Aquatic photosynthesis accounts for 50% of the global annual net primary production (NPP), despite frequent low availability and limited diffusion of CO2 in the aquatic milieu, and low affinity for CO2 by the primary carboxylating enzyme, Ribulose 1,5-bisphosphate carboxylase/oxygenase (RuBisCO). Many eukaryotic algae, and a single group of land plants, the hornworts, have an inducible carbon concentrating mechanism (CCM), to overcome these limitations. The efficiency of the CCM is improved when RuBisCO is localised to a subcellular compartment, the pyrenoid, which is hypothesised to act as a diffusion barrier for CO2 . Although the pyrenoid is a major player in global carbon balance (we estimate 10-15% of NPP), it is one of the few remaining prominent cellular features without a precise molecular or physiological definition. Under ambient CO2 , at least 90% of the cellular RuBisCO is packed into a dense matrix, together with the chaperone RuBisCO activase. Thylakoid membranes usually traverse the pyrenoid matrix, and the carboxylating substrate is thought to be delivered to the active sites of the enzyme via a carbonic anhydrase located in the lumen of these thylakoids. The mechanism of aggregation of constituents within the pyrenoid, however, still remains largely unknown. Comprehensive mutant screens have yet to reveal mutants incapable of forming pyrenoids other than those mutants with a defective RuBisCO holoenzyme, whereas DNA microarray studies uncovered little with reference to pyrenoid ultrastructure or aggregation. Taken together, this evidence raises the possibility that the basis of pyrenoid ultrastructure and aggregation lies entirely in sequence variations of RuBisCO itself. This work explored, firstly, the advantages conferred by an active CCM in hornworts and in unicellular algae, compared with the passive CO2 acquisition in most terrestrial plants. A physiological framework to CCM and pyrenoid-based photosynthesis, and isotopic discrimination, was provided by comparing the photosynthetic characteristics of selected bryophytes and algae, differing in chloroplast morphology and degrees of internalisation of gas exchanges. The results showed that on-line, carbon isotope discrimination values were a good indicator of CCM occurrence, as well as liquid-phase diffusion limitation, and biochemical limitations resulting from declining RuBisCO activity and electron transport. The methodology was used to diagnose the presence of an active CCM, and the extent of CO2 leakage. Secondly, the effect of RuBisCO sequence variations on the pyrenoid, and associated CCM, was studied using the model alga Chlamydomonas reinhardtii. The starting premise was the report by Nozaki et al. (2002) that, in some species of the family Chlamydomonaceae, a few amino acid residues within the RuBisCO large subunit (LSU) correlated strongly with pyrenoid formation. The specific roles of seven LSU residues were studied by site-directed mutagenesis. Whilst the mutations reduced the affinity of RuBisCO for CO2 and increased CO2 leakage, compared to wild-type Chlamydomonas, there was no effect on the pyrenoid phenotype. Informed by observations that Chlamydomonas mutants with a hybrid RuBisCO, composed of a native LSU, and higher plant small subunit (SSU), lacked a pyrenoid (Genkov et al., 2010), and that defined SSU alterations were neutral with respect to the pyrenoid (Genkov and Spreitzer, 2006), hitherto unexplored SSU domains were modified. A pyrenoid was successfully restored by replacing jointly the two solvent-exposed α-helices, whereas single α-helix replacements had no effect. However, leakage values indicated that the associated CCM was not fully operative, suggesting important correlates between the RuBisCO SSU and the CCM, besides the conditioning of pyrenoid formation. If the pyrenoid is partly defined by simple sequence variations in the RuBisCO SSU, as suggested by the evidence outlined in this thesis, there is the tantalising possibility that transformation of a biophysical CCM into crop plants could be a tractable approach for the future.

Abstract The gammaproteobacterium Thiomicrospira crunogena XCL–2 is a hydrothermal vent chemolithoautotroph that has a carbon concentrating mechanism (CCM), which is functionally similar to that of cyanobacteria. At hydrothermal vents, dissolved inorganic carbon (DIC) concentrations and pH values fluctuate over time, with CO2 concentrations ranging from 20 μM to greater than 1 mM, therefore having a CCM would provide an advantage when CO2 availability is very low as CCMs generate intracellular DIC concentrations much higher than extracellular, thereby providing sufficient substrate for carbon fixation. The CCM in T. crunogena includes α–carboxysomes (intracellular inclusions containing form IA RubisCO and carbonic anhydrase), and also presumably requires at least one active HCO3 µ transporter to generate the elevated intracellular concentrations of DIC. To determine whether RubisCO itself might be adapted to low CO2 concentrations, the KCO2 for purified carboxysomal RubisCO was measured (250 μM SD ±; 40) and was much greater than that of whole cells (1.03 μM). This finding suggests that the primary adaptation by T. crunogena to low–DIC conditions has been to enhance DIC uptake, presumably by energy–dependent membrane transport systems that are either ATP–dependent and/or dependent on membrane potential (δ ψ). To determine the mechanism for active DIC uptake, cells were incubated in the presence of inhibitors targeting ATP synthesis andδ ψ. After separate incubations with the ATP synthase inhibitor DCCD and the protonophore CCCP, intracellular ATP was diminished, as was the concentration of intracellular DIC and fixed carbon, despite the absence of an inhibitory effect on δ ψ in the DCCD–incubated cells. In some organisms, DCCD inhibits the NDH–1 and bc1 complexes so it was necessary to verify that ATP synthase was the primary target of DCCD in T. crunogena. Both electron transport complex activities were assayed in the presence and absence of DCCD and there was no significant difference between inhibited (309.0 μmol/s for NDH–1 and 3.4 μmol/s for bc1) and uninhibited treatments (271.7 μmol/s for NDH–1 and 3.6 μmol/s for bc1). These data support the hypothesis that an ATP–dependent transporter is primarily responsible for HCO3 µ transport in T. crunogena. The ATP–dependent transporter solute–binding protein gene (cmpA) from Synechococcus elongatus PCC 7942, was used to perform a BLAST query. Tcr_1153 was the closest match in the T. crunogena genome. However, the gene neighborhood and the result of a maximum likelihood tree suggest that Tcr_1153 is a nitrate transporter protein. Work is underway to find the genes responsible for this ATP–dependent transporter.

Gammaproteobacterium Thiomicrospira crunogena thrives at deep-sea vents despite extreme oscillations in the environmental supply of dissolved inorganic carbon (DIC; =CO2 + HCO3- + CO3-2). Survival in this habitat is likely aided by the presence of a carbon concentrating mechanism (CCM). Though CCMs are well-documented in cyanobacteria, based on this study T. crunogena is the first chemolithoautotroph to have a physiologically characterized CCM. T. crunogena is capable of rapid growth in the presence of 20 micrometers DIC, has the ability to use both extracellular HCO3- and CO2, and generates intracellular DIC concentrations 100-fold greater than extracellular, all of which are consistent with a CCM analogous to those present in cyanobacteria. Interestingly, however, the T.crunogena genome lacks apparent orthologs of many of the components of the cyanobacteria CCM (e.g., HCO3- transporters). However, despite this lack, several candidate genes were identified during genome annotation as likely to play a role in DIC uptake and fixation (three carbonic anhydrase genes: alpha-CA, beta-CA, and csoSCA, as well as genes encoding three RubisCO enzymes: cbbLS, CScbbLS, and cbbM, which encode a cytoplasmic form I RubisCO, a carboxysomal form I RubisCO, and a form II RubisCO, respectively). In order to clarify their possible roles in DIC uptake and fixation, alpha-CA, beta-CA and csoSCA transcription by low-DIC and high-DIC T. crunogena were assayed by qRT PCR, heterologous expression in E. coli, and potentiometric assays of low-DIC and high-DIC T. crunogena. Transcription of alpha-CA and beta-CA were not sensitive to the DIC concentration available during growth. When overexpressed in E.coli, carbonic anhydrase activity was detectable, and it was possible to measure the effects of the classical carbonic anhydrase inhibitors ethoxyzolamide and acetazolamide, as well as dithiothreitol (DTT; recently determined to be a carboxysomal CA inhibitor). The alpha-CA was sensitive to both of the classical inhibitors, but not DTT. Beta-CA was insensitive to all inhibitors tested, and the carboxysomal carbonic anhydrase was sensitive to both ethoxyzolamide and DTT. The observation that the CA activity measureable potentiometrically with intact T. crunogena cells is sensitive to classical inhibitors, but not DTT, strongly suggests the alpha-CA is extracellular. The presence of carbonic anhydrase activity in crude extracts of high-DIC cells that was resistant to classical inhibitors suggests that beta-CA may be more active in high-DIC cells. Incubating cells with ethoxyzolamide (which permeates cells rapidly) resulted in inhibition of carbon fixation, but not DIC uptake, while incubation with acetazolamide (which does not permeate cells rapidly) had no apparent effect on either carbon fixation or DIC uptake. The observations that inhibition of alpha-CA has no effect on DIC uptake and fixation, and that the beta-CA is not transcribed more frequently under low-DIC conditions, make it unlikely that either play a role in DIC uptake and fixation in low-DIC cells. Further studies are underway to determine the roles of alpha-CA and beta-CA in T. crunogena. To assay the entire genome for genes transcribed more frequently under low-DIC conditions, and therefore likely to play a role in the T. crunogena CCM, oligonucleotide arrays were fabricated using the T. crunogena genome sequence. RNA was isolated from cultures grown in the presence of both high (50 mM) and low (0.05 mM) concentrations of DIC, directly labeled with cy5 fluorophore, and hybridized to microarrays. Genes encoding the three RubisCO enzymes present in this organism demonstrated differential patterns of transcription consistent with what had been observed previously in Hydrogenovibrio marinus. Genes encoding two conserved hypothetical proteins were also found to be transcribed more frequently under low-DIC conditions, and this transcription pattern was verified by qRT-PCR. Knockout mutants are currently being generated to determine whether either gene is necessary for growth under low-DIC conditions. Identifying CCM genes and function in autotrophs beyond cyanobacteria will serve as a window into the physiology required to flourish in microbiallydominated ecosystems where noncyanobacterial primary producers dominate.

Chlamydomonas reinhardtii possesses a single-cell-based CO2-concentrating mechanism (CCM). The CCM is an important element of algal photosynthesis, metabolism, growth and biomass production, which works by increasing the concentration of inorganic carbon (Ci) in the pyrenoid, a dense RuBisCO-packed structure within the chloroplast. This suppresses RuBisCO oxygenase activity and associated photorespiration. The enhanced efficiency of CO2 assimilation in the pyrenoid via CCM had been modelled theoretically as a requirement for successful CCM in higher plant systems. The ultimate aim of my research is to understand the biogenesis of the pyrenoid using a set of CCM mutants with pyrenoidal defects. Immunofluorescence methods and spot growth tests under different CO2 concentrations were performed on mutants with CCM defects generated by an insertional mutagenesis screen. Morphological and physiological characterisation of these mutants revealed differences in the pyrenoid morphology, the ability for RuBisCO to aggregate into the pyrenoid and the formation of thylakoidal tubule network associated with the pyrenoid. The thylakoid tubule network may be linked to the transport of inorganic carbon into the pyrenoid as part of the CCM. Further characterisation of one of the mutants gave rise to the hypothesis that the gene of interest, Cre11.g467712 (SAGA), is a multi-functional anchor protein related to the structural formation of the pyrenoid and may be another essential component of the pyrenoid.