex formation (and DRAT activation) as well as regulating GlnJ and, through it, DRAG association to the chromatophore membrane. Understandably, the model here proposed for DRAT activity regulation is based on in vitro data and, while 2OG was shown to play a vital role, it cannot be excluded that other factors, such as the NAD+:NADH ratio and/or the uridylylation state of GlnB, may affect DRAT-GlnB interaction and DRAT activity in vivo. Nevertheless, it is now proven that DRAT interaction with GlnB greatly enhances DRAT activity. Binding of GlnB presumably

46

induces a conformational change on DRAT, increasing its affinity for Fe protein and/or β-NAD+. The previously reported DRAT variants K103E and N46D 207 were also assayed and further accentuate these effects. K103E, a constitutively active DRAT variant, interacts with GlnB in the same manner as wild-type (WT) DRAT, responding to variations in concentration of 2OG and the ADP:ATP ratio. However, this variant is active even in the presence of high concentrations of ATP and 2OG, when complex formation is disrupted. This protein has a lower Km for β-NAD+ (IV and 207) and, possibly, the conformational change induced by the substitution of a positively-charged amino acid (lysine) for a negatively charged residue (glutamate) increases the affinity for the substrate(s), obviating the need for an interaction with GlnB (IV). Conversely, the N46D variant was found to be less stable, on account of the fact that it has a lower affinity for GlnB. This protein also has no detectable activity in vitro, even though a very reduced level of activity (8%) has previously been reported in vivo 207. The substitution of residue 46 could cause a change in DRAT conformation, inhibiting its interaction with GlnB and, therefore, reducing affinity for its substrates (IV), once more demonstrating the need for DRAT-GlnB interaction for efficient DRAT ADP-ribosylation activity.

2. Regulation of energy switch-off

As previously mentioned, R. rubrum nitrogenase can also be modified in response to light withdrawal. This response, however, has different characteristics depending on the nitrogen source. Even though both glutamate-grown and N2-grown cultures will modify nitrogenase and inhibit nitrogenase activity in response to light withdrawal, the extent of inhibition is different between the two conditions, with N2-grown cultures retaining some nitrogenase activity when subjected to darkness, whereas glutamate-grown cells do not (V). As mentioned above, it was also recently shown that PII uridylylation patterns between glutamate- and N2-grown cultures are different in response to light withdrawal 148. As uridylylated GlnJ is unable to interact with AmtB1, this may also disrupt DRAG membrane association, freeing some of this protein in the cytosol where it can unmodify nitrogenase. As we have seen, membrane association of DRAG is highly dependent on AmtB1 during ammonium switch-off 144. In energy switch-off, however, AmtB1 seems to enhance but not to be required for efficient regulation of

47

DRAG membrane association in diazotrophically grown cells, thus suggesting the presence of a yet unknown membrane partner for DRAG 101. Simultaneously, the different modification patterns observed for PII proteins grown diazotrophically or with glutamate may also help explain why in AmtB1-deficient glutamate-grown R.

rubrum cultures energy switch-off is reduced to ca. 50% 144. It was not possible to assess the effect of the modification status of GlnB on DRAT-GlnB complex formation as the conditions required for GlnB modification also destabilize complex formation. DRAT is very unstable when not in complex with GlnB, forming precipitates when present by itself at the concentrations required for such studies and the in vitro modification reaction of GlnB will, therefore, also lead to DRAT precipitation. Even though using a partially modified GlnB fraction DRAT-GlnB complex formation and activity could be observed, it cannot be excluded that complete GlnB modification may affect DRAT activity. In diazotrophically-grown cells the PII proteins remain modified in response to light withdrawal and this may both destabilize DRAG membrane association, releasing some DRAG in the cytosol, at the same time that it could also affect DRAT activity. In this case, it is possible that simultaneously both DRAG and DRAT activities may be present, resulting in a fraction of the Fe protein remaining unmodified and giving measurable (albeit much lower) nitrogenase activity (V). In glutamate-grown cells, however, PII proteins are demodified in response to darkness, DRAG is found only by the membrane, DRAT is fully active and the Fe protein is completely modified, with no measurable nitrogenase activity (V). Addition of pyruvate to glutamate-grown cultures induces protein synthesis. Even though it was not possible to identify the newly synthesized protein, the outcome of pyruvate addition is the shift from a “glutamate-response” to an “N2-response”, inducing PII protein remodification after a period of 60 to 90 minutes. At the same time, and responding to this shift in PII modification, GS is also reactivated. Nitrogenase activity is regained, to the same level of N2-grown cells, with the Fe protein becoming unmodified (V). It can, therefore, be concluded that pyruvate addition will ultimately result in some DRAG release from the membrane, perhaps due to GlnJ modification, and that DRAT will become (at least partially) inactivated. Upon light withdrawal 2OG levels do decrease (from 5.34 ± 0.11 to 2.9 ± 0.23 nmol 2OG/mg protein – unpublished) in glutamate-grown cells and pyruvate addition allows them to recover to levels similar to those of cells in the light (6.49 ± 0.15 nmol 2OG/mg protein – unpublished).

48

However, all of these values are still higher than that found in diazotrophically-grown cells in the light (2.35 ± 0.34 nmol 2OG/mg protein 93). As such, variations in the intracellular 2OG pool, even though very important, are not the only factor regulating DRAT-GlnB interaction and DRAT activity. Pyruvate can be utilized by R. rubrum through different pathways and one of them includes the enzyme pyruvate formate-lyase (Pfl). This pathway (Figure 8), also present in other organisms, allows production of ATP by substrate-level phosphorylation in anaerobic conditions. In E. coli Pfl is induced upon transfer to low oxygen tensions, especially in the presence of pyruvate and expression of Pfl was shown to be dependent upon the FNR and ArcA/B redox sensors 208, 209.

Figure 8 – Pyruvate utilization pathway through Pfl; Fhl – Formate hydrogenlyase; Pta – phosphotransacetylase; Ack – Acetate kinase

R. rubrum was also shown to use this pathway when pyruvate was supplied as the carbon source under anaerobic conditions in the dark 210. In the case of diazotrophically-grown cells, light withdrawal naturally inhibits nitrogenase activity, as mentioned, but a low level of activity can still be measured immediately after transfer to the dark, with or without addition of pyruvate. However, glutamate-grown cells grown with malate as a carbon source cannot maintain measurable nitrogenase activity in the dark. Pyruvate addition, as we have discussed, induces protein synthesis and recovers nitrogenase activity to a low level under the same conditions. This newly synthesized protein(s) should already be present under diazotrophic conditions, so as to maintain anaerobic dark nitrogenase activity. Although it was not possible to assess the levels of Pfl in either condition, the enzymes in this pathway (or others regulating their activity) are likely candidates to be the proteins expressed upon pyruvate addition. Inhibition of this pathway by the addition of the specific inhibitor hypophosphite 211, 212

49

not only induces Fe protein modification, inhibiting nitrogenase activity, but it also induces demodification of the PII proteins, GS modification (V) and a slight decrease in the intracellular 2OG pool (from 6.49 ± 0.15 to 5.06 ± 0.32 nmol 2OG/mg protein – unpublished). These effects can be seen in both diazotrophically- and glutamate-grown cultures. Again, the PII modification status could account for a higher affinity of DRAG to the membrane, as under these conditions the AmtB1-GlnJ complex will once more be favoured. However, considering this pathway is also important for ATP generation, it could be that the ADP:ATP ratio, which could increase upon light withdrawal, would also play an important role on regulating DRAT-GlnB complex formation. Even though several studies have addressed this phenomenon 213-215, data on ADP:ATP ratios upon light withdrawal in diazotrophically-grown R. rubrum cultures remains scarce and sometimes contradictory. However, it seems likely that, upon transfer from light (with photosynthetical generation of ATP) to dark (with fermentative ATP production), there should be at least a transient shift in ADP:ATP ratios in the cell. This may also affect binding of 2OG by the PII proteins which, as we have discussed, requires the presence of ATP, thus influencing PII interactions with their cellular targets. Furthermore, it was shown in Rb. capsulatus that an unmodifiable GlnK variant still followed the same pattern of association to AmtB in response to switch-off effectors, further emphasizing that the binding of small effectors (ADP:ATP and 2OG) to the PII proteins, rather than modification status alone, is responsible for interaction regulation 216. Another factor to take into consideration is that the NAD(P)+: NAD(P)H ratios were also shown to influence nitrogenase switch-off, with addition of NAD+ inducing transient Fe protein modification and loss of nitrogenase activity 153 and, conversely, the addition of switch-off effectors led to the oxidation of the NAD(P) pool 152. It is not yet understood how a change in the NAD(P) pool is sensed by the DRAT/DRAG system but it could be that increased substrate availability (NAD+) plays a role in inducing DRAT activity. This could also be reflected on the redox state of the Fe protein itself, as it was shown that DRAT has a higher affinity for the oxidized form and DRAG prefers reduced Fe protein 133. A more oxidised pool of NAD(P) could then coincide with more oxidized Fe protein and, thus, stimulate DRAT activity, at the same time that DRAG activity would be inhibited. Although the full mechanism controlling DRAT and DRAG activity during energy switch-off in R. rubrum is still not fully understood, it could be that several or even all of the factors mentioned above, in combination, affect its regulation.

50

IV. Future prospects

The study of nitrogen fixation has lead to many important advances in recent years, with insights into enzyme structures (such as DRAG) and regulatory schemes and many important details regarding switch-off regulation now being known. However, as it usually happens, some of these answers only bring about new unanswered questions. The identity of the regulatory system linking redox sensing and differential protein expression in R. rubrum, for instance, is one of them. As we have discussed, this system should sense, directly or indirectly, changes in the cellular redox poise, perhaps by interaction with quinones in the chromatophore membrane, and it would be of great interest to identify its components and understand its regulation. Similarly, some of the aspects regarding regulation of DRAG membrane association still remain unresolved. Several different strategies were employed to try to identify the membrane interacting partner(s) for DRAG but so far this search has been unsuccessful, even though the AmtB1-GlnJ complex has been implicated at least in ammonium switch-off regulation. Whether such a binding partner actually exists is also still a matter of debate although, in addition to the factors described above, the fact that DRAG can also be found by the membrane in nitrogen rich conditions (when the expression of AmtB1 and GlnJ is negligible 101) gives some support to that theory. Much is still left to be understood of the mechanism regarding energy switch-off in R. rubrum. As this enzyme seems to be so important for maintenance of nitrogenase activity in dark anaerobic conditions, generating an R. rubrum strain lacking Pfl could be very informative. Also, development of a direct, accurate and real-time in vivo assay for estimation of the ADP:ATP ratios during energy (and even nitrogen) switch-off is highly desirable, to establish the importance of this factor in switch-off control. With the recent development of new in vivo sensors, such as the one using a modified version of a PII protein 91, this goal may now be closer at hand.

51

V. Concluding remarks

Efficient regulation of metabolic processes is vital for all living beings. In this thesis I have approached the strategies used by R. rubrum to adapt to different nitrogen availability conditions. As nitrogen is such an essential element for life, nitrogen fixing bacteria devote a large amount of their cellular resources to the maintenance and regulation of nitrogen fixation and assimilation. We have seen how induction of nitrogen fixation in R. rubrum represses carbon fixation and utilization pathways (I). I have also discussed how intracellular 2OG levels vary between nitrogen fixing and nitrogen rich conditions and how a manipulation of the intracellular 2OG pool will affect the co-regulation of both nitrogen and carbon fixation pathways (II). As the 2OG concentration affects PII uridylylation, the processes in which PII protein uridylylation plays an important role, such as NifA activation and GS regulation, are also affected by changes in the 2OG pool (II). Furthermore, the interconnection between carbon and nitrogen fixation pathways suggests that there is a sensor/transducer system, capable of responding to changes in the cellular redox poise, as nitrogen fixation and carbon fixation are both processes that require substantial reductive power. As the chromatophore membrane plays a vital role in many cellular processes, including nitrogen fixation, a Blue Native gel system using a new class of amphiphilic detergents was developed, in combination with LTQ-Orbitrap mass spectrometry. This system allowed identification of many chromatophore membrane protein complex components and showed that the Sdh complex is up regulated in nitrogen fixing conditions (III), providing further strength to the concept of a redox-sensing regulator. As nitrogen fixation requires large amounts of reducing equivalents, this can also be seen as evidence supporting the importance of the TCA cycle for nitrogenase activity and, perhaps, also supporting the reversed electron flow mechanism in R. rubrum as a mean to generate NADH and reoxidise the quinone pool. I have also described for the first time how DRAT and the PII protein GlnB from R. rubrum form not only a complex but a complex with measurable DRAT activity. Complex formation, and consequently DRAT activity, is regulated by binding of the small

52

effectors ADP/ATP and 2OG, presumably to the PII protein, with lower ADP:ATP ratios and the presence of 2OG destabilizing the complex and inactivating DRAT (IV). As this will undoubtedly provide a clear insight into the regulatory mechanism behind DRAT-GlnB complex formation, a project is also under way to obtain the atomic structure of this complex. While the binding of 2OG could explain regulation of DRAT activity in nitrogen switch-off, the regulation of energy switch-off is still unclear. Different nitrogen sources influence the response to light withdrawal in different manners, with diazotrophically-grown cells maintaining some activity in the dark, unlike glutamate-grown cells. Addition of pyruvate to glutamate cultures induces protein synthesis and allows recovery of partial nitrogenase activity in the dark. The work presented in this thesis also shows how the activity of the pathway starting in Pfl is vital for dark anaerobic nitrogenase activity and that the PII proteins, nitrogenase and GS modification status respond to pyruvate addition and Pfl pathway inhibition (V). It could be that this pathway is already present in diazotrophically-grown cells, providing some of the ATP needed to maintain a certain “threshold” of the ADP:ATP ratio, which maintains PII modification and affects DRAT/DRAG regulation. Further studies are still required in order to fully understand the mechanisms regulating energy switch-off and the influence of not only ADP:ATP but also of NAD(P)H:NAD(P)+ ratios for DRAT and DRAG activity regulation. Even though many new aspects have now come to light, the regulation of nitrogen fixation in R. rubrum still has countless interesting questions, waiting to be answered.

53

VI. Sumário em português

A adaptabilidade é uma das chaves do segredo das diferentes formas de vida presentes no nosso Planeta. De entre todos os diferentes grupos de organismos, as bactérias são, indubitavelmente, os de maior sucesso, sendo capazes de se adaptar às mais diversas condições. São também bactérias as responsáveis por diversos processos vitais para a manutenção da vida na Terra, como seja a fixação de azoto. Este elemento constitui cerca de 80% da atmosfera, sob a forma de N2 gasoso. Contudo, e apesar desta extrema abundância, a disponibilidade deste elemento é muitas vezes limitada. O processo de fixação de azoto, no qual o azoto atmosférico (indisponível para a maioria dos organismos) é convertido em amónia (uma forma que pode ser utilizada pela biosfera para síntese de moléculas vitais como aminoácidos, DNA ou RNA, entre outros) é pois de importância vital. A enzima responsável por este processo, nitrogenase, é extremamente sensível ao oxigénio e, dado que a sua reacção requer quantidades consideráveis de energia (sob a forma de ATP e equivalentes redutores), está sujeita a variados mecanismos de regulação, tanto pré- como pós-tradução. Modificação de uma subunidade da nitrogenase inibe a reacção de fixação de azoto, permitindo a regulação da actividade de forma instantânea. Nesta tese são descritos os mecanismos de controlo presentes na bactéria fotossintética e fixadora de azoto Rhodospirillum rubrum, presente em sedimentos de rios e lagos por todo mundo. Através de técnicas de proteómica, algumas desenvolvidas especificamente para este trabalho, foi possível demonstrar que os processos de fixação de azoto e fixação de carbono são co-regulados, possivelmente de maneira a manter o equilíbrio redox no interior da célula. Para além da regulação do proteoma, a regulação da actividade da nitrogenase por modificação pós-tradução é também objecto de estudo e, pela primeira vez, é demonstrado que a enzima responsável pela modificação da nitrogenase (DRAT) é regulada através da interacção com uma proteína pertencente à família de proteínas PII – GlnB. A formação deste complexo, a forma activa da enzima, é regulada pela razão entre as concentrações de ADP e ATP e 2-oxoglutarato. Estes resultados permitem a elaboração de um modelo de regulação pós-tradução da nitrogenase, em resposta às concentrações de amónio.

54

A enzima nitrogenase pode ser modificada tanto em resposta a um aumento das concentrações de amónio (dado que nesse caso não é necessário haver fixação de azoto) como por ausência de luz. A resposta à escuridão é também descrita nesta tese e é demonstrada pela primeira vez que a actividade da enzima piruvato formato-liase é essencial para manutenção da actividade da nitrogenase na ausência de luz e oxigénio. Em conclusão, os resultados obtidos com este estudo permitem a elaboração de um modelo de regulação do processo de fixação de azoto em R. rubrum, aumentando o nosso conhecimento acerca deste processo crucial para a vida na Terra.

55

VII. Acknowledgements

“How many roads must a man walk down

Before you call him a man?”

Bob Dylan, in “Blowin’ in the wind”

Dear reader, welcome to the most popular pages in the thesis!!! Please do take the time to peruse the previous 50 or so pages though… I did my best to write an interesting text! :o) It is no secret to anyone that the road to those three letters (PhD) is long, tricky, frustrating and many times can lead to “Insanityville”… but it is also immensely rewarding, mostly for the people one meets along the way. For reasons of space, I won’t be able to thank everyone I would like but, don’t worry, I did not forget about you…! :o) First and foremost, I thank my supervisors, Agneta Norén and Stefan Nordlund. You accepted me into the “Fixer family” without even knowing me, backed me up 200%, through thick and thin. You put up with my crazy ideas and were always there for me, with good advice, a few (needed, I know…!) “slaps in the hand” and all the help I could have ever asked for, both in and out of the lab. Thank you for “keeping my eyes on the ball” and helping me grow, both as a scientist and (I hope!) as a person. None of my words will ever repay you enough… but I still had to give it a shot! :o) Second, to my good friend, companion and “brother-in-arms”, Dr. (ah pois!!!) Pedro Teixeira. Tive a honra e felicidade de partilhar esta última década contigo, os últimos anos aqui pelo Norte – aliás, não fosses tu provavelmente nunca teria sequer ido para a Suécia... É bom ter amigos, mas melhor ainda é ter bons amigos – e felizmente que te posso contar nesse grupo! Não sei como raio fazes para saber tanta coisa... mas sem dúvida que mais que uma vez foi bom poder contar contigo, tanto para saber qualquer coisa no laboratório como também quem marcou o golo vencedor no Alguidares de Baixo F. C. vs U. D. R. Freixo de Espada à Cinta!! :oD Aqui fica, portanto, o meu abraço, por todos os bons momentos que passámos e por todos os que hão-de vir! Votos de sucesso em tudo, por toda a vida! E biba o F.C.P.!

56

To my co-workers, lab-mates and friends, Tomas Edgren, Wang He and Anders Jonsson – thanks guys! Tomas, I will always admire your positive attitude towards life and your apparently never-ending knowledge. Oh, and thanks for teaching me about beer!!! :oD Helen – ah, what can I say? Life just isn’t the same around here without you! Thanks for all the things you taught me, all the good advice, about science and other things – ;o) – all the interesting experiences and, most of all, for being a good friend. Take care of the Big Guy – and of the Small Guy/Girl too, when he/she shows up!!! And, of course, AJ!!! Thanks for your good humour and for all the stories you had for us. I have to say I’d never met anyone with a life quite as remarkable as yours! I wish you all the best! Não esquecer o resto da “Portuguese Mafia”, claro! Às nossas meninas Catarina e Salomé, com quem tive o prazer de compartilhar tantos e tão bons momentos por aqui e com quem formámos o “terrível” quarteto de Portugas, sempre pronto a cortar na casaca, fazer uma jantarada ou, simplesmente, galhofar um bocado... Muito boa sorte, não tarda muito e são vocês!! Since we’re speaking of “Portuguese Mafia”, I’d also like to extend that “honour” to our “adopted” members, Pilar, Beata, Changrong, Dimitra, Candan and Erdem. What a good time we’ve had together!! I’ll miss you guys… Also, a word of gratitude to my previous office mates Alex G., Hanna G., Hanna E.,, Kajsa and Anna W., for the great atmosphere we had in our little corner. All the best to each and every one of you! A special word of thanks to my good friends “downstairs” at the JWdG lab, especially to Mirjam K.. I am so happy to have met you, thank you so much for being there for me! Your friendly smile is something I have learned to appreciate and I can tell you that it’ll be sorely missed! I wish you nothing but the very best! To my friends from “across the yard”, Anya (Прывітаньне!!), Linnea (do the fish face again!!!) and Minttu (Hauska tavata!), thanks for all the good memories you gave me, it has been a pleasure! Also, to everyone at Martin Högbom’s lab, especially to Martin, whose enthusiasm for science is truly contagious, and Mike

and Charlotta, for all the help with DRAT, a big, big thank you! To our “cousins” from Botan, especially my friend and co-author Simina and to the “RuBisCO subunits” Irina, Nastia and Nok. Simina, I wish you and Jack all the best in the States! Большое спасибо, Иpa и Настя, my life definitely became more interesting after meeting you. I will never forget you! :o) To my students, Timo Schmidt and Veronika Henriksson, for the great times we had together and for all I learned from you. Good luck to you both! Aproveito também para deixar um grande abraço para dois grandes amigos (e parceiros de badminton!), Rui e Erika. Obrigado pelo vosso apoio, bons conselhos de “quem já passou por isso” e, especialmente ao Rui, pela ENORME ajuda! Igualmente aqui fica um beijinho para a minha amiga Nela, a quem desejo toda a

57

sorte do mundo e um valente abraço para o “xôtor” Nuno M., de quem tenho a sorte de ser amigo, mesmo depois de todos estes anos e de toda a distância! Likewise, a word of appreciation to those that followed my path from far away and always had a friendly word of encouragement. Köszkönöm szépen Zsuzsa és Zoltán! A mes très bons amis Daniel et Thérèse Ganivet, un énorme merci beaucoup! Et on continue en français pour remercier aussi à Claire, Arezki et Bertrand! Y para ti también, Ana, un gran besote - y ahora yo también soy Dotor !!! Y quizás... :o) To Bogos, for always having a friendly word for us, for always remembering the latest football news and for all the help throughout the years, thank you very much! Också, tack till Ann, Maria, Lotta och Lollo som gör att DBB alltid ”lever”. Ett speciellt tack till Eddie och Håkan för alla hjälpen med diverse (icke fungerande...) prylar! To all of you that make DBB such a wonderful place to work in - I’ll miss you all! Thanks also to Fundação para a Ciência e a Tecnologia for the finantial support. Um grande abraço a toda a minha “família adoptiva” – a família Teixeira – por todo o apoio e amizade ao longo dos anos. Finalmente, o maior agradecimento para a minha família, por todo o apoio, carinho, dedicação e ajuda ao longo deste longo percurso, especialmente aos meus pais, José e Clara Selão, às minhas avós, Alcinda Selão e Maria Madalena

Marecos, que sempre se lembram do seu neto que nunca pára em casa, ao meu tio António José Espadinha do Monte, sempre interessado em saber notícias das minhas amigas bactérias, e ao meu irmão Pedro Selão – muito obrigado por estarem aí por mim. Sem vocês nunca teria conseguido resistir, este livro não é só meu, é de todos nós no “muy grande e nobre clã Selão”! ;o)

“Caminante, son tus huellas

el camino y nada más;

caminante, no hay camino,

se hace camino al andar.

[...]

Golpe a golpe, verso a verso.“

Antonio Machado, in “Cantares”

58

VIII. References

1. Zimmer, C., "Origins. On the origin of eukaryotes". Science, 2009. 325: pp. 666-8.

2. Anderson, A.W., et al., "Studies on a radio-resistant micrococcus. I. Isola-tion, morphology, cultural characteristics, and resistance to gamma radia-tion". Food Technol., 1956. 10: pp. 575-7.

3. Whitman, W.B., Coleman, D.C., and Wiebe, W.J., "Prokaryotes: the un-seen majority". Proc Natl Acad Sci U S A, 1998. 95: pp. 6578-83.

4. Nicholson, J.K., et al., "The challenges of modeling mammalian biocom-plexity". Nat Biotechnol, 2004. 22: pp. 1268-74.

5. Lavoisier, A.L. and Kerr, R., "Elements of chemistry: in a new systematic order, containing all the modern discoveries, illustrated with thirteen cop-perplates, by Mr. Lavoisier, translated from the French by Robert Kerr". Early American imprints. Second series ; no. 799. 1801, New York :: Printed for T. & J. Swords, Medical Booksellers, and Printer to the Faculty of Physic of Columbia College.

6. Greenwood, N.N. and Earnshaw, A., "Chemistry of the Elements". 2nd edition ed. 1997, Oxford, United Kingdom: Butterworth-Heinemann.

7. Masson-Boivin, C., et al., "Establishing nitrogen-fixing symbiosis with legumes: how many rhizobium recipes?". Trends Microbiol, 2009. 17: pp. 458-66.

8. James, E.K., et al., "Further observations on the interaction between sugar cane and Gluconacetobacter diazotrophicus under laboratory and green-house conditions". J Exp Bot, 2001. 52: pp. 747-60.

9. Fu, H. and Burris, R.H., "Ammonium inhibition of nitrogenase activity in Herbaspirillum seropedicae". J Bacteriol, 1989. 171: pp. 3168-75.

10. Hurek, T., et al., "Augmented rates of respiration and efficient nitrogen fixation at nanomolar concentrations of dissolved O2 in hyperinduced Azoarcus sp. strain BH72". J Bacteriol, 1994. 176: pp. 4726-33.

11. Hartmann, A., Fu, H.A., and Burris, R.H., "Influence of amino acids on nitrogen fixation ability and growth of Azospirillum spp". Appl Environ

Microbiol, 1988. 54: pp. 87-93. 12. Winogradsky, S., "Recherches sur l’assimilation de l’azote libre de

l’atmosphere par les microbes". Arch Sci Biol, 1895. 3: pp. 297-352. 13. Schindelin, H., et al., "Structure of ADP.AIF4

- - stabilized nitrogenase com-plex and its implications for signal transduction". Nature, 1997. 387: pp. 370-6.

14. Gest, H., "A serendipic legacy: Erwin Esmarch's isolation of the first pho-tosynthetic bacterium in pure culture". Photosynthesis Research, 1995. 46: pp. 473-8.

59

15. Lundin, A., Thore, A., and Baltscheffsky, M., "Sensitive measurement of flash induced photophosphorylation in bacterial chromatophores by firefly luciferase". FEBS Lett, 1977. 79: pp. 73-6.

16. Johansson, B.C. and Baltscheffsky, M., "Subunit composition of coupling factor (ATPase) from Rhodospirillum rubrum". FEBS Lett, 1975. 53: pp. 221-4.

17. Matsuda, H., et al., "Reconstruction of photosynthetic, cyclic electron-transport system from photoreaction unit, ubiquinone-10 protein, cyto-chrome c2 and polar lipids purified from Rhodospirillum rubrum". J Bio-

chem, 1984. 95: pp. 431-42. 18. del Valle-Tascon, S., van Grondelle, R., and Duysens, L.N., "Flash-induced

photophosphorylation in Rhodospirillum rubrum chromatophores. I. The relationship between cytochrome c420 content and photophosphorylation". Biochim Biophys Acta, 1978. 504: pp. 26-39.

19. Frenkel, A.W., "Simultaneous reduction of diphosphopyridine nucleotide and oxidation of reduced flavin mononucleotide by illuminated bacterial chromatophores". J Am Chem Soc, 1958. 80: pp. 3479-80.

20. Frenkel, A.W., "Light-induced phosphorylation by cell-free preparations of photosynthetic bacteria". J Am Chem Soc, 1954. 76: pp. 5568-9.

21. Nore, B.F., "∆pH driven energy-linked NAD+ reduction in Rhodospirillum

rubrum chromatophores". Arch Biochem Biophys, 1989. 274: pp. 285-9. 22. Kamen, M.D. and Gest, H., "Evidence for a Nitrogenase System in the

Photosynthetic Bacterium Rhodospirillum rubrum". Science, 1949. 109: pp. 560.

23. Gest, H. and Kamen, M.D., "Photoproduction of Molecular Hydrogen by Rhodospirillum rubrum". Science, 1949. 109: pp. 558-9.

24. Ludden, P.W. and Burris, R.H., "Removal of an adenine-like molecule during activation of dinitrogenase reductase from Rhodospirillum rubrum". Proc Natl Acad Sci U S A, 1979. 76: pp. 6201-5.

25. Nordlund, S., Eriksson, U., and Baltscheffsky, H., "Properties of the nitro-genase system from a photosynthetic bacterium, Rhodospirillum rubrum". Biochim Biophys Acta, 1978. 504: pp. 248-54.

26. Nordlund, S. and Ludden, P.W., "Incorporation of adenine into the modify-ing group of inactive iron protein of nitrogenase from Rhodospirillum ru-

brum". Biochem J, 1983. 209: pp. 881-4. 27. Gest, H., "Metabolic patterns in photosynthetic bacteria". Bacteriol Rev,

1951. 15: pp. 183-210. 28. Carnahan, J.E., et al., "Nitrogen fixation in cell-free extracts of Clostridium

pasteurianum". Biochim Biophys Acta, 1960. 44: pp. 520-35. 29. Broda, E. and Peschek, G.A., "Nitrogen fixation as evidence for the reduc-

ing nature of the early biosphere". Biosystems, 1983. 16: pp. 1-8. 30. Seefeldt, L.C., Hoffman, B.M., and Dean, D.R., "Mechanism of Mo-

dependent nitrogenase". Annu Rev Biochem, 2009. 78: pp. 701-22. 31. Igarashi, R.Y. and Seefeldt, L.C., "Nitrogen fixation: the mechanism of the

Mo-dependent nitrogenase". Crit Rev Biochem Mol Biol, 2003. 38: pp. 351-84.

32. Raymond, J., et al., "The natural history of nitrogen fixation". Mol Biol

Evol, 2004. 21: pp. 541-54.

60

33. Lehman, L.J. and Roberts, G.P., "Identification of an alternative nitro-genase system in Rhodospirillum rubrum". J Bacteriol, 1991. 173: pp. 5705-11.

34. Davis, R., et al., "Purification and characterization of the alternative nitro-genase from the photosynthetic bacterium Rhodospirillum rubrum". J Bac-

teriol, 1996. 178: pp. 1445-50. 35. Anbar, A.D. and Knoll, A.H., "Proterozoic ocean chemistry and evolution:

a bioinorganic bridge?". Science, 2002. 297: pp. 1137-42. 36. Ribbe, M., Gadkari, D., and Meyer, O., "N2 fixation by Streptomyces ther-

moautotrophicus involves a molybdenum-dinitrogenase and a manganese-superoxide oxidoreductase that couple N2 reduction to the oxidation of su-peroxide produced from O2 by a molybdenum-CO dehydrogenase". J Biol

Chem, 1997. 272: pp. 26627-33. 37. Bulen, W.A. and LeComte, J.R., "The nitrogenase system from Azotobac-

ter: two-enzyme requirement for N2 reduction, ATP-dependent H2 evolu-tion and ATP hydrolysis". Proc Natl Acad Sci U S A, 1966. 56: pp. 979-86.

38. Mortenson, L.E., Morris, J.A., and Jeng, D.Y., "Purification, metal compo-sition and properties of molybdoferredoxin and azoferredoxin, two of the components of the nitrogen-fixing system of Clostridium pasteurianum". Biochim Biophys Acta, 1967. 141: pp. 516-22.

39. Howard, J.B. and Rees, D.C., "Nitrogenase: a nucleotide-dependent mo-lecular switch". Annu Rev Biochem, 1994. 63: pp. 235-64.

40. Peters, J.W., Fisher, K., and Dean, D.R., "Nitrogenase structure and func-tion: a biochemical-genetic perspective". Annu Rev Microbiol, 1995. 49: pp. 335-66.

41. Rubio, L.M. and Ludden, P.W., "Biosynthesis of the iron-molybdenum cofactor of nitrogenase". Annu Rev Microbiol, 2008. 62: pp. 93-111.

42. Shah, V.K. and Brill, W.J., "Isolation of an iron-molybdenum cofactor from nitrogenase". Proc Natl Acad Sci U S A, 1977. 74: pp. 3249-53.

43. Kim, J. and Rees, D.C., "Structural models for the metal centers in the nitrogenase molybdenum-iron protein". Science, 1992. 257: pp. 1677-82.

44. Chan, M.K., Kim, J., and Rees, D.C., "The nitrogenase FeMo-cofactor and P-cluster pair: 2.2 Å resolution structures". Science, 1993. 260: pp. 792-4.

45. Georgiadis, M.M., et al., "Crystallographic structure of the nitrogenase iron protein from Azotobacter vinelandii". Science, 1992. 257: pp. 1653-9.

46. Oppenheim, J., et al., "Properties of a soluble nitrogenase in Azotobacter". J Bacteriol, 1970. 101: pp. 292-6.

47. Hu, Y., et al., "Assembly of nitrogenase MoFe protein". Biochemistry, 2008. 47: pp. 3973-81.

48. Thorneley, R.N. and Lowe, D.J., "The mechanism of Klebsiella pneumo-

niae nitrogenase action. Simulation of the dependences of H2-evolution rate on component-protein concentration and ratio and sodium dithionite con-centration". Biochem J, 1984. 224: pp. 903-9.

49. Lowe, D.J. and Thorneley, R.N., "The mechanism of Klebsiella pneumo-

niae nitrogenase action. The determination of rate constants required for the simulation of the kinetics of N2 reduction and H2 evolution". Biochem J, 1984. 224: pp. 895-901.

50. Thorneley, R.N. and Lowe, D.J., "The mechanism of Klebsiella pneumo-

niae nitrogenase action. Pre-steady-state kinetics of an enzyme-bound in-

61

termediate in N2 reduction and of NH3 formation". Biochem J, 1984. 224: pp. 887-94.

51. Lowe, D.J. and Thorneley, R.N., "The mechanism of Klebsiella pneumo-

niae nitrogenase action. Pre-steady-state kinetics of H2 formation". Bio-

chem J, 1984. 224: pp. 877-86. 52. Simpson, F.B. and Burris, R.H., "A nitrogen pressure of 50 atmospheres

does not prevent evolution of hydrogen by nitrogenase". Science, 1984. 224: pp. 1095-7.

53. Dilworth, M.J., "Acetylene reduction by nitrogen-fixing preparations from Clostridium pasteurianum". Biochim Biophys Acta, 1966. 127: pp. 285-94.

54. Dixon, R. and Kahn, D., "Genetic regulation of biological nitrogen fixa-tion". Nat Rev Microbiol, 2004. 2: pp. 621-31.

55. Burgess, B.K. and Lowe, D.J., "Mechanism of molybdenum nitrogenase". Chem Rev, 1996. 96: pp. 2983-3012.

56. Edgren, T. and Nordlund, S., "The fixABCX genes in Rhodospirillum ru-

brum encode a putative membrane complex participating in electron trans-fer to nitrogenase". J Bacteriol, 2004. 186: pp. 2052-60.

57. Edgren, T. and Nordlund, S., "Two pathways of electron transport to nitro-genase in Rhodospirillum rubrum: the major pathway is dependent on the fix gene products". FEMS Microbiol Lett, 2006. 260: pp. 30-5.

58. Tsai, M.H. and Saier, M.H., Jr., "Phylogenetic characterization of the ubiq-uitous electron transfer flavoprotein families ETF-alpha and ETF-beta". Res Microbiol, 1995. 146: pp. 397-404.

59. Weidenhaupt, M., et al., "Bradyrhizobium japonicum possesses two dis-crete sets of electron transfer flavoprotein genes: fixA, fixB and etfS, etfL". Arch Microbiol, 1996. 165: pp. 169-78.

60. Edgren, T., "Electron transport to nitrogenase in Rhodospirillum rubrum", PhD thesis, Department of Biochemistry and Biophysics, Stockholm Uni-versity, Stockholm, 2006

61. Lindblad, A. and Nordlund, S., "Electron transport to nitrogenase in Rhodospirillum rubrum: role of energization of the chromatophore mem-brane". Photosynth Res, 1997. 53: pp. 23-8.

62. Haaker, H., et al., "Short-term regulation of the nitrogenase activity in Rhodopseudomonas sphaeroides". Eur J Biochem, 1982. 127: pp. 639-45.

63. Meyer, J., Kelley, B.C., and Vignais, P.M., "Effect of light nitrogenase function and synthesis in Rhodopseudomonas capsulata". J Bacteriol, 1978. 136: pp. 201-8.

64. Dupuis, A., et al., "The complex I from Rhodobacter capsulatus". Biochim

Biophys Acta, 1998. 1364: pp. 147-65. 65. Brostedt, E. and Nordlund, S., "Purification and partial characterization of a

pyruvate oxidoreductase from the photosynthetic bacterium Rhodospirillum

rubrum grown under nitrogen-fixing conditions". Biochem J, 1991. 279: pp. 155-8.

66. Lindblad, A., et al., "Identification and sequence of a nifJ-like gene in Rhodospirillum rubrum: partial characterization of a mutant unaffected in nitrogen fixation". Mol Microbiol, 1996. 20: pp. 559-68.

67. Edgren, T. and Nordlund, S., "Electron transport to nitrogenase in Rhodospirillum rubrum: identification of a new fdxN gene encoding the

62

primary electron donor to nitrogenase". FEMS Microbiol Lett, 2005. 245: pp. 345-51.

68. Nagatani, H., Shimizu, M., and Valentine, R.C., "The mechanism of am-monia assimilation in nitrogen fixing bacteria". Arch Mikrobiol, 1971. 79: pp. 164-75.

69. Reitzer, L., "Nitrogen assimilation and global regulation in Escherichia

coli". Annu Rev Microbiol, 2003. 57: pp. 155-76. 70. Helling, R.B., "Why does Escherichia coli have two primary pathways for

synthesis of glutamate?". J Bacteriol, 1994. 176: pp. 4664-8. 71. Helling, R.B., "Pathway choice in glutamate synthesis in Escherichia coli".

J Bacteriol, 1998. 180: pp. 4571-5. 72. Miller, R.E. and Stadtman, E.R., "Glutamate synthase from Escherichia

coli. An iron-sulfide flavoprotein". J Biol Chem, 1972. 247: pp. 7407-19. 73. Rhee, S.G., Chock, P.B., and Stadtman, E.R., "Regulation of Escherichia

coli glutamine synthetase". Adv Enzymol Relat Areas Mol Biol, 1989. 62: pp. 37-92.

74. Soliman, A. and Nordlund, S., "Purification and partial characterization of glutamine synthetase from the photosynthetic bacterium Rhodospirillum

rubrum". Biochim Biophys Acta, 1989. 994: pp. 138-41. 75. Carlberg, I. and Nordlund, S., "Purification and partial characterization of

glutamate synthase from Rhodospirillum rubrum grown under nitrogen-fixing conditions". Biochem J, 1991. 279: pp. 151-4.

76. Shapiro, B.M., "The glutamine synthetase deadenylylating enzyme system from Escherichia coli. Resolution into two components, specific nucleotide stimulation, and cofactor requirements". Biochemistry, 1969. 8: pp. 659-70.

77. Foor, F., Reuveny, Z., and Magasanik, B., "Regulation of the synthesis of glutamine synthetase by the PII protein in Klebsiella aerogenes". Proc Natl

Acad Sci U S A, 1980. 77: pp. 2636-40. 78. Arcondeguy, T., Jack, R., and Merrick, M., "P(II) signal transduction pro-

teins, pivotal players in microbial nitrogen control". Microbiol Mol Biol

Rev, 2001. 65: pp. 80-105. 79. Xu, Y., et al., "GlnK, a PII-homologue: structure reveals ATP binding site

and indicates how the T-loops may be involved in molecular recognition". J

Mol Biol, 1998. 282: pp. 149-65. 80. Sakai, H., et al., "Crystal structures of the signal transducing protein GlnK

from Thermus thermophilus HB8". J Struct Biol, 2005. 149: pp. 99-110. 81. van Heeswijk, W.C., et al., "An alternative PII protein in the regulation of

glutamine synthetase in Escherichia coli". Mol Microbiol, 1996. 21: pp. 133-46.

82. Thomas, G., Coutts, G., and Merrick, M., "The glnKamtB operon. A con-served gene pair in prokaryotes". Trends Genet, 2000. 16: pp. 11-4.

83. Leigh, J.A. and Dodsworth, J.A., "Nitrogen regulation in bacteria and ar-chaea". Annu Rev Microbiol, 2007. 61: pp. 349-77.

84. Dodsworth, J.A. and Leigh, J.A., "Regulation of nitrogenase by 2-oxoglutarate-reversible, direct binding of a PII-like nitrogen sensor protein to dinitrogenase". Proc Natl Acad Sci U S A, 2006. 103: pp. 9779-84.

85. Jiang, P. and Ninfa, A.J., "Escherichia coli PII signal transduction protein controlling nitrogen assimilation acts as a sensor of adenylate energy charge in vitro". Biochemistry, 2007. 46: pp. 12979-96.

63

86. Kamberov, E.S., Atkinson, M.R., and Ninfa, A.J., "The Escherichia coli PII signal transduction protein is activated upon binding 2-ketoglutarate and ATP". J Biol Chem, 1995. 270: pp. 17797-807.

87. Forchhammer, K., "P(II) signal transducers: novel functional and structural insights". Trends Microbiol, 2008. 16: pp. 65-72.

88. Ninfa, A.J. and Jiang, P., "PII signal transduction proteins: sensors of al-pha-ketoglutarate that regulate nitrogen metabolism". Curr Opin Microbiol, 2005. 8: pp. 168-73.

89. Yildiz, O., et al., "Structure of GlnK1 with bound effectors indicates regu-latory mechanism for ammonia uptake". EMBO J, 2007. 26: pp. 589-99.

90. Truan, D., et al., "A new PII protein structure identifies the 2-oxoglutarate binding site". J Mol Biol, 2010. 400: pp. 531-9.

91. Berg, J., Hung, Y.P., and Yellen, G., "A genetically encoded fluorescent reporter of ATP:ADP ratio". Nat Methods, 2009. 6: pp. 161-6.

92. Senior, P.J., "Regulation of nitrogen metabolism in Escherichia coli and Klebsiella aerogenes: studies with the continuous-culture technique". J

Bacteriol, 1975. 123: pp. 407-18. 93. Jonsson, A., Teixeira, P.F., and Nordlund, S., "The activity of adenylyl-

transferase in Rhodospirillum rubrum is only affected by alpha-ketoglutarate and unmodified PII proteins, but not by glutamine, in vitro". FEBS J, 2007. 274: pp. 2449-60.

94. Jaggi, R., et al., "The role of the T-loop of the signal transducing protein PII from Escherichia coli". FEBS Lett, 1996. 391: pp. 223-8.

95. Son, H.S. and Rhee, S.G., "Cascade control of Escherichia coli glutamine synthetase. Purification and properties of PII protein and nucleotide se-quence of its structural gene". J Biol Chem, 1987. 262: pp. 8690-5.

96. Forchhammer, K. and Tandeau de Marsac, N., "The PII protein in the cyanobacterium Synechococcus sp. strain PCC 7942 is modified by serine phosphorylation and signals the cellular N-status". J Bacteriol, 1994. 176: pp. 84-91.

97. Strosser, J., et al., "Regulation of GlnK activity: modification, membrane sequestration and proteolysis as regulatory principles in the network of ni-trogen control in Corynebacterium glutamicum". Mol Microbiol, 2004. 54: pp. 132-47.

98. Jiang, P. and Ninfa, A.J., "Sensation and signaling of alpha-ketoglutarate and adenylylate energy charge by the Escherichia coli PII signal transduc-tion protein require cooperation of the three ligand-binding sites within the PII trimer". Biochemistry, 2009. 48: pp. 11522-31.

99. Zhang, Y., et al., "Functional characterization of three GlnB homologs in the photosynthetic bacterium Rhodospirillum rubrum: roles in sensing am-monium and energy status". J Bacteriol, 2001. 183: pp. 6159-68.

100. Cheng, J., Johansson, M., and Nordlund, S., "Expression of PII and gluta-mine synthetase is regulated by PII, the ntrBC products, and processing of the glnBA mRNA in Rhodospirillum rubrum". J Bacteriol, 1999. 181: pp. 6530-4.

101. Wang, H., "Metabolic regulation of nitrogen fixation in Rhodospirillum

rubrum", PhD thesis, Department of Biochemistry and Biophysics, Stock-holm University, Stockholm, 2009

64

102. Jonsson, A. and Nordlund, S., "In vitro studies of the uridylylation of the three PII protein paralogs from Rhodospirillum rubrum: the transferase ac-tivity of R. rubrum GlnD is regulated by alpha-ketoglutarate and divalent cations but not by glutamine". J Bacteriol, 2007. 189: pp. 3471-8.

103. Teixeira, P.F., et al., "Interaction of the signal transduction protein GlnJ with the cellular targets AmtB1, GlnE and GlnD in Rhodospirillum rubrum: dependence on manganese, 2-oxoglutarate and the ADP/ATP ratio". Mi-

crobiology, 2008. 154: pp. 2336-47. 104. Hill, S., et al., "Nitrogen fixation gene (nifL) involved in oxygen regulation

of nitrogenase synthesis in K. pneumoniae". Nature, 1981. 290: pp. 424-6. 105. Ho, Y.S., Burden, L.M., and Hurley, J.H., "Structure of the GAF domain, a

ubiquitous signaling motif and a new class of cyclic GMP receptor". EMBO

J, 2000. 19: pp. 5288-99. 106. Little, R. and Dixon, R., "The amino-terminal GAF domain of Azotobacter

vinelandii NifA binds 2-oxoglutarate to resist inhibition by NifL under ni-trogen-limiting conditions". J Biol Chem, 2003. 278: pp. 28711-8.

107. Jiang, P. and Ninfa, A.J., "Regulation of autophosphorylation of Es-

cherichia coli nitrogen regulator II by the PII signal transduction protein". J

Bacteriol, 1999. 181: pp. 1906-11. 108. Zhang, Y., et al., "Mutagenesis and functional characterization of the glnB,

glnA, and nifA genes from the photosynthetic bacterium Rhodospirillum

rubrum". J Bacteriol, 2000. 182: pp. 983-92. 109. Zhang, Y., et al., "Effect of an ntrBC mutation on the posttranslational

regulation of nitrogenase activity in Rhodospirillum rubrum". J Bacteriol, 1995. 177: pp. 5322-6.

110. Zhang, Y., Pohlmann, E.L., and Roberts, G.P., "GlnD is essential for NifA activation, NtrB/NtrC-regulated gene expression, and posttranslational regulation of nitrogenase activity in the photosynthetic, nitrogen-fixing bac-terium Rhodospirillum rubrum". J Bacteriol, 2005. 187: pp. 1254-65.

111. Wu, J. and Bauer, C.E., "RegB/RegA, a global redox-responding two-component system". Adv Exp Med Biol, 2008. 631: pp. 131-48.

112. Joshi, H.M. and Tabita, F.R., "A global two component signal transduction system that integrates the control of photosynthesis, carbon dioxide assimi-lation, and nitrogen fixation". Proc Natl Acad Sci U S A, 1996. 93: pp. 14515-20.

113. Gest, H., Kamen, M.D., and Bregoff, H.M., "Studies on the metabolism of photosynthetic bacteria .5. Photoproduction of hydrogen and nitrogen fixa-tion by Rhodospirillum rubrum". J Biol Chem, 1950. 182: pp. 153-70.

114. Schick, H.J., "Regulation of photoreduction in Rhodospirillum rubrum by ammonia". Arch Mikrobiol, 1971. 75: pp. 110-20.

115. Neilson, A.H. and Nordlund, S., "Regulation of nitrogenase synthesis in intact cells of Rhodospirillum rubrum: inactivation of nitrogen fixation by ammonia, L-glutamine and L-asparagine". J Gen Microbiol, 1975. 91: pp. 53-62.

116. Kanemoto, R.H. and Ludden, P.W., "Amino acid concentrations in Rhodospirillum rubrum during expression and switch-off of nitrogenase ac-tivity". J Bacteriol, 1987. 169: pp. 3035-43.

117. Ludden, P.W. and Burris, R.H., "Purification and properties of nitrogenase from Rhodospirillum rubrum, and evidence for phosphate, ribose and an

65

adenine-like unit covalently bound to the iron protein". Biochem J, 1978. 175: pp. 251-9.

118. Pope, M.R., Murrell, S.A., and Ludden, P.W., "Purification and properties of the heat-released nucleotide-modifying group from the inactive iron pro-tein of nitrogenase from Rhodospirillum rubrum". Biochemistry, 1985. 24: pp. 2374-80.

119. Pope, M.R., Murrell, S.A., and Ludden, P.W., "Covalent modification of the iron protein of nitrogenase from Rhodospirillum rubrum by adenosine diphosphoribosylation of a specific arginine residue". Proc Natl Acad Sci U

S A, 1985. 82: pp. 3173-7. 120. Ludden, P.W., Preston, G.G., and Dowling, T.E., "Comparison of active

and inactive forms of iron protein from Rhodospirillum rubrum". Biochem

J, 1982. 203: pp. 663-8. 121. Ma, Y. and Ludden, P.W., "Role of the dinitrogenase reductase arginine

101 residue in dinitrogenase reductase ADP-ribosyltransferase binding, NAD binding, and cleavage". J Bacteriol, 2001. 183: pp. 250-6.

122. Ludden, P.W. and Burris, R.H., "Activating factor for the iron protein of nitrogenase from Rhodospirillum rubrum". Science, 1976. 194: pp. 424-6.

123. Nordlund, S., Eriksson, U., and Baltscheffsky, H., "Necessity of a mem-brane component for nitrogenase activity in Rhodospirillum rubrum". Bio-

chim Biophys Acta, 1977. 462: pp. 187-95. 124. Saari, L.L., Triplett, E.W., and Ludden, P.W., "Purification and properties

of the activating enzyme for iron protein of nitrogenase from the photosyn-thetic bacterium Rhodospirillum rubrum". J Biol Chem, 1984. 259: pp. 15502-8.

125. Triplett, E.W., Wall, J.D., and Ludden, P.W., "Expression of the activating enzyme and Fe protein of nitrogenase from Rhodospirillum rubrum". J

Bacteriol, 1982. 152: pp. 786-91. 126. Fitzmaurice, W.P., et al., "Genes coding for the reversible ADP-

ribosylation system of dinitrogenase reductase from Rhodospirillum ru-

brum". Mol Gen Genet, 1989. 218: pp. 340-7. 127. Nielsen, G.M., et al., "Purification and characterization of an oxygen-stable

form of dinitrogenase reductase-activating glycohydrolase from Rhodospirillum rubrum". Biochem J, 1994. 302 pp. 801-6.

128. Berthold, C.L., et al., "Mechanism of ADP-ribosylation removal revealed by the structure and ligand complexes of the dimanganese mono-ADP-ribosylhydrolase DraG". Proc Natl Acad Sci U S A, 2009. 106: pp. 14247-52.

129. Lowery, R.G. and Ludden, P.W., "Purification and properties of dinitro-genase reductase ADP-ribosyltransferase from the photosynthetic bacte-rium Rhodospirillum rubrum". J Biol Chem, 1988. 263: pp. 16714-9.

130. Ponnuraj, R.K., et al., "NAD-, NMN-, and NADP-dependent modification of dinitrogenase reductases from Rhodospirillum rubrum and Azotobacter

vinelandii". FEBS Lett, 2005. 579: pp. 5751-8. 131. Grunwald, S.K. and Ludden, P.W., "NAD-dependent cross-linking of dini-

trogenase reductase and dinitrogenase reductase ADP-ribosyltransferase from Rhodospirillum rubrum". J Bacteriol, 1997. 179: pp. 3277-83.

132. Zhang, Y., et al., "Comparison studies of dinitrogenase reductase ADP-ribosyl transferase/dinitrogenase reductase activating glycohydrolase regu-

66

latory systems in Rhodospirillum rubrum and Azospirillum brasilense". J

Bacteriol, 1995. 177: pp. 2354-9. 133. Halbleib, C.M., et al., "Effects of perturbations of the nitrogenase electron

transfer chain on reversible ADP-ribosylation of nitrogenase Fe protein in Klebsiella pneumoniae strains bearing the Rhodospirillum rubrum dra op-eron". J Bacteriol, 2000. 182: pp. 3681-7.

134. Yakunin, A.F., et al., "Regulation of nitrogenase in the photosynthetic bac-terium Rhodobacter sphaeroides containing draTG and nifHDK genes from Rhodobacter capsulatus". Can J Microbiol, 2001. 47: pp. 206-12.

135. Drepper, T., et al., "Role of GlnB and GlnK in ammonium control of both nitrogenase systems in the phototrophic bacterium Rhodobacter capsula-

tus". Microbiology, 2003. 149: pp. 2203-12. 136. Klassen, G., et al., "Control of nitrogenase reactivation by the GlnZ protein

in Azospirillum brasilense". J Bacteriol, 2001. 183: pp. 6710-3. 137. Klassen, G., et al., "Nitrogenase switch-off by ammonium ions in Azospiril-

lum brasilense requires the GlnB nitrogen signal-transducing protein". Appl

Environ Microbiol, 2005. 71: pp. 5637-41. 138. Coutts, G., et al., "Membrane sequestration of the signal transduction pro-

tein GlnK by the ammonium transporter AmtB". EMBO J, 2002. 21: pp. 536-45.

139. Gruswitz, F., O'Connell, J., 3rd, and Stroud, R.M., "Inhibitory complex of the transmembrane ammonia channel, AmtB, and the cytosolic regulatory protein, GlnK, at 1.96 A". Proc Natl Acad Sci U S A, 2007. 104: pp. 42-7.

140. Conroy, M.J., et al., "The crystal structure of the Escherichia coli AmtB-GlnK complex reveals how GlnK regulates the ammonia channel". Proc

Natl Acad Sci U S A, 2007. 104: pp. 1213-8. 141. Durand, A. and Merrick, M., "In vitro analysis of the Escherichia coli

AmtB-GlnK complex reveals a stoichiometric interaction and sensitivity to ATP and 2-oxoglutarate". J Biol Chem, 2006. 281: pp. 29558-67.

142. Wang, H., et al., "Reversible membrane association of dinitrogenase reduc-tase activating glycohydrolase in the regulation of nitrogenase activity in Rhodospirillum rubrum; dependence on GlnJ and AmtB1". FEMS Micro-

biol Lett, 2005. 253: pp. 273-9. 143. Wolfe, D.M., Zhang, Y., and Roberts, G.P., "Specificity and regulation of

interaction between the PII and AmtB1 proteins in Rhodospirillum ru-

brum". J Bacteriol, 2007. 189: pp. 6861-9. 144. Zhang, Y., et al., "Effect of AmtB homologues on the post-translational

regulation of nitrogenase activity in response to ammonium and energy signals in Rhodospirillum rubrum". Microbiology, 2006. 152: pp. 2075-89.

145. Huergo, L.F., et al., "ADP-ribosylation of dinitrogenase reductase in Azospirillum brasilense is regulated by AmtB-dependent membrane se-questration of DraG". Mol Microbiol, 2006. 59: pp. 326-37.

146. Yakunin, A.F. and Hallenbeck, P.C., "AmtB is necessary for NH4+-induced

nitrogenase switch-off and ADP-ribosylation in Rhodobacter capsulatus". J

Bacteriol, 2002. 184: pp. 4081-8. 147. Hartmann, A. and Burris, R.H., "Regulation of nitrogenase activity by oxy-

gen in Azospirillum brasilense and Azospirillum lipoferum". J Bacteriol, 1987. 169: pp. 944-8.

67

148. Teixeira, P.F., Wang, H., and Nordlund, S., "Nitrogenase switch-off and regulation of ammonium assimilation in response to light deprivation in Rhodospirillum rubrum are influenced by the nitrogen source used during growth". J Bacteriol, 2010. 192: pp. 1463-6.

149. Zhu, Y., et al., "Identification of Rhodospirillum rubrum GlnB variants that are altered in their ability to interact with different targets in response to ni-trogen status signals". J Bacteriol, 2006. 188: pp. 1866-74.

150. Pawlowski, A., et al., "Yeast two-hybrid studies on interaction of proteins involved in regulation of nitrogen fixation in the phototrophic bacterium Rhodobacter capsulatus". J Bacteriol, 2003. 185: pp. 5240-7.

151. Huergo, L.F., et al., "In vitro interactions between the PII proteins and the nitrogenase regulatory enzymes dinitrogenase reductase ADP-ribosyltransferase (DraT) and dinitrogenase reductase-activating glycohy-drolase (DraG) in Azospirillum brasilense". J Biol Chem, 2009. 284: pp. 6674-82.

152. Noren, A. and Nordlund, S., "Changes in the NAD(P)H concentration caused by addition of nitrogenase 'switch-off' effectors in Rhodospirillum

rubrum G-9, as measured by fluorescence". FEBS Lett, 1994. 356: pp. 43-5.

153. Noren, A., Soliman, A., and Nordlund, S., "The role of NAD+ as a signal during nitrogenase switch-off in Rhodospirillum rubrum". Biochem J, 1997. 322: pp. 829-32.

154. Kessler, P.S., Daniel, C., and Leigh, J.A., "Ammonia switch-off of nitrogen fixation in the methanogenic archaeon Methanococcus maripaludis: mechanistic features and requirement for the novel GlnB homologues, NifI1 and NifI2". J Bacteriol, 2001. 183: pp. 882-9.

155. Dodsworth, J.A., Cady, N.C., and Leigh, J.A., "2-oxoglutarate and the PII homologues NifI1 and NifI2 regulate nitrogenase activity in cell extracts of Methanococcus maripaludis". Mol Microbiol, 2005. 56: pp. 1527-38.

156. Ekman, M., Tollback, P., and Bergman, B., "Proteomic analysis of the cyanobacterium of the Azolla symbiosis: identity, adaptation, and NifH modification". J Exp Bot, 2008. 59: pp. 1023-34.

157. Gallon, J.R., et al., "A novel covalent modification of nitrogenase in a cyanobacterium". FEBS Lett, 2000. 468: pp. 231-3.

158. Dixon, R., "Tandem promoters determine regulation of the Klebsiella

pneumoniae glutamine synthetase (glnA) gene". Nucleic Acids Res, 1984. 12: pp. 7811-30.

159. Miranda-Rios, J., et al., "The complete nucleotide sequence of the glnALG

operon of Escherichia coli K12". Nucleic Acids Res, 1987. 15: pp. 2757-70. 160. Johansson, M. and Nordlund, S., "Transcription of the glnB and glnA genes

in the photosynthetic bacterium Rhodospirillum rubrum". Microbiology, 1996. 142 ( Pt 5): pp. 1265-72.

161. Woolfolk, C.A. and Stadtman, E.R., "Regulation of glutamine synthetase. 3. Cumulative feedback inhibition of glutamine synthetase from Es-

cherichia coli". Arch Biochem Biophys, 1967. 118: pp. 736-55. 162. Hammarstrom, A., Soliman, A., and Nordlund, S., "Low- and high-activity

forms of glutamine synthetase from Rhodospirillum rubrum: sensitivity to feed-back effectors and activation of the low-activity form". Biochim Bio-

phys Acta, 1991. 1080: pp. 259-63.

68

163. Stadtman, E.R., "Discovery of glutamine synthetase cascade". Methods

Enzymol, 1990. 182: pp. 793-809. 164. Kingdon, H.S., Shapiro, B.M., and Stadtman, E.R., "Regulation of gluta-

mine synthetase. 8. ATP: glutamine synthetase adenylyltransferase, an en-zyme that catalyzes alterations in the regulatory properties of glutamine synthetase". Proc Natl Acad Sci U S A, 1967. 58: pp. 1703-10.

165. Johansson, B.C. and Gest, H., "Adenylylation/deadenylylation control of the glutamine synthetase of Rhodopseudomonas capsulata". Eur J Bio-

chem, 1977. 81: pp. 365-71. 166. Mehta, R., et al., "Adenylylation and catalytic properties of Mycobacterium

tuberculosis glutamine synthetase expressed in Escherichia coli versus my-cobacteria". J Biol Chem, 2004. 279: pp. 22477-82.

167. Janssen, K.A. and Magasanik, B., "Glutamine synthetase of Klebsiella

aerogenes: genetic and physiological properties of mutants in the adenyly-lation system". J Bacteriol, 1977. 129: pp. 993-1000.

168. Nordlund, S., et al., "Properties and regulation of glutamine synthetase from Rhodospirillum rubrum". J Bacteriol, 1985. 161: pp. 13-7.

169. Bender, R.A., et al., "Biochemical parameters of glutamine synthetase from Klebsiella aerogenes". J Bacteriol, 1977. 129: pp. 1001-9.

170. Stadtman, E.R., et al., "Enzymic procedures for determining the average state of adenylylation of Escherichia coli glutamine synthetase". Anal Bio-

chem, 1979. 95: pp. 275-85. 171. Hubbard, J.S. and Stadtman, E.R., "Regulation of glutamine synthetase. II.

Patterns of feedback inhibition in microorganisms". J Bacteriol, 1967. 93: pp. 1045-55.

172. Schutt, H. and Holzer, H., "Biological function of the ammonia-induced inactivation of glutamine synthetase in Escherichia coli". Eur J Biochem, 1972. 26: pp. 68-72.

173. Jonsson, A., Nordlund, S., and Teixeira, P.F., "Reduced activity of gluta-mine synthetase in Rhodospirillum rubrum mutants lacking the adenylyl-transferase GlnE". Res Microbiol, 2009. 160: pp. 581-4.

174. Buchanan, B.B. and Arnon, D.I., "A reverse KREBS cycle in photosynthe-sis: consensus at last". Photosynth Res, 1990. 24: pp. 47-53.

175. Wilkins, M.R., et al., "From proteins to proteomes: large scale protein iden-tification by two-dimensional electrophoresis and amino acid analysis". Biotechnology (N Y), 1996. 14: pp. 61-5.

176. Anderson, N.L. and Anderson, N.G., "Proteome and proteomics: new tech-nologies, new concepts, and new words". Electrophoresis, 1998. 19: pp. 1853-61.

177. Laemmli, U.K., "Cleavage of structural proteins during the assembly of the head of bacteriophage T4". Nature, 1970. 227: pp. 680-5.

178. Kenrick, K.G. and Margolis, J., "Isoelectric focusing and gradient gel elec-trophoresis: a two-dimensional technique". Anal Biochem, 1970. 33: pp. 204-7.

179. O'Farrell, P.H., "High resolution two-dimensional electrophoresis of pro-teins". J Biol Chem, 1975. 250: pp. 4007-21.

180. Klose, J., "Protein mapping by combined isoelectric focusing and electro-phoresis of mouse tissues. A novel approach to testing for induced point mutations in mammals". Humangenetik, 1975. 26: pp. 231-43.

69

181. Righetti, P.G. and Bossi, A., "Isoelectric focusing in immobilized pH gra-dients: an update". J Chromatogr B Biomed Sci Appl, 1997. 699: pp. 77-89.

182. Gorg, A., Postel, W., and Gunther, S., "The current state of two-dimensional electrophoresis with immobilized pH gradients". Electropho-

resis, 1988. 9: pp. 531-46. 183. Weiss, W. and Gorg, A., "High-resolution two-dimensional electrophore-

sis". Methods Mol Biol, 2009. 564: pp. 13-32. 184. Chevalier, F., "Highlights on the capacities of "Gel-based" proteomics".

Proteome Sci, 2010. 8: pp. 23. 185. Santoni, V., et al., "Towards the recovery of hydrophobic proteins on two-

dimensional electrophoresis gels". Electrophoresis, 1999. 20: pp. 705-11. 186. Zahedi, R.P., Meisinger, C., and Sickmann, A., "Two-dimensional ben-

zyldimethyl-n-hexadecylammonium chloride/SDS-PAGE for membrane proteomics". Proteomics, 2005. 5: pp. 3581-8.

187. Schagger, H. and von Jagow, G., "Blue native electrophoresis for isolation of membrane protein complexes in enzymatically active form". Anal Bio-

chem, 1991. 199: pp. 223-31. 188. Reisinger, V. and Eichacker, L.A., "Analysis of membrane protein com-

plexes by Blue Native PAGE". Proteomics, 2006. 6 Suppl 2: pp. 6-15. 189. Roesli, C., Elia, G., and Neri, D., "Two-dimensional mass spectrometric

mapping". Curr Opin Chem Biol, 2006. 10: pp. 35-41. 190. Hunt, D.F., et al., "Protein sequencing by tandem mass spectrometry". Proc

Natl Acad Sci U S A, 1986. 83: pp. 6233-7. 191. Fournier, M.L., et al., "Multidimensional separations-based shotgun pro-

teomics". Chem Rev, 2007. 107: pp. 3654-86. 192. Yates, J.R., Ruse, C.I., and Nakorchevsky, A., "Proteomics by mass spec-

trometry: approaches, advances, and applications". Annu Rev Biomed Eng, 2009. 11: pp. 49-79.

193. Perry, R.H., Cooks, R.G., and Noll, R.J., "Orbitrap mass spectrometry: instrumentation, ion motion and applications". Mass Spectrom Rev, 2008. 27: pp. 661-99.

194. Tichi, M.A. and Tabita, F.R., "Maintenance and control of redox poise in Rhodobacter capsulatus strains deficient in the Calvin-Benson-Bassham pathway". Arch Microbiol, 2000. 174: pp. 322-33.

195. Romagnoli, S. and Tabita, F.R., "Carbon dioxide metabolism and its regu-lation in nonsulfur purple photosynthetic bacteria", in "The purple photo-trophic bacteria. Advances in photosynthesis and respiration", C.N. Hunter, et al., Editors. 2009, Springer: Dordrecht. pp. 563-76.

196. Dubbs, J.M. and Tabita, F.R., "Regulators of nonsulfur purple phototrophic bacteria and the interactive control of CO2 assimilation, nitrogen fixation, hydrogen metabolism and energy generation". FEMS Microbiol Rev, 2004. 28: pp. 353-76.

197. Evans, M.C., Buchanan, B.B., and Arnon, D.I., "A new ferredoxin-dependent carbon reduction cycle in a photosynthetic bacterium". Proc Natl

Acad Sci U S A, 1966. 55: pp. 928-34. 198. Buchanan, B.B., Evans, M.C., and Arnon, D.I., "Ferredoxin-dependent

carbon assimilation in Rhodospirillum rubrum". Arch Mikrobiol, 1967. 59: pp. 32-40.

70

199. Brostedt, E., et al., "Electron transport to nitrogenase in Rhodospirillum

rubrum: The role of NAD(P)H as electron donor and the effect of fluoroacetate on nitrogenase activity". FEMS Microbiol Lett, 1997. 150: pp. 263-7.

200. Park, S.J., Tseng, C.P., and Gunsalus, R.P., "Regulation of succinate dehy-drogenase (sdhCDAB) operon expression in Escherichia coli in response to carbon supply and anaerobiosis: role of ArcA and Fnr". Mol Microbiol, 1995. 15: pp. 473-82.

201. Park, S.J., Chao, G., and Gunsalus, R.P., "Aerobic regulation of the su-

cABCD genes of Escherichia coli, which encode alpha-ketoglutarate dehy-drogenase and succinyl coenzyme A synthetase: roles of ArcA, Fnr, and the upstream sdhCDAB promoter". J Bacteriol, 1997. 179: pp. 4138-42.

202. Georgellis, D., Kwon, O., and Lin, E.C., "Quinones as the redox signal for the arc two-component system of bacteria". Science, 2001. 292: pp. 2314-6.

203. Bekker, M., et al., "The ArcBA two-component system of Escherichia coli

is regulated by the redox state of both the ubiquinone and the menaquinone pool". J Bacteriol, 2010. 192: pp. 746-54.

204. Georgellis, D., et al., "Signal decay through a reverse phosphorelay in the Arc two-component signal transduction system". J Biol Chem, 1998. 273: pp. 32864-9.

205. Iuchi, S. and Lin, E.C., "arcA (dye), a global regulatory gene in Es-

cherichia coli mediating repression of enzymes in aerobic pathways". Proc

Natl Acad Sci U S A, 1988. 85: pp. 1888-92. 206. Grammel, H. and Ghosh, R., "Redox-state dynamics of ubiquinone-10

imply cooperative regulation of photosynthetic membrane expression in Rhodospirillum rubrum". J Bacteriol, 2008. 190: pp. 4912-21.

207. Zhang, Y., et al., "Isolation and characterization of draT mutants that have altered regulatory properties of dinitrogenase reductase ADP-ribosyltransferase in Rhodospirillum rubrum". Microbiology, 2001. 147: pp. 193-202.

208. Sawers, G. and Bock, A., "Anaerobic regulation of pyruvate formate-lyase from Escherichia coli K-12". J Bacteriol, 1988. 170: pp. 5330-6.

209. Shalel-Levanon, S., San, K.Y., and Bennett, G.N., "Effect of ArcA and FNR on the expression of genes related to the oxygen regulation and the glycolysis pathway in Escherichia coli under microaerobic growth condi-tions". Biotechnol Bioeng, 2005. 92: pp. 147-59.

210. Uffen, R.L., "Growth properties of Rhodospirillum rubrum mutants and fermentation of pyruvate in anaerobic, dark conditions". J Bacteriol, 1973. 116: pp. 874-84.

211. Brush, E.J., Lipsett, K.A., and Kozarich, J.W., "Inactivation of Escherichia

coli pyruvate formate-lyase by hypophosphite: evidence for a rate-limiting phosphorus-hydrogen bond cleavage". Biochemistry, 1988. 27: pp. 2217-22.

212. Solaiman, D. and Uffen, R.L., "Pyruvate-dependent diauxic growth of Rhodospirillum rubrum in light". J Bacteriol, 1982. 152: pp. 1175-87.

213. Paul, T.D. and Ludden, P.W., "Adenine nucleotide levels in Rhodospirillum

rubrum during switch-off of whole-cell nitrogenase activity". Biochem J, 1984. 224: pp. 961-9.

71

214. Li, J.D., Hu, C.Z., and Yoch, D.C., "Changes in amino acid and nucleotide pools of Rhodospirillum rubrum during switch-off of nitrogenase activity initiated by NH4

+ or darkness". J Bacteriol, 1987. 169: pp. 231-7. 215. Nordlund, S. and Hoglund, L., "Studies of the adenylate and pyridine-

nucleotide pools during nitrogenase switch-off in Rhodospirillum rubrum". Plant Soil, 1986. 90: pp. 203-9.

216. Tremblay, P.L., et al., "Membrane sequestration of PII proteins and nitro-genase regulation in the photosynthetic bacterium Rhodobacter capsula-

tus". J Bacteriol, 2007. 189: pp. 5850-9.