Global identification of regulatory RNAs in the cyanobacterium Nostoc sp. PCC 7120. Functional characterization of Yfr1 and NsiR1

Cover. Co-expression network of transcripts of Nostoc sp. PCC 7120 with expression changes after removal of combined nitrogen (left). 3D structural model of sRNAs Yfr1 (top right) and NsiR1 (bottom right). Models were generated by RNAComposer (Popenda et al. 2012. Nucleic Acids Research, 40: e112. doi: 10.1093/nar/gks339). Instituto de Bioquímica Vegetal y Fotosíntesis Departamento de Bioquímica Vegetal y Biología Molecular Universidad de Sevilla - Consejo Superior de Investigaciones Científicas TESIS DOCTORAL Global identification of regulatory RNAs in the cyanobacterium Nostoc sp. PCC 7120. Functional characterization of Yfr1 and NsiR1. Trabajo presentado por Manuel Brenes Álvarez para optar al grado de Doctor Manuel Brenes Álvarez, Sevilla, 2019 Director Directora Dr. Agustín Vioque Peña Dra. Alicia María Muro Pastor Catedrático de la Universidad de Sevilla Científica Titular del Consejo Superior de Investigaciones Científicas FUNDING This Doctoral Thesis has been carried out at the Instituto de Bioquímica Vegetal y Fotosíntesis (IBVF, CSIC-US) and has been funded by a FPU predoctoral contract from Ministry of Education, Culture and Sport (FPU014/05123). The work has been funded by projects from Ministry of Economy and Competitiveness (BFU2013-48282-C2-1-P, Non-coding RNAs involved in adaptation to nitrogen stress and cell differentiation in cyanobacteria) and from State Research Agency, Ministry of Economy, Industry and Competitiveness (BFU2016- 74943-C2-1-P, Participation of non-coding RNAs in regulatory circuits controlled by nitrogen availability in cyanobacteria) whose Principal Investigator was Dr. Alicia María Muro Pastor. Both projects were co-financed by the European Regional Development Fund (ERDF). In addition, a three-months stay at the laboratory of Dr. Wolfgang Hess (University of Freiburg, Germany) was funded by a fellowship from Ministry of Education, Culture and Sport (EST16/00088). A mis abuelos i INDEX 1. INTRODUCTION ............................................................................................................... 1 1.1 The cyanobacteria .......................................................................................................... 3 1.1.1 General characteristics ...................................................................................................... 3 1.1.2 Cyanobacterial cell envelope ............................................................................................. 7 1.1.2.1 Synthesis and remodeling of peptidoglycan ................................................................ 7 1.1.2.2 Transporters ................................................................................................................. 9 1.1.3 Metabolic characteristics ................................................................................................. 10 1.1.3.1 Carbon assimilation in cyanobacteria ........................................................................ 11 1.1.3.2 Nitrogen assimilation in cyanobacteria ..................................................................... 11 1.1.4 Adaptation to combined nitrogen deprivation ............................................................... 15 1.1.4.1 General response ....................................................................................................... 15 1.1.4.2 Heterocyst differentiation.......................................................................................... 16 1.2 RNAs regulators in bacteria .......................................................................................... 20 1.2.1 Types of RNA regulators .................................................................................................. 23 1.2.1.1 RNA regulators encoded in cis ................................................................................... 23 1.2.1.2 RNA regulators encoded in trans ............................................................................... 24 1.2.2 Identification of RNA regulators and their targets .......................................................... 27 1.2.3 RNA regulators in cyanobacteria ..................................................................................... 28 1.2.3.1 Identification of RNA regulators in cyanobacteria ..................................................... 28 1.2.3.2 Physiological processes regulated by RNA regulators in cyanobacteria .................... 29 2. OBJECTIVES.................................................................................................................... 33 3. SUMMARY OF RESULTS .................................................................................................. 37 3.1 CHAPTER I: A computational approach for the identification of conserved sRNAs in heterocyst-forming cyanobacteria........................................................... 45 3.2 CHAPTER II: Yfr1, a widely conserved sRNA, regulates the integrity of the cell wall and its remodeling during heterocyst differentiation .................................. 67 3.3 CHAPTER III: A co-expression network to dissect the complex transcriptome of Nostoc sp. PCC 7120 during heterocyst differentiation............................. 103 3.4 CHAPTER IV: NsiR1, a sRNA with multiple copies, regulates heterocyst differentiation . 157 4. GENERAL DISCUSSION .................................................................................................. 209 5. CONCLUSIONS ............................................................................................................. 221 6. GENERAL REFERENCES ................................................................................................. 225 1. INTRODUCTION Introduction 1. INTRODUCTION 1.1 The cyanobacteria 1.1.1 General characteristics Cyanobacteria are Gram-negative bacteria that form a monophyletic group inside the eubacteria (Woese, 1987). They are photosynthetic microorganisms with a photosynthetic apparatus similar to that of chloroplast of algae and higher plants (DeRuyter and Fromme, 2008) and are the only prokaryotes able to perform oxygenic photosynthesis. Nowadays, it is widely accepted that chloroplasts from algae and higher plants have evolved from ancient cyanobacteria that established a symbiotic relationship with a phagotrophic eukaryote (Margulis, 1975; Ochoa de Alda et al., 2014). However, there are some differences between the photosynthetic apparatus of cyanobacteria and algae and those of higher plants. For example, in contrast to algae and higher plants, most cyanobacteria do not have chorophyll b (Stanier and Cohen-Bazire, 1977) and contain phycobilisomes as supramolecular light-harvesting complexes (Grossman et al., 1993b). Cyanobacteria played a crucial role in Earth’s history. They are thought to be the first organisms that performed an oxygenic photosynthesis (Buick, 1992). The oldest cyanobacterial fossils were found in sedimentary rocks generated more than 3500 million years ago (Schopf and Packer, 1987) and, although with some controversy, it is now accepted that these first photosynthetic organisms may have appeared between 2350 (Kirschvink and Kopp, 2008) and 3600-3700 million years ago (Garcia-Pichel et al., 2019). The ability to use water as electron donor, and the production of oxygen as a result, contributed to one of the most drastic changes in the Biosphere, the change from an anoxygenic atmosphere to an oxygenic atmosphere, a process known as the Great Oxidation Event (GEO) (Lyons et al., 2014). This event allowed the subsequent evolution of heterotrophic aerobic metabolism (Falkowski et al., 2005). In addition, cyanobacteria are also very relevant in the actual Biosphere. Not only is a substantial amount of CO2 fixation carried out by cyanobacterial photosynthesis in the oceans, but, given that some cyanobacteria are also able to fix atmospheric nitrogen, this group of organisms plays an important role in the carbon, nitrogen and oxygen cycles (Montoya et al., 2004). Cyanobacteria have colonized diverse ecological niches, from oceans, fresh water and soil to extreme environments such as hot springs, deserts and artic regions. In addition, some species are able to differentiate several cell types or to live in a symbiotic relationship with fungi, plants or protists (Adams, 2002; Meeks and Elhai, 2002). This ability to adapt to environments 3 Introduction with variable light intensity, temperature or nutrient availability could imply the existence of specific regulatory mechanisms in these bacteria, in addition to more general mechanisms equivalent to those found in other prokaryotes (Tandeau de Marsac and Houmard, 1993). The wide variety of morphologies shown by cyanobacteria was one of the first criteria used for its classification (Rippka et al., 1979), leading to a classification in five sections (Figure 1). Figure 1. Representative cyanobacterial strains from the sections in the morphological classification of Rippka et al., 1979. Numbers indicate the taxonomic sections; (I) Chroococcales, (II) Pleurocapsales, (III) Oscillatoriales, (IV) Nostocales and (V) Stigonematales. Images are courtesy of Alicia M. Muro (I, II), José Enrique Frías, Servicio de Cultivos Biológicos del Centro de Investigaciones Científicas Isla de la Cartuja (cicCartuja) (III, V) and Elvira Olmedo (IV). Section I (Chroococcales), is formed by unicellular cyanobacteria that reproduce by binary fission or by budding (e.g. genera Synechocystis or Synechococcus). Section II (Pleurocapsales), including genus Pleurocapsa, is formed by unicellular cyanobacteria that reproduce by multiple fission. Section III (Oscillatoriales) is formed by filamentous strains that do not differentiate specialized cells (e.g. genera Oscillatoria or Spirulina). Section IV (Nostocales), including genera Nostoc and Anabaena, and Section V (Stigonematales), including genus Fischerella, are formed by filamentous strains that have the capacity of cellular differentiation. Strains with cell division taking place in one plane, therefore unbranched, are classified in Section IV whereas branched strains, in which cell division may take place in more than one plane, are classified in Section V. Strains in these two last sections are able to differentiate several cell types. They can differentiate heterocyst (specialized cells that fix atmospheric nitrogen) under nitrogen deprivation or akinetes (a spore-like cell) under certain stress conditions. In addition, some strains can differentiate hormogonia, small motile filaments 4 Introduction that participate in the spreading of the organism and the establishment of symbiotic relationships (Flores and Herrero, 2010). More recent classifications are based on methods of molecular phylogeny. These methods have shown that the morphological classification is not coherent with the phylogeny (Giovannoni et al., 1988; Turner et al., 1999; Sánchez-Baracaldo et al., 2005; Tomitani et al., 2006) and only sections IV and V, that include the strains with the capacity of cellular differentiation, are phylogenetically coherent (Sánchez-Baracaldo et al., 2005; Tomitani et al., 2006; Schirrmeister et al., 2015). In 2016, when our approach for the prediction of sRNAs was carried out, the genomes available were basically those included in the analysis of Shih et al., 2013. Therefore, we have used the phylogenetic tree in Shih et al., 2013 as reference (Figure 2). Some conclusions derived from inspection of this phylogenetic tree are a) multicellularity has emerged several times along the evolution of cyanobacteria, and b) some unicellular species seem to be the result of a loss of multicellularity (Shih et al., 2013). Cyanobacterial genomes show diverse sizes, organization and ploidy. Genes conserved in all cyanobacterial genomes are known as core genes. These genes are related to important and essential physiological process, such as translation, photosynthesis, CO2 fixation and synthesis of chlorophyll and ATP. Along the evolution of cyanobacteria, strains with more complex physiology have acquired specific genes by duplication and genetic drift or by horizontal transfer (Shi and Falkowski, 2008). As a consequence of the acquisition and duplication of genes along the evolution of cyanobacteria, the size of cyanobacterial genomes varies from 1.66 Mb in Prochlorococcus MED4, considered the minimal size of the genome for a non-symbiotic photosynthetic organism (Rocap et al., 2003), to 9.2 Mb in Nostoc punctiforme. The GC content also varies widely between different strains, from 32% to 71%, similar to the GC content range found in prokaryotes as a whole (Herdman et al., 1979). Some cyanobacteria, like Prochlorococcus or Synechococcus, have only one circular chromosome. Other cyanobacteria, like our model organism Nostoc sp. PCC 7120, have their genomic content arranged in a circular chromosome and several plasmids, and some strains even have linear chromosomes, as Cyanothece 51142 (Hess, 2011). Finally, polyploidy is another source of genome heterogeneity. There are monoploid, oligoploid and polyploid strains with more than 100 copies of their chromosome, as some strains of Synechocystis (Griese et al., 2011). 5 Introduction Figure 2. Cyanobacterial phylogenetic tree according to Shih et al., 2013. Branches are coloured according to the sections in the classification of Rippka et al., 1979 (colour code top right). Framed amplified area contains heterocystforming strains, including branched (purple bar) or non-branched strains (pink bar). Nostoc sp. PCC 7120, the model organism used in this work, is indicated by a red arrow. The image, adapted from Shih et al., 2013, shows cyanobacterial genomes sequenced in that work (red). Morphological transitions are indicated by blue triangles. 6 Introduction The cyanobacterium used as model organism in this Thesis is Nostoc sp. PCC 7120 (a.k.a Anabaena sp. PCC 7120), a non-branched filamentous cyanobacteria with the capacity to differentiate heterocysts (Section IV). The genome of Nostoc sp. PCC 7120 is distributed in a circular chromosome of 6,4 Mb and six plasmids with a size between 405 kb and 5.5 kb, named α, β, γ, δ, ε and ζ (Kaneko et al., 2001). 1.1.2 Cyanobacterial cell envelope The cell wall of cyanobacteria resembles a Gram-negative architecture, with a peptidoglycan layer between an inner cytoplasmic membrane and an outer membrane (Silhavy et al., 2010). However, cyanobacterial cell envelopes have some unusual features. The peptidoglycan layer is relatively thick and has a high-degree of crosslinking, typical features of Gram-positive strains (Hoiczyk and Hansel, 2000). In addition, the outer membrane in multicellular cyanobacteria does not enter the septa between adjacent cells (Wilk et al., 2011), generating, as a consequence, a continuous periplasm (Mariscal et al., 2007). 1.1.2.1 Synthesis and remodelling of peptidoglycan Peptidoglycan is a strong and flexible mesh that surrounds the cell, contributes to its shape and protects it against osmotic pressure. Peptidoglycan contains glycan strands of alternating N-acetylglucosamine (GluNAc) and N-acetylmuramic acid (MurNAc) cross-linked by short peptides (Vollmer et al., 2008a). A pentapeptide substitutes the D-lactoyl group of each MurNAc residue and, in Gram-negative strains, usually has the sequence L-Ala-γ-D-Glu-meso- 2,6-diaminopimelic-acid-D-Ala-D-Ala. The last D-Ala residue is lost when the D-Ala at position 4 of the pentapeptide in one glycan strand is crosslinked to the diaminopimelic acid at position 3 of the pentapeptide in another glycan strand. The biosynthesis of peptidoglycan involves three main steps; the synthesis of the disaccharide with the pentapeptide in the cytosol, the translocation of this disaccharidepentapeptide unit across the plasma membrane and the transglycosylation and transpeptidation of new units in the nascent peptidoglycan strands (Figure 3). The first step of peptidoglycan synthesis is the conversion of GluNAc to MurNAc. Later, the sequential addition of amino acids (including D-Ala generated by alanine racemase) to the nascent pentapeptide 7 Introduction depends on NtcA but also on HetR, the master regulator of heterocyst differentiation (see section 1.1.4.2). Cyanobacteria can also use organic compounds as nitrogen source. For example, Nostoc sp. PCC 7120 can use glutamine, arginine (Burnat and Flores, 2014; Burnat et al., 2018) or urea (Valladares et al., 2002). However, nitrogen sources are used hierarchically, with inorganic nitrogen sources, as nitrate or nitrite, used preferentially. Nitrate or nitrite enter the cells by high affinity transporters and are reduced in two successive reactions that use reduced ferredoxin as electron donor, linking nitrogen assimilation to photosynthesis (Flores et al., 2005). Nitrate is reduced to nitrite, nitrite is reduced to ammonium and ammonium is finally assimilated by the GS-GOGAT cycle (Figure 4). Gene products involved in nitrate assimilation are encoded in the nir operon, whose transcription is activated by NtcA (Figure 5) (Frías et al., 1997). This operon encodes nitrite reductase (NirA), the subunits of one of the main nitrate/nitrite transporters (NrtABCD) and nitrate reductase (NarB). In addition, NtcA also activates the expression of another operon that encodes NtcB and CnaT, two proteins necessary for the full induction of the nir operon. NtcB is a transcription factor of the LysR family that binds to the promoter region and promotes the transcription of the nir operon (Frías et al., 2000) and CnaT boosts the expression of the nir operon by an unknown mechanism (Frías et al., 2003). The activity of nitrate assimilation enzymes is also regulated by other factors. For example, proteins NirB and NarM seem to interact with NirA and NarB respectively (Frías and Flores, 2015), and these interactions are necessary for full activity of these enzymes (Frías and Flores, 2010; 2015). Finally, the NirA/NirB complex seems to negatively affect the nir operon expression when nitrate/nitrite are not present in the media (Frías and Flores, 2010). Figure 5. Regulation of the nitrate assimilation cluster. NtcA bound to 2-oxoglutarate (2-OG) activates the expression of the nir operon encoding nitrite reductase (nirA), nitrate reductase (narB) and nitrate/nitrite transporter (nrtABCD). ntcB and cnaT are also induced by NtcA+2-OG and their products are necessary for the full activation of the nir operon; NtcB binds to the promoter of the nir operon and activates its expression whereas CnaT enhances the expression of the nir operon by an unknown mechanism. NirB and NarM are also necessary for the proper activity of nitrite reductase (NirA) and nitrate reductase (NarB), respectively. Finally, in the absence of nitrate or nitrite, the complex NirA/NirB seems to negatively affect the expression of the nir operon. Taken from Herrero and Flores, 2019. 14 Introduction 1.1.4 Adaptation to combined nitrogen deprivation 1.1.4.1 General response When cyanobacteria sense a lack of combined nitrogen, a general response takes place to try to overcome such lack of nutrients. This response is under control of NtcA and common to all cyanobacteria, whether or not they are ultimately able to fix atmospheric nitrogen, a process that in some strains requires differentiation of heterocysts (see below). This general response involves induction of the expression of transporters of ammonium (Flaherty et al., 2011; Mitschke et al., 2011b) nitrate/nitrite (Frías et al., 1997) and urea (Valladares et al., 2002). Expression of the genes required for the assimilation of these alternative sources is also induced, in order to use traces of nitrate, nitrite, urea, amino acids or even ammonium that may remain in the media. In addition, the activity of glutamine synthetase (GS, encoded by glnA) is induced by several mechanisms. NtcA promotes the transcription of glnA (Herrero et al., 2001), inhibits the transcription of the inactivating factors of glutamine synthetase IF7 and IF17 (encoded by gifA and gifB) (García-Domínguez et al., 2000) and promotes the transcription of NsiR4, a sRNA that inhibits translation of IF7 (Klähn et al., 2015). In addition, a glutamine riboswitch regulates the expression of the other inactivating factor (IF17) (Klähn et al., 2018). In the absence of combined nitrogen, cyanobacteria activate the degradation of proteins (Fleming and Haselkorn, 1974). Therefore, another aspect of the general response to nitrogen deprivation is the recycling of the amino acids present in the proteins that constitute the phycobilisomes, that represent more than 40% of the cellular protein and constitute a major storage of carbon and nitrogen. Under nitrogen deprivation, these supramolecular complexes are dismantled, in a process known as “bleaching” (Grossman et al., 1993b; 1993a), and their amino acids are recycled and incorporated in other metabolic pathways. NblA (non-bleaching phenotype A) is a small protein that plays an important role in this process mediating the interaction between the phycobiliproteins and the proteases that degrade them (Collier and Grossman, 1994; Karradt et al., 2008; Sendersky et al., 2014). Phycobilisome degradation is a general response common to all cyanobacteria but differs in strains that are eventually able to fix atmospheric nitrogen from those that are not. Whereas in non-N2-fixing cyanobacteria, phycobilisome degradation persists in time, in N2-fixing cyanobacteria, such as our model organism, phycobilisome degradation is only a transient response until heterocysts start to fix N2 (Baier et al., 2004). 15 Introduction 1.1.4.2 Heterocyst differentiation Some cyanobacteria have the capacity of fix atmospheric nitrogen by nitrogenase, a multiproteic complex that reduces N2 to ammonium in a reaction that consumes a high amount of energy (Seefeldt et al., 2009). Because nitrogenase is extremely sensitive to oxygen, cyanobacteria have developed several strategies to solve the incompatibility between nitrogen fixation and oxygenic photosynthesis (Fay, 1992). Non heterocyst-forming cyanobacteria usually alternate these processes, fixing nitrogen during darkness periods and performing oxygenic photosynthesis during daytime (Huang et al., 1999; Reade et al., 1999; Misra and Tuli, 2000). Heterocyst-forming cyanobacteria can spatially separate these two processes. The differentiation of some vegetative cells into heterocysts in a semi-regular pattern generates a multicellular organism with two specialized cell types, vegetative cells and heterocysts. Vegetative cells perform photosynthesis and CO2 fixation whereas heterocysts provide a microoxic environment for N2 fixation. So, both cell types need to exchange carbon and fixed nitrogen in order to sustain the growth of the filament as whole (Figure 6) (Muro-Pastor and Hess, 2012). Figure 6. Heterocyst differentiation and metabolic exchange between vegetative cells and heterocysts. (A) Filaments of Nostoc sp. PCC 7120 growing on top of plates without combined nitrogen differentiate new heterocysts (white triangles) between two mature heterocysts (black triangles). (B) Approximate timing of heterocyst differentiation. (C) Division of labour and metabolic exchange between heterocysts (nitrogen fixation) and vegetative cells (oxygenic photosynthesis and carbon fixation). Taken from Muro-Pastor and Hess, 2012. 16 Introduction Heterocysts are terminally differentiated non-dividing cells devoted to nitrogen fixation. Its morphology and physiology have evolved to generate a micro-oxic environment that protects and supports nitrogenase activity. Heterocyst differentiation involves a complex developmental program that includes morphological changes, metabolic reprogramming, regulation of pattern and timing of differentiation and commitment (Flores et al., 2019). Heterocyst are morphologically different from vegetative cells. They are larger and have a thick envelope that restricts gas diffusion and constraints the contact area between heterocysts and vegetative cells (Fay, 1992). Their thylakoid membranes are reorganized at the poles in a structure called “honeycomb” (Lang and Fay, 1971), they don’t have carboxysomes (Winkenbach and Wolk, 1973; Stewart and Codd, 1975) and they accumulate fixed nitrogen in granules of cyanophycin localized at the poles (Sherman et al., 2000). The micro-oxic environment of heterocysts is achieved through several strategies. Heterocyst envelopes contain two extra-layers outside the cell wall; an outer polysaccharide layer (HEP) (Bryce et al., 1972) and an inner glycolipid layer (HGL) (Cardemil and Wolk, 1979). Genes required for the synthesis of the HEP and HGL layers are mainly encoded in clusters called “gene expression islands”. Most of the genes related to synthesis of the HEP layer are encoded in one expression island (alr2825-alr2841) (Huang et al., 2005), but some are also outside of the cluster (Maldener et al., 2003; Wang et al., 2007), and are transcribed at early stages in the heterocyst differentiation process (Ehira et al., 2003; Flaherty et al., 2011; Mitschke et al., 2011b). Most of the genes involved in the biosynthesis and deposition of the components of the HGL layer are clustered in another “gene expression island” (all5341- all5359) (Fan et al., 2005; Awai and Wolk, 2007) and are transcribed later during the differentiation process (Ehira et al., 2003; Flaherty et al., 2011). Thus, the HEP layer is deposited first, and at an early stage, and the HGL layer is deposited beneath it. In addition to the specialized heterocyst envelope, several enzymatic activities contribute to reducing the intracellular concentration of oxygen. Heterocysts have a different organization of their thylakoid membranes. Vegetative cells usually have their thylakoid membranes concentrically arranged in the periphery whereas, in heterocysts, membranes are localized near the poles (Wildon and Mercer, 1963; Lang and Fay, 1971). Terminal respiratory oxidases localized in these membranes (Cox oxidases) (Valladares et al., 2003) and cytoplasmic flavodiiron proteins (Flv1B and Flv3B) (Ermakova et al., 2014) consume O2 in the heterocyst, contributing to its micro-oxic ambient. In addition, a specific peroxidase that is induced in heterocysts (RbrA) protects these cells from the Reactive Oxygen Species (ROS) (Zhao et al., 2007). Finally, photosystem II (PSII) is not functional in heterocysts (Donze et al., 1972) and there is no photosynthetic CO2 fixation because of lack of RuBisCO 17 Introduction (Winkenbach and Wolk, 1973; Stewart and Codd, 1975). However, given that nitrogenase requires a high amount of energy and reductants, heterocysts maintain photosystem I (PSI) and perform photophosphorylation (Magnuson and Cardona, 2016). Nitrogenase is a multiproteic complex. Genes encoding factors involved in the maturation of nitrogenase, its enzymatic activity or its specific electron donors are encoded in a large cluster (nifB-fdxN-nifS-nifU-nifH-nifD-nifK-nifE-nifN-nifX-all1435-asl1434-nifW-hesA-hesB- fdxH) that is transcribed in the last steps of heterocyst differentiation (Ehira et al., 2003; Flaherty et al., 2014). Nitrogenase reduces N2 to ammonium that is later assimilated by glutamine synthetase, producing glutamine (Wolk et al., 1976). Because heterocysts don’t have GOGAT activity (Thomas et al., 1977; Martín-Figueroa et al., 2000), glutamine must be transferred to vegetative cells in exchange for glutamate in order to close the GS-GOGAT cycle. So, heterocysts feed vegetative cells with glutamine and other nitrogen compounds (Burnat et al., 2014) and in return, vegetative cells feed heterocysts with reduced carbon compounds. For example, vegetative cells feed heterocysts with sucrose that is hydrolysed in the heterocysts by invertase (InvB) (López-Igual et al., 2010; Vargas et al., 2011), generating fructose and glucose that can be later oxidised by the oxidative pentose-phosphate cycle generating energy and reductant (Magnuson and Cardona, 2016). This exchange of metabolites could not be possible without the structures and pathways of exchange commented in section 1.1.2. The regulatory cascade that orchestrates heterocyst differentiation is complex and not fully understood. The two essential transcription factors involved in differentiation of heterocysts are NtcA and HetR. Whereas NtcA is the global regulator of nitrogen assimilation (Herrero et al., 2001; Herrero et al., 2004), HetR is specifically involved in cellular differentiation (Buikema and Haselkorn, 1991; Black et al., 1993). At the initiation of heterocyst differentiation, induction of the expression of ntcA and hetR depends on each other (Muro-Pastor et al., 2002) and leads to an overproduction of both factors in cells fated to become heterocysts (Black et al., 1993; Olmedo-Verd et al., 2006). This specific accumulation of HetR and NtcA in some cells may trigger the regulatory cascade that activates the differentiation program. The NtcA-dependent induction of hetR may be mediated by NrrA, a response regulator-like factor whose expression is induced by NtcA (Muro-Pastor et al., 2006) that binds in vitro to the promoter region of hetR (Ehira and Ohmori, 2006). The expression of hetR is regulated a several levels. hetR is transcribed from a complex promoter that includes constitutive TSS but also one TSS specifically induced in cells becoming heterocysts (Rajagopalan and Callahan, 2010). HetR is a DNA-binding protein (Kim et al., 2011; 18 Introduction Kim et al., 2013) for which a binding in vitro to its own promoter region and those of several genes specifically induced in heterocyst has been reported (Huang et al., 2004). HetR positively regulates its own expression and the expression of its inhibitors, patS (Yoon and Golden, 1998) and hetN (Callahan and Buikema, 2001), both essential for the maintenance of the pattern of heterocysts along the filaments (Borthakur et al., 2005; Rivers et al., 2018). PatS and HetN contain a RGSGR pentapeptide that binds to HetR, blocks its DNA-binding activity (Huang et al., 2004) and signals its degradation (Risser and Callahan, 2009). The diffusion of PatS (Corrales- Guerrero et al., 2013) or a HetN-derived product may generate a gradient of morphogen that regulates the expression of hetR in the semi-regular pattern that intercalates heterocyst and vegetative cells. HetR is also an auto-proteolytic protein (Zhou et al., 1998) and is degraded by a protease (HetF) (Wong and Meeks, 2001; Risser and Callahan, 2008). In addition, HetR seems to interact with HetP and HetP homologs, proteins that regulate the establishment of the commitment point of heterocysts (Videau et al., 2016a). After this point, pro-heterocysts become irreversibly committed to differentiate while their division machinery gets downregulated at transcriptional and translational levels (Kuhn et al., 2000; Wang and Xu, 2005). NtcA and HetR also regulate the early expression of devH, a transcriptional activator of the CRP family that regulates the expression of hgl genes (Hebbar and Curtis, 2000; Ramírez et al., 2005). At an advanced state of maturation, the transcriptional regulator CnfR (previously known as PatB) is specifically induced in heterocysts (Jones et al., 2003). CnfR may be activated by the low oxygen tension inside of heterocyst (Liang et al., 1993), it binds to the promoter region of the nif operon and promotes the expression of nitrogenase genes (Pratte and Thiel, 2016). In addition, the regulation of heterocyst differentiation involves many other mechanisms, including the use of alternative sigma factors that are specifically expressed in heterocysts at an early (sigC) or late (sigE) stage (Aldea et al., 2007; Muro-Pastor et al., 2017). Finally, several sRNAs that are specifically expressed in heterocysts have been identified (Ionescu et al., 2010; Mitschke et al., 2011b), suggesting a regulatory role of this type of molecules in the complex regulatory network leading to heterocyst differentiation. 19 Introduction 1.2 RNAs regulators in bacteria The central dogma of molecular biology postulated that the transfer of information in a biological system was unidirectional and finished with the synthesis of proteins (Crick, 1970). In this scenario, the RNA was considered a mere messenger in the transmission of information from the DNA to the synthesis of proteins. However, the use of next generation sequencing techniques has revealed an unexpected amount of non-coding transcripts, apart from ribosomal RNAs (rRNAs) or transfer RNAs (tRNAs). Although the function of some of these non-coding RNAs remains unknown, many of them regulate gene expression (Wade and Grainger, 2014). So, the ubiquity of these molecules and the wide variety of processes that they may regulate suggest an important role for these molecules in the mechanisms of genetic regulation. The mechanisms for the biogenesis of RNA regulators are diverse (Figure 7). Figure 7. Biogenesis of RNA regulators. (A-C) Antisense RNAs (asRNAs) may be complementary to the coding region, 5’UTR or 3’UTR of their target gene (A). Two mRNAs encoded in opposite orientation may regulate the expression of each other, a type of regulation known as “excludons” (B-C). (D-F) sRNAs may be generated in intergenic regions as independent genes (D), by processing of a mRNA by RNases (orange) (E) or from the 5’UTR, 3’UTR or even inside the coding region of a gene (F). DNA strands are indicated by parallel black lines. TSS are indicated by vertical arrows. Shine Dalgarno sequences (SD) are indicated by green squares. 20 Introduction One of the main groups of non-coding regulatory RNAs are antisense RNAs (asRNAs) (Lasa et al., 2012), that are transcribed in antisense orientation to the molecule that they regulate (Figure 7A-C). The other main group of non-coding regulatory RNAs are small RNAs (sRNAs), molecules with sizes between 50 and 350 nucleotides that are commonly transcribed from a different genomic region than the molecule(s) that they regulate (Storz et al., 2011). sRNAs may be transcribed from independent genes in intergenic regions (Figure 7D) or they could be generated by the processing of another RNA molecule (Figure 7E), as SdsN (Hao et al., 2016). In addition, they could be transcribed from TSS that are located in the 3’ UTR (UTR, untranslated region) of a gene, as DapZ (Chao et al., 2012), in the coding region of a gene, as MicL (Guo et al., 2014), or they could also share a promoter with a gene, as actuatons (Figure 7F) (Kopf et al., 2015a). RNA regulators usually exert a post-transcriptional regulation that complements the regulation carried out by transcription factors, playing a role in the fine-tuning necessary for the adaptation of living organism to an ever-changing environment (Wagner and Romby, 2015). The advantages of post-transcriptional regulation exerted by sRNAs versus classical regulation exerted by transcription factors have been widely reviewed (Figure 8) (Beisel and Storz, 2010; Mandin and Guillier, 2013; Wagner and Romby, 2015; Nitzan et al., 2017). Figure 8. Kinetic properties of sRNA regulation. (A) The regulation exerted by sRNA is faster than the regulation exerted by transcription factors. (B) The 1:1 interaction that occurs in many sRNA-mRNA interactions provokes an inverse relation between free sRNA and free mRNA target. (C) The inverse relation between free sRNA and mRNA target generates a threshold-linear response that filters weak or transient signals. The effective expression of a target does not start until it overpasses the transcription rate of the sRNA (*). After this point, the effective transcription rate of the target depends on the difference between the transcription rate of sRNA and target mRNA. Adapted from Nitzan et al., 2017. First, because sRNA expression only involves transcription whereas the expression of transcription factors involves transcription and translation, sRNA regulation may be faster than regulation exerted by transcription factors (Figure 8A). In addition, in the case of negative regulation, the binding of a transcription factor to the DNA prevents the transcription of new 21 Introduction mRNA molecules but the pool of available mRNA continues to be translated. In contrast, because of the direct regulation exerted by regulatory RNAs, sRNAs may destabilize the mRNA pool directly, achieving a faster regulation. Second, sRNA regulation seems to achieve an inverse relation between the amount of free sRNA and free mRNA target (Figure 8B). The interaction of a transcription factor with the DNA usually involves one or a few molecules of transcription factors that regulate the synthesis of many molecules of mRNA. However, the interaction of a sRNA with its target mRNA may involve the interaction of one molecule of sRNA per one molecule of mRNA, achieving the inverse correlation between free sRNA and free mRNA target. Third, this inverse correlation is used in some cases to achieve a threshold-linear response (Figure 8C). Some sRNA may be constitutively transcribed, leading to sequestration of every target mRNA molecule until the transcription rate of the target mRNA passes a certain threshold. However, when the transcription rate of the target mRNA overpasses the amount of sRNA, a linear response occurs that depends on the difference between the transcription rate of sRNA and target mRNA, a feature that could be used to filter weak or transient signals. All these features of sRNA regulation have been validated by several dynamic simulations (Shimoni et al., 2007; Mehta et al., 2008). Finally, the combination of sRNA regulation and the regulation exerted by transcription factors may generate complex regulatory circuits. One of the simplest regulatory loops emerges when a transcription factors regulates the expression of a target mRNA and the expression of a sRNA that, in addition, regulates the same target mRNA (Mangan and Alon, 2003). These coherent or incoherent loops, depending on the logical relationships between their components, are thought to be essential, for example, in the filtering of spurious stimulus or the generation of pulses of expression (Alon, 2007). The overlay of several loops and layers of regulation generates an intricate network that fine-tunes the physiological responses. In fact, some analysis suggest that the post-transcriptional regulation exerted by sRNAs may play an important part (even more than classical transcriptional regulation) in the determination of the final concentration of gene products (Vogel and Marcotte, 2012). 22 Introduction 1.2.1 Types of RNA regulators 1.2.1.1 RNA regulators encoded in cis RNA regulators that are transcribed from regions adjacent to the RNA molecule that they regulate are referred to as RNA regulators encoded in cis. The main group of cis encoded RNAs are asRNAs. These molecules are transcribed in antisense orientation to the mRNA target that they regulate, and, because of their genomic position, they show perfect, sometimes extensive, base complementarity to the coding region, 5’ UTR or 3’UTR of their mRNA target (Figure 7A). In some cases, the antisense regulation may be produced by the interaction of two mRNAs that regulate each other, a type of regulation referred to as “excludons” (Figure 7C-D) (Sesto et al., 2013; Georg and Hess, 2018). The mechanisms of asRNA regulation are based on the antisense orientation of the interacting molecules. asRNAs may interfere in the transcription of their mRNA targets by transcription interference or they may post-transcriptionally regulate the stability or translation efficiency of their mRNA targets (Figure 9) (Thomason and Storz, 2010; Georg and Hess, 2011; 2018). Figure 9. Some asRNA regulatory mechanisms. The interaction by perfect base complementarity between an asRNA (red) and its target mRNA (blue) may provoke termination of transcription (A), mRNA degradation by RNases (orange) (B) or changes in the translation efficiency, generally by a block in the translation initiation mediated by interaction of the asRNA with the region containing the Shine-Dalgarno sequence (green) that prevents binding of the ribosome (grey) (C). Antisense transcription was first observed in the context of mechanisms controlling plasmid copy number or associated to transposons, phages or toxin-antitoxin systems (Wagner and Simons, 1994). Nowadays, the use of transcriptomic data reveals that antisense 23 Introduction RpaB and PsrR1 are part of a regulatory coherent loop that regulates acclimation to changing light conditions. Ncr0700/PmgR1 (photomixotrophic growth RNA 1) is another sRNA with an expression profile opposite to PsrR1. PmgR1 accumulates in the dark (Beck et al., 2014), regulates the accumulation of glycogen and is essential for mixotrophic growth (de Porcellinis et al., 2016). Photosynthesis is also regulated by asRNAs in Synechocystis sp. PCC 6803. The psbA2 and psbA3 mRNAs, encoding the D1 protein of the reaction center of PSII, are degraded by RNase E in the dark at the AU-rich motifs located in the 5’ UTRs of these mRNAs (Sakurai et al., 2012). The maintenance of the levels of psbA mRNAs in the light is achieved through a positive regulation exerted by two asRNAs complementary to the 5’UTR of these mRNAs. In the light, the asRNAs are upregulated, interact with the AU-rich motifs and prevent the cleavage of the mRNAs by RNase E (Sakurai et al., 2012). Flv4 and Flv2 are flavodiiron proteins that protect PSII against oxidative stress (Zhang et al., 2012) and IsiA is a chlorophyll binding protein that, under iron deficiency, forms an additional antenna around PSI (Chauhan et al., 2011). The expression of the operon flv4-2 and the isiA mRNA are negatively regulated by the asRNAs As1_flv4 (Eisenhut et al., 2012) and IsrR (Dühring et al., 2006), respectively. In both cases, the constitutive expression of the asRNAs and the co-degradation with their target mRNAs establish a certain threshold for the protein synthesis (Legewie et al., 2008), ensuring that Flv4 and Flv2 or IsiA are only translated under low CO2 conditions or iron deficiency, respectively. Some asRNAs and sRNAs have also been described in the heterocyst-forming strain Nostoc sp. PCC 7120. One of the first asRNA described in cyanobacteria was an antisense complementary to furA (all1691). This antisense is produced from the transcription of alr1690, an adjacent gene in antisense orientation, it partially covers the sequences encoding FurA and post-transcriptionally regulates the expression of this transcriptional regulator involved in iron metabolism (Hernández et al., 2006). Because the adaptation to nitrogen deprivation in Nostoc sp. PCC 7120 is a wellstudied process, most sRNAs described in this organism so far are regulated by nitrogen availability. Some sRNAs, as NsiR3 (Mitschke et al., 2011b), NsiR4 (Klähn et al., 2015) or NsrR1 (Álvarez-Escribano et al., 2018), may be involved in the general response to nitrogen deprivation whereas transcription of other sRNAs, such as NsiR1 (Ionescu et al., 2010) is specifically induced in heterocyst, suggesting it might participate in the regulation of heterocyst differentiation. NsiR1 is transcribed from multiple copies in the genome of heterocyst-forming cyanobacteria 30 Introduction (Ionescu et al., 2010) and is expressed at a very early stage in the differentiation process (Muro- Pastor, 2014) but its function has remained elusive. NsiR4 (nitrogen stress-induced RNA 4) is a sRNA that participates in the regulation of the assimilation of nitrogen through the post-transcriptional regulation of IF7, one of the inactivating factors of glutamine synthetase in Synechocystis sp. PCC 6803 (Klähn et al., 2015). Because the expression of NsiR4 and IF7 are regulated by NtcA, NsiR4 is involved in a coherent regulatory loop that regulates one of the key enzymes in the assimilation of nitrogen (Klähn et al., 2015). NsrR1 (nitrogen stress-repressed RNA 1) is a sRNA directly repressed by NtcA under nitrogen deprivation and interacts with the 5’UTR of the nblA mRNA promoting its degradation (Álvarez-Escribano et al., 2018). As it is the case of NsiR4, because NtcA also regulates the expression of nblA, NsrR1, NtcA and NblA also form a coherent regulatory loop that regulates the recycling of phycobilisomes under changes in nitrogen availability (Álvarez-Escribano et al., 2018). Some sRNAs and asRNAs are specifically induced in heterocyst and could be involved in the regulation of heterocyst differentiation. A recent report identified an asRNA (As_glpX) specifically induced in heterocysts that regulates the degradation of glpX mRNA (Olmedo-Verd et al., 2019). glpX encodes sedoheptulose-1,7-bisphosphatase/fructose-1,6-bisphosphatase (SBPase), a key enzyme in the Calvin cycle. Thus, this asRNA contributes to the metabolic reprogramming of heterocyst by locally diminishing the levels of one of the enzymes necessary for CO2 fixation. The main goal of this Thesis has been the identification and characterization of asRNAs and sRNAs in Nostoc sp. PCC 7120 that may be involved in the regulation of heterocyst differentiation or in general physiological process that may affect this differentiation process. 31 2. OBJECTIVES Objectives 2. OBJECTIVES The aim of this Thesis is the identification and characterization of sRNAs and asRNAs that may be involved in the regulation of the responses to nitrogen deprivation in Nostoc sp. PCC 7120. The following specific objectives are addressed: 1. Identification of new sRNAs by a bioinformatic approach. 2. Functional characterization of Yfr1, a phylogenetically conserved sRNA. 3. Generation of a co-expression network, which includes non-coding RNAs, to dissect the complex changes in the transcriptome taking place during heterocyst differentiation. 4. Functional characterization of NsiR1, a sRNA specifically induced in heterocysts. 35 3. SUMMARY OF RESULTS Summary of results 3. SUMMARY OF RESULTS Regulation of gene expression cannot be understood without the participation of RNA regulators. Because these molecules are difficult to detect and usually exert a subtle posttranscriptional regulation, their functions have remained elusive for many years. The emergence of “omic” technologies and the use of global approaches have increased the number of new RNA regulators identified. The global objective of this Thesis is the use of computational approaches for the identification of new sRNAs and asRNAs in the heterocyst-forming cyanobacterium Nostoc sp. PCC 7120 and the functional characterization of some of those RNAs. The content of this dissertation is structured in four chapters, two of them already published as research articles. 3.1 Chapter I: A computational approach for the identification of conserved sRNAs in heterocyst-forming cyanobacteria. Brenes-Álvarez, M., Olmedo-Verd, E., Vioque, A., & Muro-Pastor, A. M. (2016). Identification of conserved and potentially regulatory small RNAs in heterocystous cyanobacteria. Front Microbiol, 7, 48. doi:10.3389/fmicb.2016.00048 Previous computational approaches to the identification of sRNAs had been focused in the conservation of sequences located in intergenic regions flanked by predicted transcriptional signals (-10, -35 boxes) and putative Rho-independent terminators. Whereas the software for the prediction of Rho-independent terminators (a stem loop followed by a poly uracil stretch) is quite accurate, the prediction of promoter motifs and other transcriptional signals is not very accurate. In our work, we overcome this drawback by using an experimental dataset from a previous dRNAseq study that determined most of the TSS in Nostoc sp. PCC 7120 (Mitschke et al., 2011b). The use of this dataset together with sequence information from 89 cyanobacterial genomes allowed us to predict 327 putative sRNAs. Basically, we searched in the genome of Nostoc sp. PCC 7120 for sequences located in intergenic regions that were flanked by a TSS and a Rho-independent terminator and with a size range of 45 - 350 nucleotides. Sequences that fulfilled those criteria were extracted and 39 CHAPTER I. A computational approach for the identification of conserved sRNAs in heterocyst-forming cyanobacteria Brenes-Álvarez, M., Olmedo-Verd, E., Vioque, A., and Muro-Pastor, A.M. (2016). Identification of conserved and potentially regulatory small RNAs in heterocystous cyanobacteria. Frontiers in Microbiology 7, 48. doi: 10.3389/fmicb.2016.00048. ! 3.2 CHAPTER II Yfr1, a widely conserved sRNA, regulates the integrity of the cell wall and its remodeling during heterocyst differentiation Integrity of the cell wall and its remodeling during heterocyst differentiation is regulated by phylogenetically conserved sRNA Yfr1 in Nostoc sp. PCC 7120 Manuel Brenes-Álvarez, Agustín Vioque and Alicia M. Muro-Pastor# Instituto de Bioquímica Vegetal y Fotosíntesis, Consejo Superior de Investigaciones Científicas and Universidad de Sevilla, E-41092 Sevilla, Spain #Corresponding author: Alicia M. Muro-Pastor, Instituto de Bioquímica Vegetal y Fotosíntesis, Consejo Superior de Investigaciones Científicas and Universidad de Sevilla, E-41092 Sevilla, Spain Tel. +34-954489521 Email: [email protected] ORCID iD: 0000-0003-2503-6336 Running Title: Yfr1 sRNA regulation in Nostoc 7120 Word count (abstract): 233 Word count (text): 4892 69 ABSTRACT Yfr1 is a strictly conserved small RNA in Cyanobacteria. A bioinformatic prediction to identify possible interactions of Yfr1 with mRNAs was carried out using the sequences of Yfr1 from several heterocyst-forming strains, including Nostoc sp. PCC 7120. The results of the prediction were enriched in genes encoding outer membrane proteins and enzymes related to peptidoglycan biosynthesis and turnover. Heterologous expression assays in Escherichia coli demonstrated direct interactions of Yfr1 with mRNAs of eleven of the candidate genes. The expression of ten of them (alr2458, alr4550, murC, all4829, all2158, mraY, alr2269, alr0834, conR, patN) was repressed by interaction with Yfr1, whereas the expression of amiC2, encoding an amidase, was increased. The interactions between Yfr1 and the eleven mRNAs were confirmed by site-directed mutagenesis of Yfr1. Furthermore, a Nostoc strain with reduced levels of Yfr1 had higher amounts of mraY and murC mRNAs, supporting a role for Yfr1 in the regulation of those genes. Nostoc strains with either reduced or increased expression of Yfr1 showed anomalies in cell wall completion and were more sensitive to vancomycin than the wild type strain. Furthermore, growth in the absence of combined nitrogen, that involves differentiation of heterocysts, was compromised in the strain overexpressing Yfr1, and filaments were broken at the connections between vegetative cells and heterocysts. These results indicate that Yfr1 is an important regulator of cell wall homeostasis and correct cell wall remodeling during heterocyst differentiation. IMPORTANCE Bacterial small RNAs (sRNAs) are important players affecting the regulation of essentially every aspect of bacterial physiology. The cell wall is a highly dynamic structure that protects bacteria from their fluctuating environment. Cell envelope remodeling is particularly critical for bacteria that undergo differentiation processes such as spore formation or differentiation of heterocysts. Heterocyst development involves the deposition of additional layers of glycolipids and polysaccharides outside the outer membrane. Here we show that a cyanobacterial phylogenetically conserved small regulatory RNA, Yfr1, coordinates expression of proteins involved in cell wall-related processes, including peptidoglycan metabolism and transport of different molecules, as well as expression of several proteins involved in heterocyst differentiation. 70 INTRODUCTION Bacterial cell envelopes are multi-layered structures that delimit the interior of the cell from its environment. The essential component of bacterial cell walls is peptidoglycan, a strong and flexible mesh that protects the cell against osmotic pressure and contributes to the shape of the cell. According to the architecture of the envelopes, bacteria are classified in two groups. Gram-positive strains have a thick peptidoglycan layer (30 to 100 nm) surrounding the cytoplasmic membrane, while Gram-negative strains have a thin peptidoglycan layer (only a few nm) between the inner, cytoplasmic membrane, and a second, outer membrane (1). Cyanobacteria are Gram-negative. However, the peptidoglycan layer between the inner and outer membranes is relatively thick (15-30 nm in filamentous strains) and characterized by extensive crosslinking, rather resembling the architecture of Gram-positive bacteria (2). The sacculus of peptidoglycan is a dynamic structure that must adapt to the growth of the cells, the separation of daughter cells during cell division, the turnover of peptidoglycan or the assembly of large trans-envelope complexes (e. g. secretion systems). In addition, cellular differentiation processes that affect cell envelopes involve peptidoglycan remodeling, performed by different murein hydrolases, such as N-Acetylmuramoyl-L-Ala amidases (Ami enzymes), reviewed in (3). Nostoc sp. PCC 7120 is a filamentous cyanobacterium that under nitrogen deprivation differentiates heterocysts (specialized cells devoted to N2 fixation) in a semiregular pattern (4). Heterocyst differentiation involves biochemical and morphological changes that provide a micro-oxic environment for nitrogenase, an O2-labile enzyme. In addition to the inactivation of the O2-producing photosystem II and the increase in the O2 consumption rate, heterocysts have a special cellular envelope with two extra layers, an external one composed of polysaccharides (HEPs) and a laminated, internal layer composed of glycolipids (HGLs) that acts as a barrier to gas diffusion (5). Both heterocyst-specific layers are deposited outside the outer membrane of these cells. Under nitrogen-fixing conditions there is a metabolic division of labor between vegetative cells and heterocysts in cyanobacterial filaments. Heterocysts feed vegetative cells with fixed nitrogen and obtain fixed carbon in return. Growth of nitrogen-fixing filaments depends on the transport and exchange of metabolites between vegetative cells and heterocysts (4). One route for this exchange of metabolites could consist of diffusion from cytoplasm to cytoplasm through septal junctions, protein structures that allow the 71 intercommunication between the cytoplasms of adjacent cells (6-9). These structures traverse the septal peptidoglycan disks through perforations called “nanopores” (7) that may be made by murein hydrolases. Mutants in murein amidases AmiC1 (alr0092) (10) and AmiC2 (alr0093, hcwA) (11, 12) show a significant reduction in the number of nanopores and defects in diazotrophic growth. In addition, mutants in peptidoglycan synthesis enzymes MurC (alr5065) and MurB (alr5066) (13) and in some penicillin binding proteins (all2981, alr4579 and alr5051) (14, 15) show alterations in heterocyst differentiation and growth defects in media without combined nitrogen. All this evidence points to synthesis and remodeling of peptidoglycan being essential for proper heterocyst differentiation and diazotrophic growth. In fact, several genes related to peptidoglycan metabolism are transcribed from HetR-dependent, heterocyst-specific promoters (16, 17). The maintenance and assembly of the outer membrane relies on proteins of the Omp85 family. Omp85 is essential for outer membrane maturation in E.coli (18) and a mutant of alr2269 (encoding a Omp85 homolog in Nostoc sp. PCC 7120) showed a disturbed outer membrane and, as a consequence, higher sensitivity to harmful substances, such as erythromycin (19). Molecules enter the periplasm through different transporters, such as porinlike proteins and TonB-dependent transporters. Porins usually allow the diffusion of hydrophilic molecules of a size up to 600 Da with low selectivity (20). The genome of Nostoc sp. PCC 7120 encodes a general porin (Alr0834), OprB-like porins (Alr4550 and All4499) (21), and several TonB-dependent transporters (All2158, All3310, All4026) (22). Yfr1 is one of the first small RNAs (sRNA) identified in Cyanobacteria and was initially described in unicellular picocyanobacteria (23). Bioinformatic prediction and the use of a heterologous reporter system in E.coli demonstrated that Yfr1 can interact with the mRNAs encoding two porins in Prochlorococcus MED4 (24). Recently, a global approach allowed the identification of the targetome of Yfr1 from Prochlorococcus MED4 (25). In Synechococcus elongatus PCC 6301, Yfr1 is highly expressed, with slight abundance changes in cells exposed to high salt stress or oxidative stress. A Yfr1 mutant showed reduced growth under iron limitation, high salt stress or oxidative stress (26). Yfr1 has been identified in all cyanobacterial genomes analyzed, from the minimal genomes of unicellular strains of the Prochlorococcus-Synechococcus lineages to much larger genomes of complex filamentous strains, such as Nostoc, able to undergo cellular differentiation processes (27). This broad occurrence suggests a widely conserved function for Yfr1. 72 Interestingly, a fully conserved sequence motif is present in all Yfr1 homologs across the different cyanobacterial clades (27, 28). We have identified and verified the interaction of Yfr1 with several mRNAs whose products are involved in peptidoglycan metabolism and envelope biogenesis and maintenance in Nostoc sp. PCC 7120, including several proteins required for proper differentiation of functional heterocysts and therefore N2 fixation. Our results suggest that Yfr1 may regulate the composition and remodeling of envelopes of heterocyst-forming cyanobacteria. RESULTS Validation of putative Yfr1 targets in E. coli. According to previous studies, Yfr1 is expressed constitutively, with only slight expression changes under different growth conditions in the unicellular cyanobacteria Synechococcus elongatus PCC 6301 (26) or Prochlorococcus MED4 (25). We tested expression of Yfr1 in Nostoc sp. PCC 7120 (Fig. S1). Similar to the observations made in the case of unicellular strains, Yfr1 had a relatively strong expression and its levels changed only slightly due to nitrogen deficiency or high light stress, two conditions that lead to pronounced changes in cyanobacterial gene expression. As previously shown, Yfr1 accumulates in the form of two transcripts with slightly different size (28). The sequence of Yfr1 from Nostoc sp. PCC 7120 is shown in Fig. 1A. Because prediction of sRNA targets can be improved when comparative phylogenetic information is taken into account (29), we used CopraRNA, a software to predict potential interactions between sRNAs and mRNA targets that are conserved among a set of organisms (29, 30). Using the sequences of Yfr1 homologs from ten different heterocystous strains (see Materials and Methods for details), the resulting list of predicted targets (Table S1) showed a significant enrichment in transporters, enzymes related to cell wall synthesis or remodeling and proteins located in the outer membrane. Amongst the predicted targets were several mRNAs encoding enzymes involved in peptidoglycan metabolism such as alr2458 (alanine racemase), alr5065 (murC), all4316 (mraY), alr0093 (hcwA, amiC2) or all3826 (a penicillin binding protein, PBP). Several of the predicted targets corresponded to proteins known to be located in the outer membrane (21), including alr0834, encoding a homolog of the two previously validated targets of Yfr1 in Prochlorococcus (24), alr4550 and all4499 (OprB-like porins), all2158 and all4026 (TonB dependent transporters), proteins related to the biogenesis of outer membrane (alr2269, 73 envelopes as well as for the proper communication between vegetative cells and heterocysts (10-12, 14, 15). In order to analyze completion of heterocyst differentiation in strains OE_Yfr1 and OE_as_Yfr1, we tested their growth in plates containing different nitrogen sources (Fig. 5A). Whereas strain OE_as_Yfr1 showed no difference with respect to the OE_C strain, strain OE_Yfr1 grew worse than the control strain in plates without combined nitrogen, a nutritional condition that requires differentiation of functional nitrogen-fixing heterocysts. Strain OE_Yfr1 was unable to grow in liquid media without combined nitrogen (data not shown). The visualization of filaments of strain OE_Yfr1 streaked on top of plates of media lacking combined nitrogen showed patterned differentiation of heterocysts but the filaments appeared broken between heterocysts and adjacent vegetative cells (Fig. 5B). Figure 5. Growth of strains with altered levels of Yfr1. (A) Growth of OE_C, OE_Yfr1 and OE_as_Yfr1 strains in media lacking nitrogen (N2), or containing nitrate (NO3-) or ammonium (NH4+). Pictures were taken after 10 days of incubation at 30ºC. (B) Bright field images of filaments from the OE_C strain and from two independent clones of strain OE_Yfr1 streaked on top of BG110 plates. The broken connections between heterocysts and vegetative cells are indicated with black triangles. 80 DISCUSSION Yfr1 is an sRNA conserved in all cyanobacterial genomes (27, 28). The conservation of an sRNA in organisms with such a wide variety of morphologies, ecological niches and developmental processes, may suggest a regulatory function of a general aspect of the physiology of cyanobacteria. The prediction of targets by CopraRNA (29, 30) for Yfr1 homologs in heterocyst-forming cyanobacteria showed a remarkable enrichment in transporters and proteins located in the outer membrane as well as proteins involved in the synthesis or remodeling of peptidoglycan (Table S1). Target prediction was performed using genomes of heterocystous strains and in fact, among the TOP 50 predicted targets there are some genes of unknown function, such as alr4714, asl4743, all0997 and alr0255, that appear conserved only in this cyanobacterial clade (39). Using a heterologous assay in E. coli we have demonstrated the effect of Yfr1 on expression of eleven selected predicted targets and have verified that Yfr1 mostly downregulates the expression of the targets (ten cases), but it can also exert a positive regulation, as in the case of alr0093 (Fig. 1B). We could demonstrate that regulation was exerted via base-pairing of the highly conserved motif of Yfr1 with the mRNAs (Fig. 1 and 2). Most of the predicted interaction sites were located in the translation initiation region (Fig. S2). However, the interaction between Yfr1 and alr0093 was located far upstream from the start codon (Fig. S2), opening the possibility that the positive regulation of this particular mRNA is operated through a conformational change in its 5’-UTR that might improve translation. Using compensatory mutations, we have mapped the interactions between Yfr1 and the 5’-UTR of all4316 and alr5065 to the ribosome binding region (Fig. 2). These interactions are consistent with a negative regulation based on interference with ribosome access. In fact, the reduction in the corresponding RNA levels is only around 30-50% (Fig. S3A-B) whereas the fluorescence of translational reporters fused to the 5’ UTRs of all4316 and alr5065 is reduced more than 90% in the presence of Yfr1 (Fig. S3C), further suggesting a mechanism involving translational interference rather than alteration of mRNA stability. In order to assess the effects of Yfr1 in Nostoc, we overexpressed Yfr1 and anti-Yfr1 RNA under the strong and constitutive trc promoter (Fig. 3). Indeed, by analyzing Nostoc strains with altered levels of Yfr1 we observed a higher accumulation of full-length mRNAs for all4316 and alr5065 in strain OE_as_Yfr1 (Fig. 3C-F), that is depleted of Yfr1 (Fig. 3B), than in the control OE-C strain. This result was consistent with the negative regulation exerted by Yfr1 on these targets as validated in the E. coli system (Fig. 2). 81 Vancomycin binds to the D-alanyl-D-alanine terminus of nascent glycan chains preventing the cross-linking between two glycan strands (40). Strains OE_Yfr1 and OE_as_Yfr1 were more sensitive to vancomycin than strain OE_C strain (Fig. 4A), suggesting that both strains have alterations in peptidoglycan integrity, consistent with altered amounts of murC and mraY mRNAs. In addition, the poor growth of the OE_Yfr1 strain in plates containing SDS (a detergent) or erythromycin (a macrolide antibiotic) suggested additional defects in the integrity and permeability of the envelopes of this strain. This result is consistent with a negative regulation of Yfr1 on alr2269 mRNA (Omp85), since an alr2269 mutant showed greater sensitivity to SDS and erythromycin (19). When we measured the incorporation of Van-FL, a fluorescent derivative of vancomycin, we observed opposite phenotypes for the OE_Yfr1 and OE_as_Yfr1 strains (Fig. 4B). OE_Yfr1 had wider, frequent nascent septa in the middle of cells that had not finished its previous division whereas OE_as_Yfr1 had narrower septa and only some new septa were starting its completion. These results might be consistent with altered peptidoglycan synthesis and remodeling mediated by a negative regulation by Yfr1 of all4316 (murC) and alr5065 (mraY) and a positive regulation of alr0093 (amiC2). Modulation of the amounts of the corresponding enzymes could facilitate a faster completion of the peptidoglycan layer in OE_as_Yfr1 and a slower completion of the peptidoglycan layer in OE_Yfr1 strain, although these strains could differ also in other aspects of cell division. In fact, ConR, a protein of the LytR-CpsA-Psr superfamily involved in septum formation (33), is one of the targets of Yfr1 verified in this work (Fig. 1). Heterocyst differentiation involves important morphological changes in which synthesis and remodeling of peptidoglycan play an important role. Two of the validated targets of Yfr1 (amiC2 and murC) are necessary for proper diazotrophic growth (11-13). In addition, mraY (also among the targets of Yfr1 validated in this study) has a complex promoter region in which one transcription start site is specifically upregulated in heterocysts (41), suggesting a role of this enzyme in heterocyst differentiation. Strain OE_Yfr1 was unable to grow in liquid medium without combined nitrogen and grew very poorly on solid medium without combined nitrogen (Fig. 5A). Filaments of OE_Yfr1 plated on top of medium without combined nitrogen showed that although the strain was able to differentiate heterocysts with a normal pattern, the heterocyst-vegetative cells connections appeared frequently disrupted (Fig. 5B). These results suggest that the regulation exerted by Yfr1 at the level of a general aspect of the physiology of Nostoc sp. PCC 7120 (bacterial envelopes), is also critical for the differentiation of functional mature heterocyst. 82 Finally, expression of most previously studied cyanobacterial sRNAs, such as NsiR1 (42, 43), NsiR4 (44), IsaR1 (45), PsrR1 (46) or NsrR1 (47) transiently changes in response to certain environmental conditions, including light or availability of different nutrients. Expression of Yfr1 however is relatively high in Nostoc sp. PCC 7120 under standard laboratory conditions and its expression did not significantly change in response to high light or nitrogen availability, the two conditions we tested (Fig. S1). In Synechococcus elongatus PCC 6301, Yfr1 accumulates up to 18,000 copies per cell and its expression barely changes under different stresses (26). Yfr1 expression only slightly changes under tested stress conditions in Prochlorococcus MED4 (25). Therefore, the question arises how the regulation exerted by Yfr1 on its targets is modulated. One possibility is that the regulatory effects of Yfr1 depend on its controlled sequestration by an RNA-binding protein or another sRNA that could act as a trap. In the unicellular cyanobacterium Prochlorococcus, the regulation of Yfr1 occurs through sequestration by another conserved sRNA, Yfr2, that contains a conserved region partially complementary to Yfr1 and whose accumulation was found to respond to changes in the nitrogen availability (25). In Nostoc, Yfr2 is also upregulated under nitrogen deprivation (16) and we have carried out EMSA assays showing that Nostoc Yfr2 also interacts with Yfr1 in a similar way to that described in Prochlorococcus (Fig. S4). Thus, under nitrogen deprivation or other stresses Yfr2 transcribed from one or several of the four repeats found in the genome of Nostoc sp. PCC7120 could bind to the conserved region of Yfr1 preventing the interaction of Yfr1 with its target mRNAs (Fig. 6). Overexpression of Yfr1 in strain OE_Yfr1 might be buffered by interaction with Yfr2, therefore leading to changes that are less evident than those observed in the E. coli system, in which Yfr2 is absent, while Yfr1 is expressed from a high copy number plasmid. The regulatory model proposed here implies that under nitrogen stress Yfr2 would reduce the inhibitory effect of Yfr1 on a number of proteins required for cell wall changes occurring during heterocyst development. It is possible that additional regulatory mechanisms ensure specific enrichment of those proteins in heterocysts, as discussed above for mraY. The results presented here shed light on a general regulatory network that assures that proper amounts of different proteins related to cell wall biosynthesis and remodeling are present in Nostoc cells under different circumstances, and specifically during heterocyst development. The use of regulatory RNAs to control the amounts of outer membrane proteins is well described in Enterobacteria (48, 49). Enzymes involved in peptidoglycan biosynthesis have also been found to be controlled by a small RNA (50). Our work shows that also in Cyanobacteria similar mechanisms operate for the coordinated regulation of outer membrane proteins and peptidoglycan biosynthesis enzymes through Yfr1. 83 Finally, though Yfr1 is universally conserved in cyanobacteria, it appears that different regulatory functions are ascribed to Yfr1 in evolutionary distant strains. While in Prochloroccocus Yfr1 regulates genes involved in carbon and nitrogen metabolism (25), here we show that in a heterocyst-forming strain Yfr1 regulates genes related to cell wall synthesis and remodeling. Because most unicellular cyanobacteria seem to have evolved from non-heterocystous filamentous strains (51) it would be interesting to analyze the targetome of Yfr1 in nonheterocystous filamentous strains. Figure 6. Yfr1 regulatory network. The sRNAs Yfr1 and Yfr2 (orange ovals) and the proteins encoded by Yfr1 target mRNAs (blue ovals) are shown within a schematic drawing of a cell (the thick light gray or dark gray lines indicate cytoplasmic and outer membrane, respectively). Proteins are located in different areas depending on its function or the physiological processes in which they may be involved; top right, peptidoglycan biosynthesis or remodeling, bottom right, transporters, bottom left, proteins related to heterocyst differentiation and top left, other functions. Positive regulation is indicated by a blue arrow, negative regulation by black lines with blunt ends. The hypothetical negative regulation of Yfr1 by Yfr2 is indicated by a red line with blunt end and a red question mark. Possible regulation of Yfr2 by environmental signals is indicated with a dashed arrow. OM: outer membrane; CM: cytoplasmic membrane. MATERIAL AND METHODS Strains and growth conditions. Nostoc sp. PCC 7120 wild type and OE_C, OE_Yfr1 and OE_as_Yfr1 strains (Table S2) were grown photoautotrophically at 30ºC in BG11 medium (52) lacking NaNO3 but containing 84 3.5 mM NH4Cl and 7 mM N-[Tris(hydroxymethyl)methyl]-2-aminoethanesulfonic acid-NaOH buffer (pH 7.5). For Northern blot analysis of Yfr1 expression under different conditions, cultures of Nostoc sp. PCC 7120 were bubbled with an air/CO2 mixture (1% v/v) and grown photoautotrophically at 30ºC in BG11 medium (52) supplemented with 10 mM NaHCO3 (BG11C) lacking NaNO3 but containing 6 mM NH4Cl and 12 mM N-[Tris(hydroxymethyl)methyl]-2- aminoethanesulfonic acid-NaOH buffer (pH 7.5) (BG11C + NH4+). To induce nitrogen deficiency, filaments were collected by filtration, washed and resuspended in nitrogen-free BG11 medium containing 10 mM NaHCO3 (BG110C). High light stress (HL stress) was induced by increasing light intensity from 50 µE m-2 s-1 to 500 µE m-2 s-1. To test growth of strains with altered levels of Yfr1 under different conditions, liquid cultures of these strains growing in BG11 media were diluted to A750 0.17 and 10 µL of serial 5- fold dilutions were spotted on plates containing different nitrogen sources and/or different harmful compounds. OE_C, OE_Yfr1 and OE_as_Yfr1 strains were grown in the presence of appropriate antibiotics at the following concentrations: streptomycin (Sm) and spectinomycin (Sp), 2 µg/mL each (liquid medium) or 5 µg/mL each (solid medium). E. coli strains (Table S2) were grown in LB medium, supplemented with appropriate antibiotics (53). Generation of Nostoc strains with altered levels of Yfr1. Plasmids and oligonucleotides used in this work are described in Table S3 and Table S4, respectively. We have used pMBA37 (36) as a vector for overexpression of Yfr1 or an antisense to Yfr1 (as_Yfr1). pMBA37 contains the trc promoter and the T1 terminator of the rrnB gene of E. coli as transcriptional terminator and allows the overexpression of a cloned sequence flanked by NsiI-XhoI sites. Sequences corresponding to Yfr1 and as_Yfr1 (Yfr1 in reverse orientation) were amplified using genomic DNA as template and oligonucleotides 575 and 576 or 577 and 578, respectively. After digestion of the PCR products with NsiI and XhoI at the sites provided by the oligonucleotides, the fragments were cloned into NsiI-XhoI digested pMBA37, rendering pMBA48 (Yfr1) and pMBA49 (as_Yfr1). pMBA51 (a plasmid that overexpresses a control RNA corresponding only to the T1 terminator under the trc promoter) (36), pMBA48 and pMBA49 85 were introduced in Nostoc sp. PCC 7120 by conjugation (54) generating strains OE_C (control), OE_Yfr1 and OE_as_Yfr1, respectively (Table S2). RNA isolation, Northern blot analysis and primer extension assays. Total RNA was isolated using hot phenol as described (55) with modifications (27). Northern blot detection of Yfr1 was performed using 10% urea-polyacrylamide gels as described (56) and 7.5 µg of total RNA. Northern blot hybridization for mRNAs (all4316 and alr5065) was performed using 1% agarose denaturing formaldehyde gels and 10 µg of total RNA. All RNA samples were then transferred to Hybond-N+ membrane (GE Healthcare) with 20XSSC buffer. Strand-specific 32P-labeled probes were prepared with Taq DNA polymerase using a PCR fragment as template and oligonucleotides specified in Table S4 in a reaction with α-32P-dCTP and one single oligonucleotide as primer (corresponding to the complementary strand of the sRNA or mRNA to be detected). PCR fragments used as templates for Yfr1, all4316, and alr5065 probes were amplified from genomic DNA using oligonucleotides pairs 368/369, 430/431, and 448/449, respectively. Hybridization to rnpB (57) or 5S rRNA was used as a loading and transfer control. Fluorescent vancomycin conjugate staining. Filaments from 1 mL of liquid cultures growing in BG11 medium for 5 days were pelleted, washed, resuspended in 50 µL of PBS buffer, mixed thoroughly and incubated with vancomycin-FL, (Van-FL; BODIPYTM FL Conjugate, Invitrogen) or uncoupled vancomycin at 1 µg/mL for 1 hour in the dark. After the incubation, unlinked Van-FL or vancomycin were removed by washing twice with PBS buffer. Fluorescence was analyzed using a Leica HCX PLAN-APO 63X 1.4 NA oil immersion objective attached to a Leica TCS SP2 confocal laser-scanning microscope. Van-FL was excited at 488 nm by an argon ion laser and the fluorescent emission was monitored in the range of 500 to 530 nm. Samples incubated with uncoupled vancomycin were used to set a threshold to measure the specific fluorescence of Van-FL. Computational methods. Sequences of homologs of Yfr1 were taken from (27). CopraRNA (29, 30) was used for the prediction of the targets of Yfr1, using the homologs in the genomes of Nostoc sp. PCC 7120, Nostoc sp. PCC 7524, Anabaena variabilis ATCC 29413, Nostoc sp. PCC 7107, Calothrix sp. PCC 86 7507, Nostoc azollae 0708, Nostoc punctiforme PCC 73102, Calothrix sp. PCC 6303, Cylindrospermum stagnale PCC 7417 and Anabaena cylindrica PCC 7122. Prediction of the interaction site between Yfr1 and the 5’-UTR of several predicted targets in Nostoc sp. PCC 7120 was performed using IntaRNA (35). Alignment of Yfr2 homologs was made using Clustal Omega (58). Secondary structures of Yfr1 and Yfr2-1 and their interaction were predicted by RNAcofold (59). Reporter assay for in vivo verification of targets. We used the reporter assay described in (60) and fusions to the gene encoding superfolder GFP (sfGFP) in plasmids pXG10-SF or pXG30-SF (34) for experimental target verification in E.coli (Table S3). In this system both the GFP fusions and Yfr1 are transcribed constitutively. 5’-UTR of monocistronic targets were cloned in pXG10-SF from their corresponding TSS according to (16) to 30-60 nucleotides within the coding region. For targets that could be cotranscribed with a gene located upstream, the last 60 nucleotides of the upstream gene, together with the whole intergenic region plus 60 nucleotides within the coding region of the target gene were cloned in pXG30-SF. To facilitate translation in E. coli, GTG start codons were replaced by ATG using overlapping PCR and oligonucleotides specified in Table S4. PCR fragments containing the region to be cloned were amplified using genomic DNA as template and oligonucleotides specified in Table S4. Fragments were digested with NsiI and NheI and cloned into pXG10-SF or pXG30-SF treated with the same enzymes, resulting in translational fusions of the different targets to sfGFP (Table S5). To express Yfr1 in E. coli, the sequence encoding Yfr1 was amplified from genomic DNA using primers 422 (5’ phosphorylated) and 423. The PCR product was digested with XbaI and fused to a plasmid backbone that was amplified from pZE12-luc with primers PLlacOB and PLlacOD (60) and digested with XbaI, rendering pMBA1 (Table S6). For the mutagenesis of Yfr1 and the 5’-UTRs of all4316 and alr5065, mutations were introduced by overlapping PCR with primers containing the desired changes (Table S4) and the fragments were cloned in the same way as the corresponding wild type versions. The specific mutations were designed based on changes in the hybridization energies predicted by IntaRNA (35). 87 Combinations of plasmids bearing fragments encoding Yfr1 (or its mutated versions) and the 5’ UTRs of target genes (or mutated versions) were introduced in E. coli DH5α. Plasmid pJV300 was used as a control expressing an unrelated RNA. Fluorescence measurements were done with a microplate reader (Varioskan) using liquid cultures from eight individual colonies of cells carrying each plasmid combination, as previously described (29). In vitro transcription of RNA and EMSA essay. RNAs were transcribed in vitro with MEGAscript High Yield Transcription Kit (AM1333, Ambion). The DNA templates for the transcription of Yfr1 and Yfr2 were generated by PCR with a primer that includes a T7 promoter sequence upstream the 5’-end of the corresponding RNA, and a primer matching the 3’-end of the RNA (Table S4). The template used for these PCR amplifications was genomic DNA. After in vitro transcription, RNAs were treated with DNase I and purified by sequential phenol, phenol/chloroform and chloroform extractions, ethanolprecipitated at −20 °C, and washed with 70% ethanol. For EMSA, 200 ng of each in vitro transcribed Yfr1 and Yfr2 were combined in a volume of 5 µL, denatured for 1 min at 95 °C and chilled on ice for 5 min. After denaturing and chilling steps, 10x structure buffer (AM7118, Ambion) was added and the samples were incubated further for 15 min at 37 °C, before adding 1 µL of 50 % glycerol. Samples were run on 2.5 % agarose gel with TBE buffer 0.5 % at 50 V in a cold chamber. ACKNOWLEDGEMENTS We thank Iris Maldener and Rebeca Pérez (University of Tübingen, Germany) for the protocol for the fluorescent vancomycin assay, Claudia Steglich (University of Freiburg, Germany) for making available to us the information on the interaction between Yfr1 and Yfr2 prior to publication, and Claudia Steglich and Wolfgang Hess (University of Freiburg, Germany) for valuable discussions. This work was supported by grants BFU2013-48282-C2-1 from Ministerio de Economía y Competitividad, and BFU2016-74943-C2-1-P from Agencia Estatal de Investigación (AEI), Ministerio de Economía, Industria y Competitividad, both cofinanced by Fondo Europeo de Desarrollo Regional (FEDER), to AMMP. MBA is the recipient of a predoctoral contract from Ministerio de Educación, Cultura y Deporte, Spain (FPU014/05123 and EST16- 00088). 88 REFERENCES 1. Silhavy TJ, Kahne D, Walker S. 2010. The bacterial cell envelope. Cold Spring Harb Perspect Biol 2:a000414. 2. Hoiczyk E, Hansel A. 2000. Cyanobacterial cell walls: news from an unusual prokaryotic envelope. J Bacteriol 182:1191-1199. 3. Vollmer W, Joris B, Charlier P, Foster S. 2008. Bacterial peptidoglycan (murein) hydrolases. FEMS Microbiol Rev 32:259-286. 4. Muro-Pastor AM, Hess WR. 2012. Heterocyst differentiation: from single mutants to global approaches. Trends Microbiol 20:548-557. 5. Kumar K, Mella-Herrera RA, Golden JW. 2010. Cyanobacterial heterocysts. Cold Spring Harb Perspect Biol 2:a000315. 6. Mullineaux CW, Mariscal V, Nenninger A, Khanum H, Herrero A, Flores E, Adams DG. 2008. Mechanism of intercellular molecular exchange in heterocyst-forming cyanobacteria. Embo j 27:1299-1308. 7. Nürnberg DJ, Mariscal V, Bornikoel J, Nieves-Morión M, Krauss N, Herrero A, Maldener I, Flores E, Mullineaux CW. 2015. Intercellular diffusion of a fluorescent sucrose analog via the septal junctions in a filamentous cyanobacterium. mBio 6:e02109-14. 8. Nieves-Morión M, Mullineaux CW, Flores E. 2017. Molecular diffusion through cyanobacterial septal junctions. mBio 8:e01756-16. 9. Weiss GL, Kieninger A-K, Maldener I, Forchhammer K, Pilhofer M. 2019. Structure and function of a bacterial gap junction analog. Cell 178:374-384.e315. 10. Bornikoel J, Staiger J, Madlung J, Forchhammer K, Maldener I. 2018. LytM factor Alr3353 affects filament morphology and cell-cell communication in the multicellular cyanobacterium Anabaena sp. PCC 7120. Mol Microbiol 108:187-203. 11. Zhu J, Jager K, Black T, Zarka K, Koksharova O, Wolk CP. 2001. HcwA, an autolysin, is required for heterocyst maturation in Anabaena sp. strain PCC 7120. J Bacteriol 183:6841-6851. 12. Bornikoel J, Carrión A, Fan Q, Flores E, Forchhammer K, Mariscal V, Mullineaux CW, Perez R, Silber N, Wolk CP, Maldener I. 2017. Role of two cell wall amidases in septal junction and nanopore formation in the multicellular cyanobacterium Anabaena sp. PCC 7120. Front Cell Infect Microbiol 7:386. 13. Videau P, Rivers OS, Ushijima B, Oshiro RT, Kim MJ, Philmus B, Cozy LM. 2016. Mutation of the murC and murB genes impairs heterocyst differentiation in Anabaena sp. strain PCC 7120. J Bacteriol 198:1196-1206. 89 Figure S2. Interactions predicted by IntaRNA between Yfr1 and several mRNAs. Nucleotides of 5’UTRs are numbered with respect to the start of the coding sequence (start codons are indicated in bold and underlined). Positions of the mutations introduced in Yfr1_UG are shown in red. Positions of the mutations introduced in Yfr1_AAAA are shown in purple. 96 Figure S3. Comparison of the effects of Yfr1 on expression of all4316 and alr5065 at the mRNA (A, B) and protein level (C) in E. coli. (A) Northern blot with RNA isolated from E. coli cells bearing plasmid pMBA4 (all4316-sfgfp) or pMBA7 (alr5065-sfgfp) and a plasmid expressing Yfr1 (pMBA1) or a control RNA (pJV300) hybridized with probes for all4316, alr5065, Yfr1 and 5S RNA as loading control. (B) Quantification of signals in Northern blots. Data are presented as the mean +/- standard deviation of the signal normalized to the 5S signal (three individual experiments). (C) Fluorescence of the E. coli cultures analyzed in (A-B). The data are presented as mean +/- standard deviation of the results from three cultures after subtraction of fluorescence in cells bearing pXG0 and normalized for cell density (A600). (*) p<0,05: (**) p<0,0001, t-test. 97 Figure S4. Interaction between Yfr1 and Yfr2. (A) Clustal Omega alignment of the four Yfr2 RNAs encoded in the genome of Nostoc sp. PCC 7120. The region predicted to interact with Yfr1 is shaded in orange. (B) Secondary structures of Yfr1 (black) and Yfr2-1 (red) and their interaction as predicted by RNAcofold. (C) EMSA showing the interaction between in vitro transcribed Yfr1 and Yfr2. Equimolar amounts of Yfr1 and Yfr2 were combined and subjected to electrophoresis as described in the Material and Methods. 98 Table S1. Top 50 predicted targets of Yfr1. Gene Annotationa fdr p-value alr2458 alanine racemase 7,52E-09 2,14E-12 alr4550 porin-like, OprB-like 1,51E-08 8,60E-12 alr5065 (murC) UDP-N-acetylmuramate-L-alanine ligase 2,82E-08 2,40E-11 all4829 undecaprenyl-phosphate galactosephosphotransferase 2,64E-07 3,00E-10 all2158 ferrichrome-iron receptor (TonB-dependent transporter) 3,29E-07 4,68E-10 all4316 (mraY) phospho-N-acetylmuramoyl-pentapeptide transferase 5,12E-07 8,72E-10 alr2269 chloroplastic outer envelope membrane protein homolog, Omp85, Toc75 6,58E-07 1,31E-09 all5094 oxidoreductase 7,30E-07 1,66E-09 all1835 unknown protein 2,39E-06 6,11E-09 alr0834 porin, ortholog of PMM1119 and PMM1121 3,05E-06 8,65E-09 all4499 porin-like, OprB-like 4,93E-06 1,54E-08 all2352 WD-40 repeat protein. 6,88E-06 2,35E-08 alr4077 hypothetical protein 7,09E-06 2,62E-08 all0187 (conR) LytR-CpsA-Psr superfamily 9,05E-06 3,68E-08 all1587 hypothetical protein 9,05E-06 3,86E-08 all0481 unknown protein 9,32E-06 4,24E-08 alr0068 adenylate kinase 1,37E-05 6,85E-08 all4220 hypothetical protein 1,37E-05 6,99E-08 alr3380 dolichol-phosphate mannosyltransferase 3,32E-05 1,81E-07 alr3654 uncharacterized RNA methyltransferase 3,32E-05 1,89E-07 alr2887 (hgdD) outer membrane channel, TolC-like, 4,54E-05 2,78E-07 alr2057 (aroE) shikimate 5-dehydrogenase 4,54E-05 2,96E-07 alr1805 hypothetical protein 4,54E-05 2,97E-07 all3826 peptidoglycan-binding protein 4,92E-05 3,36E-07 alr4812 (patN) heterocyst differentiation related protein 5,49E-05 3,90E-07 all4026 similar to TonB-dependent receptor 7,32E-05 5,41E-07 asr3279 unknown protein 8,67E-05 6,65E-07 alr0093 (amiC2, hcwA) N-acetylmuramoyl-L-alanine amidase 9,51E-05 7,57E-07 alr5049 unknown protein, putative OmpA/MotB system 9,87E-05 8,13E-07 all3669 unknown protein 1,05E-04 8,97E-07 alr4123 (prk) phosphoribulokinase 1,40E-04 1,23E-06 all1340 unknown protein 1,44E-04 1,31E-06 alr5030 hypothetical protein 1,99E-04 1,86E-06 all3902 hypothetical protein 2,35E-04 2,27E-06 all2116 hypothetical protein 3,54E-04 3,58E-06 alr4602 type I site-specific deoxyribonuclease chain S 3,54E-04 3,62E-06 alr2479 LytR-CpsA-Psr superfamily 3,80E-04 4,00E-06 alr4714 unknown protein, specific to heterocyst clade 3,85E-04 4,15E-06 alr4269 hypothetical protein 3,91E-04 4,33E-06 99 alr3004 unknown protein 4,43E-04 5,03E-06 all0089 hypothetical protein 4,59E-04 5,34E-06 all0571 cyanophycinase 4,77E-04 5,69E-06 all0211 hypothetical protein 5,09E-04 6,21E-06 asl4743 unknown protein, specific to heterocyst clade 5,36E-04 6,70E-06 alr3955 thioredoxin 6,42E-04 8,21E-06 all3310 hypothetical outer membrane receptor for ferritin, (TonB-dependent transporter) 6,66E-04 8,71E-06 all0997 unknown protein, specific to heterocyst clade 7,50E-04 1,00E-05 all4106 probable oxidoreductase 7,91E-04 1,08E-05 alr0255 unknown protein, specific to heterocyst clade 7,98E-04 1,11E-05 aAnnotation based on NCBI (https://www.ncbi.nlm.nih.gov/), Cyanobase (http://genome.microbedb.jp/cyanobase/), CopraRNA results (1, 2) and previous studies. Interactions verified in this work are highlighted in gray. Table S2. Strains Strain Description Reference Escherichia coli DH5α Used for routine transformation (3) Nostoc sp. PCC 7120 Wild type Pasteur Culture Collection OE_C SmRSpR, pMBA51 inserted in plasmid alpha. T1 terminator of E.coli rrnB gene expressed constitutively from trc promoter. (4) OE_Yfr1 SmRSpR, pMBA48 inserted in plasmid alpha. Yfr1 expressed constitutively from trc promoter. This work OE_as_Yfr1 SmRSpR, pMBA49 inserted in plasmid alpha. Antisense to Yfr1 expressed constitutively from trc promoter. This work 100 Table S3. Plasmids Name Description Reference pMBA37 ApRSmRSpR, plasmid for the overexpression of transcripts from the trc promoter and followed by the T1 terminator of E.coli rrnB gene, used as transcriptional terminator. (4) pMBA51 ApRSmRSpR, Control plasmid, expresses a 56 nt transcript derived from the T1 terminator of E.coli rrnB gene from trc promoter. (4) pMBA48 ApRSmRSpR, plasmid based on pMBA37 for the overexpression of Yfr1. This work pMBA49 ApRSmRSpR, plasmid based on pMBA37 for the overexpression of an antisense to Yfr1. This work pJV300 Control plasmid, expresses a ~50 nt transcript derived from T1 terminator of E.coli rrnB gene. (5) pXG0 CmR, control plasmid without sfgfp. (6) pXG10-SF CmR, plasmid for construction of translational sfGFP fusions of monocistronic targets. (7) pXG30-SF CmR, plasmid for construction of translational sfGFP fusions of dicistronic targets. (7) pZE12-luc ApR, plasmid to express sRNAs under control of PLlac-O promoter. (8) pMBA1 ApR, plasmid based on pZE12-luc expressing Yfr1. This work pMBA13 ApR, same as pMBA1 but with AC to UG change at positions 31-32 of Yfr1 (mut 31_32). This work pMBA15 ApR, same as pMBA1 but with CCUC to AAAA change at positions 27-30 of Yfr1 (mut AAAA). This work pMBA2 CmR, sfGFP reporter plasmid based on pXG10-SF containing the all0187 5’UTR plus sequences encoding the first 20 aminoacids (with mutation in start codon GTG -> ATG). This work pMBA3 CmR, sfGFP reporter plasmid based on pXG30-SF containing the all2158 5’UTR plus sequences encoding the first 20 aminoacids. This work pMBA4 CmR, sfGFP reporter plasmid based on pXG30-SF containing the all4316 5’UTR plus sequences encoding the first 20 aminoacids (with mutation in start codon GTG -> ATG). This work pMBA5 CmR, sfGFP reporter plasmid based on pXG10-SF containing the all4829 5’UTR plus sequences encoding the first 17 aminoacids. This work pMBA6 CmR, sfGFP reporter plasmid based on pXG30-SF containing the alr0093 5’UTR plus sequences encoding the first 20 aminoacids (with mutation in start codon GTG -> ATG). This work pMBA7 CmR, sfGFP reporter plasmid based on pXG10-SF containing the alr5065 5’UTR plus sequences encoding the first 20 aminoacids. This work pMBA8 CmR, sfGFP reporter plasmid based on pXG10-SF containing the alr2458 5’UTR plus sequences encoding the first 14 aminoacids. This work pMBA9 CmR, sfGFP reporter plasmid based on pXG10-SF containing the alr4550 5’UTR plus sequences encoding the first 20 aminoacids. This work pMBA10 CmR, sfGFP reporter plasmid based on pXG10-SF containing the patN 5’UTR plus sequences encoding the first 20 aminoacids. This work pMBA11 CmR, sfGFP reporter plasmid based on pXG10-SF containing the alr0834 5’UTR plus sequences encoding the first 20 aminoacids. This work pMBA17 CmR, same as pMBA7 but with a GT to CA change at position -21 and -22 of the 5'-UTR with respect to the start codon (alr5065 mut). This work pMBA19 CmR, same as pMBA4 but with a GT to CA change at position -22 and -23 of the 5'-UTR with respect to the start codon (all4316 mut). This work pMBA90 CmR, sfGFP reporter plasmid based on pXG10-SF containing the alr2269 5’UTR plus sequences encoding the first 20 aminoacids (with mutation in start codon GTG -> ATG). This work 101 Table S4. Oligonucleotides Name Sequence (5’-3’) Used for 189 CGCACTGACCGAATTCATTAA Amplification of plasmid backbone from pZE12-luc 190 GTGCTCAGTATCTTGTTATCCG 356 AATCAGAACGCAGAAGCGGT Probe for 5S of E. coli 357 CGGCGGATTTGTCCTACTCA 368 AAAGGGAATAGGGAACGGCG Probe for Yfr1 369 CGGCTGCTGACATGCCTAAA 422 5’P-AGCGGAGACGCATGTTTCCGTTCAC Cloning of Yfr1 in pZE12-luc 423 GTTTTTTCTAGATACATTTGCAAAGAAGCTAAAGAATT 424 GTTTTATGCATAAACGGTAGAACATTTCAAGTG Cloning of all0187 5’-UTR with mutation in start codon (GTG- >ATG) in pXG10-SF 425 GTTTTGCTAGCGAGATCCGGTATTTGTTGAGGC 426 TTAATTACAATTATCTCCCCACTCAACCA 427 TAATTATGATTAAACAGGTTCAATGGTCA 428 GTTTTATGCATGCAGGTCACAAGACGGTTTT Cloning of all2158 5’-UTR in pXG30-SF 429 GTTTTGCTAGCTCCACAAAGGCTGATGATAA 430 GTTTTATGCATGACGACGATATCACCATAGA Cloning of all4316 5’-UTR with mutation in start codon (GTG- >ATG) in pXG30-SF 431 GTTTTGCTAGCAGAAGCTAGACCAATACCAG 432 GCGTCCATAAAAGTTTCCCTTCACTCCAC 433 CTTTTATGGACGCTAAATTATCGCCTA 434 GTTTTATGCATATGTGTGGTGTGGATTATTAAGG Cloning of all4829 5’-UTR in pXG10-SF 435 GTTTTGCTAGCATCTTGCCGTAAGCGTCGTTT 436 GTTTTATGCATCTTCAGTACCTCAAACGATAG Cloning of alr0093 5’-UTR with mutation in start codon (GTG- >ATG) in pXG30-SF 437 GTTTTGCTAGCTGGTGACGATAGTAAGAAGA 438 AATTTCATAATCTTTTCTCCTGATGTGCG 439 AGATTATGAAATTACACTGGTTACTATC 440 GTTTTATGCATATACTTTTGAAGTAAATATTACTTC Cloning of all4829 5’-UTR in pXG10-SF 441 GTTTTGCTAGCATTTGTTTCTGTTATGGCAG 442 GTTTTATGCATGTAATTGATCCTGTCGGTTA Cloning of alr2458 5’-UTR in pXG10-SF 443 GTTTTGCTAGCTTGATGAGAAGCCATACTTG 444 GTTTTATGCATATATGGGTACGTTTTTGCAT Cloning of alr4550 5’-UTR in pXG10-SF 445 GTTTTGCTAGCTAGTGTTGCGCCTAAAAC 446 GTTTTATGCATATCAGAATATTCATCTGTAAAAAA Cloning of patN 5’-UTR in pXG10-SF 447 GTTTTGCTAGCCAAGAGAGCCGCCAGTCCAG 448 GTTTTATGCATGTTATTTGTGATTGATGGTGTG Cloning of alr5065 5’-UTR in pXG10-SF 449 GTTTTGCTAGCTATACCGCCGATTCCGATAA 470 CTCCTCTGACCACACTCCGCCCGGACTA Mutagenesis of Yfr1 to generate Yfr1_UG 471 TGTGGTCAGAGGAGTGAACGGAAACATG 484 ATTAAACAGTGGAGTGAAGGGAAACTTT Mutagenesis of all4316 5’-UTR to generate all4316 mut 485 CTCCACTGTTTAATAAAACTGAATTATA 486 TGTGGTCAGGTGATAAGAGGAGTAATCA Mutagenesis of alr5065 5’-UTR to generate alr5065 mut 487 ATCACCTGACCACACCATCAATCACAAA 491 TCACTAAAAACACCACACTCCGCCCGGAC Mutagenesis of Yfr1 to generate Yfr1_AAAA 492 GGTGTTTTTAGTGAACGGAAACATGCGTC 575 GTTTTATGCATAGCGGAGACGCATGTTTCC Cloning of Yfr1 in pMBA37 576 GTTTTCTCGAGGGAACCGCCCGAACAGTAGT 577 GTTTTATGCATGGAACCGCCCGAACAGTAGT Cloning of as_Yfr1 in pMBA37 578 GTTTTCTCGAGAGCGGAGACGCATGTTTCCG 874 GTTTTATGCATACCTTTGGCATAAACAAGCAAT Cloning of alr2269 5’-UTR with mutation in start codon (GTG- >ATG) in pXG10-SF 875 GTTTTGCTAGCAGTTTGGGCATTTGCAGTTAATGA 876 CAGCCGCCATTAATACGGGAGATAAACGCA 877 TCCCGTATTAATGGCGGCTGTAGCAATCAC 903 TAATACGACTCACTATAGGGAGCGGAGACGCATGTTTCC G 102 904 AAGGAACCGCCCGAACAGTAG PCR of template for in vitro transcription of Yfr1 905 TAATACGACTCACTATAGGGTGTGAGGAGCAAGTTGAAA PCR of template for in vitro transcription of Yfr2 906 AAAAAGCATCTGGGAGTGTTGC Restriction sites used for cloning (underlined), point mutations introduced (red) and 5’ monophosphate (5’P) are indicated. Table S5. Sequences of inserts in the sfGFP fusion plasmids. Plasmid Sequence Description pMBA2 atgcatAAACGGTAGAACATTTCAAGTGTACAGTGCGGTCTTACAC TGGCAAGATTGGTAGCTTGCTTTAACATAAATTTTAGGGAGTGGTT GAGTGGGGAGATAATTATGATTAAACAGGTTCAATGGTCAGATAAT CAGGTCGCGCCTCAACAAATACCGGATCTCgctagc all0187 (mutation in start codon) pMBA3 atgcatGCAGGTCACAAGACGGTTTTTGGATCATAACTGGCGGTTT TTGGATAGACACCCTGCGAGTAACTTCCCTATACTACTTTAACTGC TATTAAGAATTATTAGCAGTAATAGGCTGACTTTGATAAGTTTATT GTGTGGTGTGAGGAACGATGAAGCGTTGGTGTTTGTCCCCCAGTAT TCATTTATGGTTTATCATCAGCCTTTGTGGAgctagc all2158 pMBA4 atgcatGACGACGATATCACCATAGACGAAGATTAGTCTTTTGAGA CTGGGACTCTGTTATATATAGTCCCTATAATTCAGTTTTATTAAAG TGTGGAGTGAAGGGAAACTTTTATGGACGCTAAATTATCGCCTAAT CAAGGATTAAATATTTCTGGTATTGGTCTAGCTTCTgctagc all4316 (mutation in start codon) pMBA5 atgcatATGTGTGGTGTGGATTATTAAGGAGTATGATGACTGCCCA GAGTTCACTCCTCTCCGGCAAACGACGCTTACGGCAAGATgctagc all4829 pMBA6 atgcatCTTCAGTACCTCAAACGATAGATAATCAGTCAGTCTTCAC AAACTTATAATCATCCTGATTGCCACACTATCCCTGTTATTCCTGA AGATTTTGTGTGAATACTAACCACATCCTTATATTCTGAACCTGGA TCTTATGCCTTCACGACATGGCAAAATCAAAAGGGCGTGTTAAGAT GTGTCGCCATTAGTTGTATTTTTATACCAGCGCAAAATTTTTATTG CAGTAAAGCTGAAGTTTGAACGGTGATGGTGTGAGGATTAGCCAAC ATTTTGGAGCCTTAACGGCTCTCATGAACGCACATCAGGAGAAAAG ATTATGAAATTACACTGGTTACTATCGTCACCAgctagc alr0093 (mutation in start codon) pMBA7 atgcatGTTATTTGTGATTGATGGTGTGGTGTGGTGATAAGAGGAG TAATCAATGAATAATAGTGTAGATTTTGGCGGTAGACCATTTCATT TTATCGGAATCGGCGGTATAgctagc alr5065 pMBA8 atgcatGTAATTGATCCTGTCGGTTATTGGTAATGGCTCAGAACCC CTATTAATTGGCTATTACCTCAGTCAAAAACAGAAGTTTAGCTCTT ATTACTGCTAGTTATTAGTAGCAAAAAGAACCTGTTAATTTTAATT AAGAGGTGGAGTATTGAGTAAAATACACTAGGTTCCAGTCTCTAAC CCCCTCCCTGATCAACATCTGAACGAGAAAATTTCAAAATTTTACT CAAGTATGTTGTATCTGTTGCGAAATGCGAGTTAAACTTCCTGATT AACTCTGATGTGGCTATAAAACCAATTCGCAATAGCCTGCGGCAAG GCTAACGCCAACGCAATTCGGAATAAATACGCATTTAAAAAGCAGT GGTGTGATAAGGGGGAGTGAAATGTTAAGTCGCCAACAAGCCTCAA GTATGGCTTCTCATCAAgctagc alr2458 pMBA9 atgcatATATGGGTACGTTTTTGCATAAGTTAATTGCATTAATTTT TAATACTAAGTTAATCAATCATCAGCTTGGCAAAAACTAAAATTGT AATAGTGGCAACTATGCCTCATCAAAAGATTTCAATAAGATGTCAC ATTTTTGAGAGTAGCTGCAACAAGAACAATTATCTATACAAGGTGT GAGGAGAAAAGTAAAAATGTCTAATCTATTGTGGAAATCCCTAGTG GTTAGCCCTGCTGTTTTAGGCGCAACACTAgctagc alr4550 pMBA10 atgcatATCAGAATATTCATCTGTAAAAAAACTAGCTGAAACAGCA GCTTAATATTTTGGGTATTATGTGAATTAAGTAGGCTAAAAGATTA alr4812 103 TGAGTGTGAGAGTAGGTCAGTCAAACGTTATGGAATTTACATCGCA GATGAGGGGTGAAGGCGATGAGTGAAAGTATGGCATTTATCGGCGG TGTCGCCGTAGCTGGACTGGCGGCTCTCTTGgctagc pMBA11 atgcatATACTTTTGAAGTAAATATTACTTCAAATAAACTTGGTAA ATCTATTGCCAAGAAATAAGATGTGTGTGAGTAAATTGTGAGGAGA AAAATGCAGAAAATATGGAAGTATTGGCTGGTTAACCCAGTAATTT GCAGCACAATGCTATTTTCTGGCGCTGCTGCCTTTGCAGGAGAAAC CCCTGCCATAACAGAAACAAATgctagc alr0834 pMBA17 atgcatGTTATTTGTGATTGATGGTGTGGTCAGGTGATAAGAGGAG TAATCAATGAATAATAGTGTAGATTTTGGCGGTAGACCATTTCATT TTATCGGAATCGGCGGTATAgctagc alr5065 mut pMBA19 atgcatGACGACGATATCACCATAGACGAAGATTAGTCTTTTGAGA CTGGGACTCTGTTATATATAGTCCCTATAATTCAGTTTTATTAAAC AGTGGAGTGAAGGGAAACTTTTATGGACGCTAAATTATCGCCTAAT CAAGGATTAAATATTTCTGGTATTGGTCTAGCTTCTgctagc all4316 mut pMBA90 atgcatACCTTTGGCATAAACAAGCAATTGCCTGGTAATGGTGTGT GGATGTGAAGAGGAAGAAAGGAATTAAAGACGATGCGTTTATCTCC CGTATTAATGGCGGCTGTAGCAATCACAGCACCCTTGAGTAGTTCA TTAACTGCAAATGCCCAAACTgctagc alr2269 (mutation in start codon) Nostoc sequences are capitalized, with black letters corresponding to 5’UTR and green letters corresponding to coding sequences, respectively. NsiI and NheI sites that were used for cloning are highlighted in magenta and yellow, respectively. Nucleotide changes with respect to wild type sequences are indicated in red. Table S6. Sequences of inserts in plasmids used for expression of Yfr1 and its mutated versions. Plasmid Sequence Description pMBA1 AGCGGAGACGCATGTTTCCGTTCACTCCTCACACCACACTCCGCCCGGAC TACTGTTCGGGCGGTTCCTTATGTATTACATAAATTCTTTAGCTTCTTTG CAAATGTATCTAGA Yfr1 pMBA13 AGCGGAGACGCATGTTTCCGTTCACTCCTCTGACCACACTCCGCCCGGAC TACTGTTCGGGCGGTTCCTTATGTATTACATAAATTCTTTAGCTTCTTTG CAAATGTATCTAGA Yfr1_UG pMBA15 AGCGGAGACGCATGTTTCCGTTCACTAAAAACACCACACTCCGCCCGGAC TACTGTTCGGGCGGTTCCTTATGTATTACATAAATTCTTTAGCTTCTTTG CAAATGTATCTAGA Yfr1_AAAA Grey shadowed letters indicate the Yfr1 sequence. Modified nucleotides are marked in red. XbaI restriction site used for cloning is shown in blue. 104 SUPLEMENTARY REFERENCES 1. Wright PR, Georg J, Mann M, Sorescu DA, Richter AS, Lott S, Kleinkauf R, Hess WR, Backofen R. 2014. CopraRNA and IntaRNA: predicting small RNA targets, networks and interaction domains. Nucleic Acids Research 42:W119-W123. 2. Wright PR, Richter AS, Papenfort K, Mann M, Vogel J, Hess WR, Backofen R, Georg J. 2013. Comparative genomics boosts target prediction for bacterial small RNAs. Proc Natl Acad Sci U S A 110:E3487-3496. 3. Hanahan D. 1983. Studies on transformation of Escherichia coli with plasmids. J Mol Biol 166:557-580. 4. Olmedo-Verd E, Brenes-Álvarez M, Vioque A, Muro-Pastor AM. 2019. A heterocystspecific antisense RNA contributes to metabolic reprogramming in Nostoc sp. PCC 7120. Plant Cell Physiol 60:1646-1655. 5. Sittka A, Pfeiffer V, Tedin K, Vogel J. 2007. The RNA chaperone Hfq is essential for the virulence of Salmonella typhimurium. Mol Microbiol 63:193-217. 6. Urban JH, Vogel J. 2009. A green fluorescent protein (GFP)-based plasmid system to study post-transcriptional control of gene expression in vivo. Methods Mol Biol 540:301-319. 7. Corcoran CP, Podkaminski D, Papenfort K, Urban JH, Hinton JC, Vogel J. 2012. Superfolder GFP reporters validate diverse new mRNA targets of the classic porin regulator, MicF RNA. Mol Microbiol 84:428-445. 8. Lutz R, Bujard H. 1997. Independent and tight regulation of transcriptional units in Escherichia coli via the LacR/O, the TetR/O and AraC/I1-I2 regulatory elements. Nucleic Acids Res 25:1203-1210. 105 Classification: Biological Sciences, Microbiology Title: NsiR1, a small RNA with multiple copies, regulates heterocyst differentiation in Nostoc sp. PCC 7120 Authors: Manuel Brenes-Álvarez, Marina Mingueta, Agustín Vioque1, and Alicia M. Muro-Pastor ORCID MBA: 0000-0002-6452-5557 ORCID AMMP: 0000-0003-2503-6336 ORCID AV: 0000-0002-3975-7348 Authors affiliation: Instituto de Bioqumica Vegetal y Fotosntesis, Consejo Superior de Investigaciones Cientficas and Universidad de Sevilla, Sevilla, Spain. aPresent address: Centro de Biología Molecular Severo Ochoa, CSIC-UAM, Nicolas Cabrera 1, 28049, Madrid, Spain. 1Corresponding Author: Agustín Vioque, Instituto de Bioqumica Vegetal y Fotosntesis, Américo Vespucio 49, 41092, Sevilla, Spain Phone: +34-954489519; Email: vi[email protected] Keywords: cyanobacteria, differentiation commitment, Anabaena, HetR, Author contributions: M.B.A., A.V., and A.M.M.P. designed research; M.B.A., M.M., A.V., and A.M.P. performed research; M.B.A., A.V., and A.M.P. analyzed data; and M.B.A., A.V., and A.M.P. wrote the paper. The authors declare no conflict of interest. 163 Abstract Upon nitrogen starvation filamentous cyanobacteria develop heterocysts, specialized cells devoted to fixation of atmospheric nitrogen. Differentiation of heterocyst at semi-regular intervals requires complex structural and functional changes that are under control of the master transcriptional regulator HetR. NsiR1 (nitrogen stress-induced RNA 1) is a HetR- dependent non-coding RNA that is expressed in early steps of heterocyst development. In Nostoc sp. PCC 7120 there are twelve tandem copies of nsiR1 (nsiR1.1 to nsiR1.12), seven of them with identical sequence (nsiR1.3 to nsiR1.9) and the others slightly divergent. nsiR1.1 is transcribed antisense to the 5’ UTR of hetF, a gene required for heterocyst development. Here we show that binding of NsiR1.1 inhibits translation of the hetF mRNA by inducing structural changes in its 5’ UTR. Constitutive overexpression of NsiR1.1 or NsiR1.4 results in different phenotypes related to heterocyst development, suggesting that they regulate non-overlapping targets. We have identified gene alr3234, a hetP-like gene involved in regulation of commitment to heterocyst differentiation, as a target of NsiR1.4. A strain overexpressing NsiR1.4 commits to heterocyst development earlier than the wild type. The post-transcriptional regulation by NsiR1 of two genes related to the process of heterocyst differentiation and commitment, hetF and alr3234, adds a new level of complexity to the network of transcriptional regulation and proteinprotein interactions involved in heterocyst differentiation. Significance statement Heterocysts are nitrogen-fixing specialized cells that appear at semi-regular intervals along cyanobacterial filaments upon nitrogen starvation. Heterocyst differentiation and patterning is a model system to study cell differentiation in multicellular prokaryotes. The differentiation process has a complex regulation, only partially understood, that includes transcriptional changes, factor diffusion between cells, and protein-protein interactions. This work adds to this complexity a non-coding RNA, NsiR1, that negatively regulates, at a posttranscriptional level, hetF and alr3234 (encoding a HetP-like protein), two genes that are part of the heterocyst differentiation regulatory network. Thus, regulation by a small RNA is a new element involved in the control of heterocyst differentiation. 164 Introduction Cyanobacteria are the only prokaryotes able to perform oxygenic photosynthesis and important contributors to global primary productivity. They constitute a coherent phylogenetic group with diverse morphology and wide ecological distribution. Nostoc sp. PCC 7120 (a.k.a Anabaena sp. PCC 7120) is a filamentous cyanobacterium that under nitrogen deprivation differentiates heterocysts, specialized cells devoted to fixation of atmospheric nitrogen. Heterocyst differentiation involves a complex developmental program (1-4) that is ultimately under control of NtcA, the global regulator of nitrogen assimilation (5), but also under control of HetR, a specific regulator of cell differentiation (6, 7). At the initiation of heterocyst differentiation, induction of the expression of NtcA and HetR depends on each other (8) and leads to increased expression of HetR in cells becoming heterocysts. Expression of hetR is regulated at several levels. hetR is transcribed from a complex promoter including several transcriptional start sites (TSSs), one of them specifically induced in cells becoming heterocysts (9, 10). HetR is a DNA-binding protein that positively regulates its own expression and the expression of many genes, including negative elements essential for the maintenance of the pattern of heterocysts along the filaments such as patS (11), hetN (12) and patX (13). The binding of the pentapeptide RGSGR, that is part of the sequence of PatS, blocks the DNA-binding activity of HetR (14) and signals its degradation (15). HetR is an autoproteolytic protein (16) that can also be degraded by the HetF protease (17, 18). A hetF mutant does not differentiate heterocysts, has an altered morphology with enlarged cells and accumulates a high amount of HetR (18). However, since hetF transcription seems unpatterned along the filaments (18), the question of how HetF may contribute to the patterned accumulation of active HetR remains open. Because heterocysts are terminally differentiated non-dividing cells, a key aspect of differentiation is the so-called commitment point, a stage of differentiation from which the process is no longer halted by the addition of combined nitrogen. Several genes have been identified related to the regulation of commitment in Nostoc sp. PCC 7120, including hetP (19) and its homologs asl1930, alr2902 and alr3234 (20). Asl1930, Alr2902, and Alr3234 would interact with HetR and with each other, regulating commitment via protein-protein interactions (20). Mutant strains with genes alr3234 or asl1930 deleted reach the commitment point earlier than the wild-type strain, which suggests that Alr3234 and Asl1930 might function as negative regulators delaying commitment (20). 165 Small non-coding RNAs (sRNAs) are important regulators of gene expression in bacteria, including cyanobacteria (21). Most frequently they bind to their target mRNAs close to the ribosome binding site therefore inhibiting translation, but a large diversity of alternative regulatory mechanisms by sRNAs are known (22). They are involved in the regulation of diverse processes related to response to stress or adaptation to a changing environment, included cell differentiation. In Bacillus several sRNAs have been identified that are sporulation-specific and with expression restricted to the forespore (23-26). In heterocyst-forming cyanobacteria sRNAs and antisense RNAs have been identified with heterocyst-specific expression (27-29). A heterocyst-specific antisense RNA is involved in carbon fixation shutdown in heterocysts (30). NsiR1 is a small RNA of about 60 nucleotides encoded in 12 tandem copies in the region upstream of hetF in Nostoc sp. PCC 7120 (27) (Fig. 1A). NsiR1 is expressed specifically in cells becoming heterocyst before any sign of morphological differentiation is observed (31). The different copies are transcribed from individual promoters that contain the DIF1 sequence motif associated to heterocyst-specific transcription (10, 28). Central copies nsiR1.3 to nsiR1.9 are identical whereas the other copies, located at the periphery of the array, have some sequence divergence. According to the position of the TSS of hetF (10, 27), nsiR1.1 overlaps the hetF 5’ UTR. NsiR1.1 could therefore function as an antisense of the hetF transcript. As the seven identical copies (3-9, thereafter nsiR1.4) are very similar to nsiR1.1, they could also potentially regulate the expression of hetF or other targets in Nostoc sp. PCC 7120. In this work we show that Nostoc strains constitutively overexpressing nsiR1.1 or nsiR1.4 have distinct phenotypes related to heterocyst differentiation, indicating that they have non-overlapping targets. NsiR1.1 negatively regulates hetF, and NsiR1.4 negatively regulates the hetP homolog alr3234. Therefore, sequence divergence between the different versions of NsiR1 allows regulation of diverse aspects of heterocyst differentiation through their effects on the amounts of HetF (influencing HetR accumulation) and Alr3234 (influencing commitment) specifically in heterocysts. Results NsiR1 interacts with the 5’ UTR of hetF. NsiR1.1 is transcribed antisense to the 5’ UTR of hetF in Nostoc sp. PCC 7120 and this arrangement is conserved along genomes of heterocystous cyanobacteria (27) (SI Appendix, Fig. S1 and S2), therefore we hypothesized that NsiR1.1 might regulate hetF expression. To explore this possibility the 5’-ends of hetF transcripts were analyzed by retrotranscription using an 166 oligonucleotide complementary to hetF 5’ UTR. 5’-ends corresponding to the TSS of hetF, position 4273166f (10, 27) in the Nostoc sp. PCC 7120 genome, were in fact observed in RNA samples isolated from both the wild type strain and the hetR mutant growing in the presence of combined nitrogen (indicated with an asterisk in Fig. 1B). However, when the cells were subjected to nitrogen deficiency, a HetR-dependent stop appeared in a position corresponding to that of the transcriptional start site of nsiR1.1 (4273242r, black triangle in Fig. 1B). Retrotranscriptase stops when it encounters an obstacle, for example, regions of strong secondary structure or a double stranded RNA such as a sRNA-mRNA duplex (32). Our results therefore suggest that the additional HetR-dependent stop might be due to binding of NsiR1, whose expression is HetR-dependent (27, 28), to hetF 5’ UTR. If NsiR1.1 negatively regulates hetF, we would expect less HetF protein in heterocysts than in vegetative cells because nsiR1 transcription is heterocyst-specific. To test this hypothesis, we constructed Nostoc sp. PCC 7120 strains that express, under the control of a constitutive promoter (rnpB), a translational fusion of the hetF 5’ UTR plus 60 nucleotides of its coding region to gfpmut2 (plasmid pMBA93, Fig. 1A). A strain expressing gfpmut2 and its original 5’UTR from the rnpB promoter was constructed as a control (plasmid pMBA92, Fig. 1A). As expected the strain bearing the control plasmid pMBA92 showed similar green fluorescence in vegetative cells or heterocysts (Fig. 1C). However, the strain bearing pMBA93 showed detectable green fluorescence in vegetative cells but not in heterocysts (Fig. 1C), suggesting heterocyst-specific regulation at the level of the hetF 5’ UTR. We then tested the interaction of NsiR1.1 with hetF 5’ UTR using a heterologous reporter assay in E. coli (33) in which the 5’ UTR of hetF plus 60 nucleotides of its coding region was fused to the gene sfgfp encoding superfolder GFP. Because identical copies (nsiR1.3 to nsiR1.9) are very similar to nsiR1.1 (27) (SI Appendix, Fig. S3), we also tested the interaction between the hetF 5’ UTR and the sRNA encoded in one of the identical copies (nsiR1.4). We coexpressed in E. coli the hetF::sfgfp fusion together with NsiR1.1, NsiR1.4 or a control, unrelated RNA. Cells carrying the hetF::sfgfp translational fusion showed significant fluorescence, indicating that the translational initiation region of hetF was functional in E. coli (Fig. 1D and E). The fluorescence of cells carrying the hetF::sfgfp fusion decreased about 70% or 40% of the control when the fusion was co-expressed with NsiR1.1 or NsiR1.4, respectively, suggesting that each of these versions of NsiR1 were able to interact with the 5’ UTR of hetF (Fig. 1D and E) when transcribed in a heterologous system. 167 Figure 1. NsiR1 interacts with hetF 5’ UTR. (A) Schematic representation of the genomic region encoding NsiR1 in Nostoc sp. PCC 7120 and constructs in plasmids pMBA92 and pMBA93. Bent arrows indicate the transcriptional start sites. Identical repeats of nsiR1 (3-9, nsiR1.4) are shown as red arrows. Repeats of nsiR1 with some sequence divergence (1, 2, 10, 11 and 12) are shown in light blue except repeat 1 (nsiR1.1), antisense to hetF, that is shown in green. (B) Primer extension analysis of the hetF transcript using RNA isolated from Nostoc sp. PCC 7120 wild type strain (WT) or a hetR mutant strain at different times after removal of combined nitrogen. The position of the TSS of hetF (4273166f) is indicated with an asterisk. The position corresponding to the TSS of nsiR1.1 (4273242r) is indicated with a black triangle. On the right, a sequencing reaction performed with the same oligonucleotide used for primer extension. (C) Top, confocal fluorescence images (red channel, green channel and bright field image) of filaments from strains bearing plasmids pMBA92 (control, gfp gene expressed from the rnpB promoter) or pMBA93 (translational fusion of the hetF 5’ UTR to the gfp gene expressed from the rnpB promoter) growing on top of nitrogenfree medium. Bottom, quantification of the signals for the green (GFP) and red (autofluorescence) channels. Heterocysts are indicated with numbers. (D) GFP fluorescence of E. coli cultures bearing different combinations of a plasmid expressing 5’ UTR hetF::sfgfp with plasmids expressing NsiR1.1, NsiR1.4 or an unrelated control RNA. The data are represented as mean and standard deviation of results from cultures corresponding to four independent colonies after normalization for cell density and subtraction of background fluorescence in cells bearing pXG-0. (E) FACS-based reporter assays. The plot shows the distribution of fluorescence in 10,000 E. coli cells bearing plasmids expressing 5’ UTR hetF:sfgfp and the unrelated control RNA (black), NsiR1.1 (green) or NsiR1.4 (red). The inset shows the average fluorescence of each sample. T-test P-value <0.05*;< 0.01*. 168 Taken together, the results obtained in Nostoc sp. PCC 7120 and E. coli suggested a post transcriptional negative regulation exerted by NsiR1 on hetF expression. The overlapping of NsiR1.1 and the hetF 5’ UTR extends between positions 20 and 78 of the UTR, 180 nucleotides upstream the start codon. To analyze the regulatory mechanism, we carried out in vitro footprinting assays of NsiR1.1 and hetF 5’ UTR interaction (Fig. 2). 32P-labeled hetF 5’ UTR (a fragment extending from the TSS at position −257 to +60 with respect to the start of the coding sequence) was incubated with unlabeled NsiR1.1 and probed with RNase T1, RNase A, or lead(II) acetate (Fig. 2A). Figure 2. In vitro footprinting assay of the interaction between NsiR1.1 and the hetF 5’ UTR. (A) 5’-labeled hetF 5’ UTR RNA was treated with RNase T1, RNase A, or lead(II) acetate (Pb+2) in the absence (-) or presence (+) of NsiR1.1. Samples were loaded on two gels that were run to different lengths. Vertical bars indicate the NsiR1.1 binding site (red), a region with increased sensitivity to the probes in the presence of NsR1.1 (green) and the region including the 169 start codon (orange). Nucleotides with differential sensitivity to RNase T1 or RNase A in the presence or absence of NsiR1.1 are indicated with colored triangles. Nucleotide positions of hetF 5’ UTR are shown on the left. C, untreated control; OH, alkaline ladder; T1s, RNase T1 sequencing ladder. (B, C) Models of the secondary structure of hetF 5’ UTR from Nostoc sp. PCC 7120 without (B) or with bound NsiR1.1 (C). Nucleotides corresponding to the areas highlighted by vertical bars in the gels are colored in the same way. Nucleotides with differential sensitivity to RNase T1 or RNase A outside the NsiR1 binding site are indicated with colored triangles. Start codon is in purple with a purple frame. A footprint was detected between positions 23 and 81 of hetF 5’ UTR (Fig. 2A, red bar), in agreement with the perfect antisense sequence complementary to NsiR1.1. In addition to this footprinting, we also detected some structural changes when hetF 5’ UTR was combined with NsiR1.1; a region between positions 98 and 104 of hetF 5’ UTR became more sensitive to RNase T1, RNase A and lead(II) acetate (Fig. 2A, green bar) and the region including the start codon became more protected (Fig. 2A, orange bar). These results indicate that upon binding of NsiR1 there is a structural reorganization of the 5’ UTR of hetF resulting in reduced access to the initiation codon (Fig. 2B and C) that could explain the reduction of HetF expression in heterocysts. Strains with altered levels of different versions of NsiR1. We have constructed strains with constitutive high levels of NsiR1.1 (OE_NsiR1.1) or NsiR1.4 (OE_NsiR1.4) transcribed from the strong trc promoter to the T1 terminator of the rrnB gene of E. coli (Fig. 3A). In addition, because we have introduced an NsiI restriction site (6 bp) at the 5’ end of nsiR1.1 or nsiR1.4 to facilitate cloning, this 6 nt tag allows the discrimination, based on its length, of molecules of NsiR1.1 and NsiR1.4 transcribed from their native promoters from molecules expressed from the trc promoter in the constructs introduced in Nostoc sp. PCC 7120. The location of nsiR1.1 in the 5’ UTR of hetF precludes the deletion of all copies of nsiR1 without removing a segment of hetF 5’ UTR. To deplete NsiR1 without altering the genomic region upstream of hetF we have followed a strategy similar to (34) and generated a Nostoc strain (OE_as_NsiR1) that constitutively expresses the sequence complementary to nsiR1.4 from the trc promoter (Fig. 3A). This antisense RNA would act as a sponge neutralizing the accumulation of nsiR1. 170 Figure 3. Nostoc sp. PCC 7120 strains with altered levels of NsiR1. (A) Scheme of DNA fragments cloned in the plasmids used to generate the different strains. Transcription start site (bent arrows), Rho-independent terminator of NsiR1 (small stem loop), T1 terminator (large stem loop), trc promoter, probe for nsiR1.1 (green line), probe for nsiR1.4 (red line), oligonucleotide used for strand-specific probe labeling (black triangle) and sequences corresponding to nsiR1.1 (green arrow) and nsiR1.4 (red arrows) are indicated. (B) Northern blots using RNA extracted from the different strains at different times after nitrogen removal hybridized with probes for NsiR1.1 (top), NsiR1.4 (middle) and 5S rRNA (bottom) as a loading control. Endogenous NsiR1 (black triangles) and the three RNA species constitutively expressed from the trc promoter, NsiR1.1 (green triangle), NsiR1.4 (red triangle) and as_NsiR1 (grey triangles), are indicated. (C) Primer extension analysis of hetF transcript with RNA isolated from the mutant strains at different time points after combined nitrogen removal. The position of the TSS of hetF (4273166f) is indicated with an asterisk. The position corresponding to the TSS of nsiR1.1 (4273242r) is indicated with a black triangle. Stops provoked by NsiR1.1 (green triangle) and NsiR1.4 (red triangles) constitutively expressed from the trc promoter are indicated. On the right, a sequencing reaction performed with the same oligonucleotide used for primer extension. (D) Accumulation of HetR in the mutant strains 18 h after combined nitrogen removal. Western blots were carried out with proteins extracted from three independent cultures of each strain using antibodies against HetR or GroEL (used as loading control). (E) Quantification of Western blot shown in (D). The data are presented as the mean ± standard deviation of the HetR signal normalized to the GroEL signal of the three independent clones of each strain shown in (D). The mean of HetR accumulation of the OE_C strain in each gel is used as 100%. **T-test P-value < 0.01. 171 Figure 6. Post-transcriptional regulation of alr3234 in Nostoc sp. PCC 7120 and commitment assay in strains with altered levels of NsiR1. (A) Top, confocal fluorescence images (red channel, green channel and bright field image) of filaments growing on top of nitrogen-free medium and carrying under the control of the rnpB promoter the gfp gene (plasmid pMBA92) (left) or a translational fusion of the alr3234 5’ UTR plus 60 nucleotides of the coding sequence to the gfp gene (plasmid pMB94) (right). Bottom, quantification of the signals for the green (GFP) and red (autofluorescence) channels. Heterocysts are indicated with numbers. (B) Determination of time of heterocyst commitment. After combined nitrogen removal (t=0h), ammonium was added at the indicated times and the frequency of heterocyst was determined 24-36 h after ammonium addition. For each strain the data are presented as the mean and standard deviation of two cultures from two independent clones. 178 Discussion Heterocyst differentiation is a complex process that requires precise regulatory mechanisms in order to establish exclusive transcriptional programs for different cell types (vegetative cells, immature pro-heterocysts and mature heterocysts). Heterocyst differentiation is triggered by mutual induction of expression of the transcriptional regulator NtcA and the master regulator HetR in vegetative cells becoming heterocysts (7, 8, 38). The accumulation of active HetR in pro-heterocysts seems to trigger the downstream cascade. NsiR1 was the first HetR-dependent sRNA identified in Nostoc sp. PCC 7120 (27), but it is also well conserved in nonheterocyst-forming filamentous cyanobacteria (29). However, while most non-heterocyst strains contain a single copy of NsiR1, heterocystous strains have multiple tandem copies with slight sequence divergence (SI Appendix, Fig. S1, Fig. S2). The increase in the number of copies of NsiR1, that correlates with the acquisition of heterocysts in evolution, and their sequence divergence, could allow some degree of functional specialization by regulating a variety of targets. According to the position of the TSS of hetF in Nostoc sp. PCC 7120 (10, 27), the nsiR1 copy closest to hetF (nsiR1.1) could function as an antisense of hetF 5’ UTR. This function of NsiR1.1 may be conserved in heterocyst-forming cyanobacteria because most of them have a nsiR1 homolog in the upstream region of hetF in antisense orientation (SI Appendix, Fig. S2). Co-expression experiments in E. coli indicate that NsiR1.1, and to a lower extent NsiR1.4, can repress the expression of hetF (Fig. 1D and E). A translational fusion of hetF 5’ UTR to GFP had reduced expression in heterocysts (Fig. 1C), as expected if its translation is repressed by NsiR1, that is expressed at an early stage during heterocyst development (31). In addition, primer extension analysis in Nostoc sp. PCC 7120 also suggested a putative interaction of NsiR1.1 with hetF 5’ UTR based on the stop of the retrotranscriptase (Fig. 1B). Two observations point to post-transcriptional regulation of translation rather than transcriptional interference as the mechanism for hetF regulation by NsiR1. First, negative regulation of hetF expression is observed in the E. coli system, in which NsiR1 and the hetF 5’ UTR are transcribed from E. coli promoters in separate plasmids. Second, the observation that NsiR1.1, which perfectly matches the hetF 5’ UTR, and NsiR1.4 which contains several nucleotide changes, repress hetF translation to a different extent (Fig. 1D), is indicative of post-transcriptional regulation by perfect (NsiR1.1) or imperfect (NsiR1.4) base pairing. Finally, the quantity of hetF mRNA did not significantly decrease when NsiR1 was present (Fig. 1B), indicating that NsiR1 did not affect hetF transcription. 179 The interaction site of NsiR1 with the hetF 5’ UTR was located relatively far upstream from the hetF start codon, precluding repression of translation by direct interaction of NsiR1 with the translation initiation region and pointing to a translational regulation based on a conformational change of the hetF 5’ UTR. In vitro footprinting data and computational analysis showed that the hetF 5’ UTR is highly structured and the interaction with NsiR1.1 provokes a conformational change in which the region of the start codon of hetF gets more protected (Fig. 2), providing a mechanistic explanation to the negative regulation of hetF expression by NsiR1.1. In order to study the function of NsiR1, we created strains of Nostoc sp. PCC 7120 with altered levels of different versions of NsiR1 (Fig. 3). The overexpression of NsiR1.1 or NsiR1.4 under the control of the trc promoter was irrespective of the nitrogen source and strong enough to overpass the endogenous expression of NsiR1, that was induced upon nitrogen removal. The OE_as_NsiR1 strain had reduced amounts of NsiR1, implying that the interaction between NsiR1 and its antisense results in its degradation (Fig. 3B). Primer extension assays with RNA extracted from strains with altered levels of different versions of NsiR1 confirmed the interaction of NsiR1.1 and NsiR1.4 with hetF 5’ UTR (Fig. 3C). Expression of hetF is low and the transcripts are unstable (17, 18), therefore we were unable to quantify changes in hetF expression in our mutant strains. Instead, because one of the features described for a hetF mutant is the accumulation of higher levels of HetR protein (18), we analyzed HetR protein levels as an indication of HetF function. In fact, strains OE_NsiR1.1 and, to a lesser extent, OE_NsiR1.4, had higher amounts of HetR, as expected for a strain that would have less HetF (Fig. 3D and E). This higher amount of HetR was not due to increased hetR transcription in strains OE_NsiR1.1 and OE_NsiR1.4 (SI Appendix, Fig. S4). Finally, OE_NsiR1.1 had a phenotype partially similar in some aspects to that described for a hetF mutant. A hetF mutant grows poorly in media without combined nitrogen, has enlarged cells and cannot differentiate heterocysts (18). OE_NsiR1.1 showed some filaments with larger cells in liquid cultures growing without combined nitrogen (SI Appendix, Fig. S5) and was delayed in the timing of commitment (Fig. 6B). Taken together, all these in vivo and in vitro results strongly support a role for NsiR1 in reducing the amount of HetF in cells becoming heterocysts. A computational analysis (27) identified alr3234, a hetP homolog described as a negative modulator of heterocyst commitment (20), as a potential target of NsiR1.4. We confirmed the interaction between NsiR1.4 and alr3234 mRNA by in vitro footprinting and coexpression in E.coli (Fig. 5). NsiR1.4 binds to a region that includes the start codon of alr3234, and thus could inhibit translation initiation by hampering ribosome access. One of the phenotypic features described for an alr3234 mutant is earlier commitment to heterocyst 180 differentiation (20). Timing of commitment in OE_NsiR1.4 was in fact shifted to earlier time points (about 2 hours earlier than OE_C) while OE_as_NsiR1 had a slight delay in commitment (Fig. 6B), supporting an inhibitory effect of NsiR1.4 on alr3234. The putative positive effect of NsiR1.4 on heterocyst commitment is also in line with the observation that OE_NsiR1.4 differentiates heterocyst even in the presence of nitrate (Fig. 4B). Alr3234 represents a brake that delays commitment, providing a lag time to avoid unnecessary differentiation under fluctuating conditions (20). We hypothesize that only when enough NsiR1 has accumulated as a consequence of constant nitrogen deficiency is the amount of Alr3234 reduced and the brake on commitment released. Despite the slight difference in sequence between NsiR1.1 and NsiR1.4 (SI Appendix, Fig. S3), strains OE_NsiR1.1 and OE_NsiR1.4 had different phenotypes. They grew differentially in solid media with different nitrogen sources (Fig. 4A). OE_NsiR1.4, but not OE_NsiR1.1, differentiated heterocysts with a normal pattern in solid media with nitrate as nitrogen source, a condition that suppresses heterocyst development in the wild type (Fig. 4B). Finally, these strains differ in the timing of commitment to heterocyst differentiation (Fig. 6B). Thus, these results suggest a divergence in the functions of NsiR1.1 and NsiR1.4, that could be explained by their different effects on hetF and alr3234 or by interactions with additional unknown targets specific for each of them. The positive effect of NsiR1.4 on differentiation was in accordance with the observation that OE_as_NsiR1 (a strain with a lower amount of the NsiR1 pool, mostly formed by NsiR1.4) was delayed in differentiation (Fig. 4D) and slightly delayed in commitment (Fig. 6B). In this work we describe how different versions of one sRNA regulate different targets. Other instances of sRNAs encoded by multiple repeats have been described in diverse bacteria (39). In most cases there is a regulatory redundancy and the different copies of one given sRNA act additively to regulate one particular target. In other cases they are non-redundant, regulating different targets. NsiR1.1, on one side, and the identical copies NsiR1.3-NsiR1.9 on the other, represent divergent versions of a sRNA with partially redundant targets but possibly also different targets. Present models of events leading to heterocyst differentiation and patterning are based on transcriptional regulation and protein-protein interactions. Beyond those mechanisms, we here identify a role for a sRNA that specifically accumulates in heterocysts. We propose that, even though hetF and alr3234 are transcribed constitutively, upon nitrogen deprivation the early and specific expression of NsiR1.1 and NsiR1.4 in pro-heterocysts would 181 provoke the post-transcriptional repression of hetF and alr3234, resulting on a local accumulation of HetR in pro-heterocyst (triggering the downstream cascade of heterocyst differentiation) and releasing the brake of alr3234 inhibition of the commitment of proheterocyst into differentiation (Fig. 7). So, the specific post-transcriptional regulation exerted by different versions of NsiR1 in the multicellular cyanobacterium Nostoc sp. PCC 7120 is a good example of how complexity can emerge from the superposition of simple patterns, constitutive transcription (hetF, alr3234) plus a cell-type specific negative post-transcriptional regulation mediated by a small RNA (NsiR1). Further study of additional potential targets for NsiR1, and the characterization of additional heterocyst-specific non-coding RNAs (28) could reveal the importance of these molecules in the regulation of this developmental process. Figure 7. Model of the function of NsiR1.1 and NsiR1.4 in the regulation of heterocyst differentiation and commitment. Transcriptional regulation is indicated by black arrows. Negative post transcriptional regulation by NsiR1 is indicated by red lines with blunt ends. HetF regulation of HetR is indicated by blue lines. Thin lines indicate weaker regulation than thick lines. Different letter size portraits the different relative amounts of the proteins or RNAs displayed in the two types of cells. Question marks indicate possible additional targets of NsiR1.1 and NsiR1.4. Green dashed line delimits events occurring in heterocysts (left) from events occurring in vegetative cells (right). 182 Materials and Methods Strains and growth conditions. The different Nostoc sp. PCC 7120 strains used in this work (SI Appendix,Table S1) were grown in BG11 medium (40) as detailed in SI Appendix, SI Materials and Methods. Escherichia coli strains (SI Appendix, Table S1) were grown in LB medium, supplemented with appropriate antibiotics (41). Construction of Nostoc sp. PCC 7120 derived strains. Details on the construction of plasmids (SI Appendix, Tables S2) are given in SI Appendix, SI Materials and Methods. In vitro synthesis and labelling of RNA. RNAs were in vitro transcribed (MEGAscript Kit; Life Technologies, #AM1330) from PCR-generated DNA templates using the primers indicated in SI Appendix, Table S3. Details of the transcription, purification and labeling procedures are given in SI Appendix, SI Materials and Methods. In vitro structure probing and footprinting. RNase T1, RNase A and lead(II) acetate were used for structure probing with in vitro transcribed RNAs (SI Appendix, Table S4) essentially as described (42). Details of the procedures are given in SI Appendix, SI Materials and Methods RNA isolation, Northern blot analysis and primer extension assays. Total RNA was isolated using hot phenol as described (43) with modifications (29). Strand specific 32P probe labeling and Northern blots were performed as detailed in SI Appendix. Primer extension of 5´ ends of hetF was performed as previously described (44) using oligonucleotide 105 (SI Appendix, Table S3) labeled with [ -32P] ATP and 10 µg of total RNA. 183 Heterocyst differentiation and commitment assay. Heterocysts were counted after staining with Alcian blue as described (45). Number of heterocysts and commitment time were determined as described in SI Appendix, SI Materials and Methods. Nitrate and nitrite reductase assay. Nitrate and nitrite reductase were performed as described (46, 47). More details in SI Appendix, SI Materials and Methods. Western blot analysis. For HetR quantitation in Nostoc sp. PCC 7120 filaments, protein extracts were prepared using glass beads and Western blots were performed as detailed in SI Appendix, SI Materials and Methods. Antibodies against Nostoc sp. PCC 7120 HetR (48) and E.coli GroEL (Sigma-Aldrich) were used. Quantitation of alr3234::GFP in E. coli cells was performed as detailed in SI Appendix, SI Materials and Methods. Antibodies against GFP (monoclonal, Roche) and E. coli GroEL (Sigma-Aldrich) were used. Reporter assay for in vivo verification of targets. We used the GFP reporter system described in (49) with the superfolder GFP (sfGFP) and plasmid pXG10-SF (33) for the experimental target verification in E.coli. The construction of the different GFP fusions used (SI Appendix, Table S5) and site directed mutagenesis are described in SI Appendix, SI Materials and Methods. Fluorescence microscopy. Fluorescence of Nostoc sp. PCC 7120 filaments carrying plasmid pMBA92, pMBA93 and pMBA94 (SI Appendix, Table S6) growing on top of solidified nitrogen-free medium, was analyzed and quantified as described (31) using a Leica HCX PLAN-APO 63X 1.4 NA oil immersion objective attached to a Leica TCS SP2 confocal laser-scanning microscope. Samples were excited at 488 nm by an argon ion laser and the fluorescent emission was monitored by collection across windows of 500–538 nm (GFP) and 630–700 nm (cyanobacterial autofluorescence). 184 Computational and statistical methods. Computational and statistical methods are described in detail in SI Appendix, SI Materials and Methods. Acknowledgements This work was supported by grants BFU2013-48282-C2-1 from Ministerio de Economa y Competitividad, and BFU2016-74943-C2-1-P from Agencia Estatal de Investigacin (AEI), Ministerio de Economa, Industria y Competitividad, both cofinanced by the European Regional Development Fund, to A.M.P. M.B.A is the recipient of a predoctoral contract from Ministerio de Educacin, Cultura y Deporte, Spain (FPU014/05123 and EST16-00088). We are grateful to Wolfgang R. Hess (Freiburg University, Germany) for valuable comments on the manuscript. References 1. E. Flores, A. Herrero, Compartmentalized function through cell differentiation in filamentous cyanobacteria. Nat. Rev. Microbiol. 8, 39-50 (2010). 2. E. Flores, S. Picossi, A. Valladares, A. Herrero, Transcriptional regulation of development in heterocyst-forming cyanobacteria. Biochim. Biophys. Acta. Gene. Regul. Mech. 1862, 673- 684 (2019). 3. A. Herrero, J. Stavans, E. Flores, The multicellular nature of filamentous heterocyst-forming cyanobacteria. FEMS Microbiol Rev 40, 831-854 (2016). 4. A. M. Muro-Pastor, W. R. Hess, Heterocyst differentiation: from single mutants to global approaches. Trends Microbiol. 20, 548-557 (2012). 5. A. Herrero, A. M. Muro-Pastor, E. Flores, Nitrogen control in cyanobacteria. J. Bacteriol. 183, 411-425 (2001). 6. W. J. Buikema, R. Haselkorn, Characterization of a gene controlling heterocyst differentiation in the cyanobacterium Anabaena 7120. Genes Dev. 5, 321-330 (1991). 7. T. A. Black, Y. Cai, C. P. Wolk, Spatial expression and autoregulation of hetR, a gene involved in the control of heterocyst development in Anabaena. Mol. Microbiol. 9, 77-84 (1993). 8. A. M. Muro-Pastor, A. Valladares, E. Flores, A. Herrero, Mutual dependence of the expression of the cell differentiation regulatory protein HetR and the global nitrogen regulator NtcA during heterocyst development. Mol. Microbiol. 44, 1377-1385 (2002). 185 9. R. Rajagopalan, S. M. Callahan, Temporal and spatial regulation of the four transcription start sites of hetR from Anabaena sp. strain PCC 7120. J. Bacteriol. 192, 1088-1096 (2010). 10. J. Mitschke, A. Vioque, F. Haas, W. R. Hess, A. M. Muro-Pastor, Dynamics of transcriptional start site selection during nitrogen stress-induced cell differentiation in Anabaena sp. PCC 7120. Proc. Natl. Acad. Sci. U.S.A 108, 20130-20135 (2011). 11. H. S. Yoon, J. W. Golden, Heterocyst pattern formation controlled by a diffusible peptide. Science 282, 935-938 (1998). 12. S. M. Callahan, W. J. Buikema, The role of HetN in maintenance of the heterocyst pattern in Anabaena sp. PCC 7120. Mol. Microbiol. 40, 941-950 (2001). 13. J. Elhai, I. Khudyakov, Ancient association of cyanobacterial multicellularity with the regulator HetR and an RGSGR pentapeptide-containing protein (PatX). Molecular Microbiology 110, 931-954 (2018). 14. X. Huang, Y. Dong, J. Zhao, HetR homodimer is a DNA-binding protein required for heterocyst differentiation, and the DNA-binding activity is inhibited by PatS. Proc. Natl. Acad. Sci. U.S.A 101, 4848-4853 (2004). 15. D. D. Risser, S. M. Callahan, Genetic and cytological evidence that heterocyst patterning is regulated by inhibitor gradients that promote activator decay. Proc. Natl. Acad. Sci. U.S.A 106, 19884-19888 (2009). 16. R. Zhou et al., Evidence that HetR protein is an unusual serine-type protease. Proc. Natl. Acad. Sci. U.S.A 95, 4959-4963 (1998). 17. F. C. Wong, J. C. Meeks, The hetF gene product is essential to heterocyst differentiation and affects HetR function in the cyanobacterium Nostoc punctiforme. J. Bacteriol. 183, 2654- 2661 (2001). 18. D. D. Risser, S. M. Callahan, HetF and PatA control levels of HetR in Anabaena sp. strain PCC 7120. J. Bacteriol. 190, 7645-7654 (2008). 19. K. C. Higa, S. M. Callahan, Ectopic expression of hetP can partially bypass the need for hetR in heterocyst differentiation by Anabaena sp. strain PCC 7120. Mol. Microbiol. 77, 562-574 (2010). 20. P. Videau et al., The heterocyst regulatory protein HetP and its homologs modulate heterocyst commitment in Anabaena sp. strain PCC 7120. Proc. Natl. Acad. Sci. U.S.A 113, E6984–E6992 (2016). 21. M. Kopf, W. R. Hess, Regulatory RNAs in photosynthetic cyanobacteria. FEMS Microbiol. Rev. 39, 301-315 (2015). 22. E. G. H. Wagner, P. Romby, Small RNAs in bacteria and archaea: who they are, what they do, and how they do it. Adv. Genet. 90, 133-208 (2015). 186 23. J. M. Silvaggi, J. B. Perkins, R. Losick, Genes for small, noncoding RNAs under sporulation control in Bacillus subtilis. J. Bacteriol. 188, 532-541 (2006). 24. A. Marchais, S. Duperrier, S. Durand, D. Gautheret, P. Stragier, CsfG, a sporulation-specific, small non-coding RNA highly conserved in endospore formers. RNA Biol. 8, 358-364 (2011). 25. R. A. Mars, P. Nicolas, E. L. Denham, J. M. van Dijl, Regulatory RNAs in Bacillus subtilis: a Gram-positive perspective on bacterial RNA-mediated regulation of gene expression. Microbiol. Mol. Biol. Rev. 80, 1029-1057 (2016). 26. M. Schmalisch et al., Small genes under sporulation control in the Bacillus subtilis genome. J. Bacteriol. 192, 5402-5412 (2010). 27. D. Ionescu, B. Voβ, A. Oren, W. R. Hess, A. M. Muro-Pastor, Heterocyst-specific transcription of NsiR1, a non-coding RNA encoded in a tandem array of direct repeats in cyanobacteria. J. Mol. Biol. 398, 177-188 (2010). 28. M. Brenes-Álvarez et al., Elements of the heterocyst-specific transcriptome unravelled by co-expression analysis in Nostoc sp. PCC 7120. Environ. Microbiol. 21, 2544-2558 (2019). 29. M. Brenes-Álvarez, E. Olmedo-Verd, A. Vioque, A. M. Muro-Pastor, Identification of conserved and potentially regulatory small RNAs in heterocystous cyanobacteria. Front. Microbiol. 7, 48 (2016). 30. E. Olmedo-Verd, M. Brenes-Álvarez, A. Vioque, A. M. Muro-Pastor, A heterocyst-specific antisense RNA contributes to metabolic reprogramming in Nostoc sp. PCC 7120. Plant Cell Physiol. 60, 1646-1655 (2019). 31. A. M. Muro-Pastor, The heterocyst-specific NsiR1 small RNA is an early marker of cell differentiation in cyanobacterial filaments. mBio 5, e01079-14 (2014). 32. J. Jagodnik, A. Brosse, T. N. Le Lam, C. Chiaruttini, M. Guillier, Mechanistic study of basepairing small regulatory RNAs in bacteria. Methods 117, 67-76 (2017). 33. C. P. Corcoran et al., Superfolder GFP reporters validate diverse new mRNA targets of the classic porin regulator, MicF RNA. Mol. Microbiol. 84, 428-445 (2012). 34. U. Dühring, I. M. Axmann, W. R. Hess, A. Wilde, An internal antisense RNA regulates expression of the photosynthesis gene isiA. Proc. Natl. Acad. Sci. U.S.A. 103, 7054-7058 (2006). 35. J. E. Frías, E. Flores, Negative regulation of expression of the nitrate assimilation nirA operon in the heterocyst-forming cyanobacterium Anabaena sp. strain PCC 7120. J. Bacteriol. 192, 2769-2778 (2010). 36. J. E. Frías, E. Flores, A. Herrero, Nitrate assimilation gene cluster from the heterocystforming cyanobacterium Anabaena sp. strain PCC 7120. J. Bacteriol. 179, 477-486 (1997). 187 air/CO2 mixture (1% v/v). After 4 hours of incubation, cells were concentrated by centrifugation to a final concentration of 50 µg chlorophyll/mL. Nitrate reductase (16) and nitrite reductase (17) assays were performed with an amount of cells corresponding to 5 and 25 µg chlorophyll respectively. Dithionite-reduced methyl viologen was used as the reductant in cells made permeable with mixed alkyltrimethylammonium bromide. Western blot analysis For HetR Western blot analysis, crude extracts were prepared using glass beads. Cells from 25 mL cultures of OE_C, OE_NsiR1.1, OE_NsiR1.4 and OE_as_NsiR1 subjected to 18 h of nitrogen deficiency were harvested by filtration, washed with TE50/100 buffer (50 mM Tris pH 8, 100 mM EDTA) and resuspended in 500 µL of resuspension buffer (50 mM Tris-HCl pH 8.0, 2 mM 2-mercaptoethanol) containing a protease inhibitor cocktail (Roche). The cellular suspension was mixed with glass beads (SIGMA, 200 µm) in an Eppendorf tube and subjected to 7 cycles of 1 min vortexing plus 1 min of cooling on ice. Cell extract was separated from cell debris and unbroken cells by centrifugation (3 min at 3000 x g at 4 °C). The soluble fraction was obtained by centrifugation of the crude extract at 16.000 x g for 30 min at 4 °C. The protein concentration was determined by the Lowry procedure (18). 40 µg of proteins were fractionated on 10% SDS-PAGE. Antibodies against Nostoc sp. PCC 7120 HetR (19) and E. coli GroEL (Sigma- Aldrich) were used. For Alr3234::GFP Western blot analysis, E.coli cells from stationary phase cultures were harvested and resuspended in SDS-PAGE loading buffer. Proteins were fractionated in 10% SDS-PAGE. Antibodies against GFP (Roche) and E. coli GroEL (Sigma-Aldrich) were used. ECL Plus immunoblotting system (GE Healthcare) was used to detect the different primary antibodies using anti-rabbit (Sigma-Aldrich) or anti-mouse (Bio-Rad) horse-radish peroxidase conjugated secondary antibodies. Reporter assay for in vivo verification of targets We used the GFP reporter system described in (20) with the superfolder GFP (sfGFP) and plasmid pXG10-SF (21) for the experimental verification of targets in E.coli. The 5’UTR of hetF from the TSS at position –257 (4273166f) plus 60 nucleotides within its coding region was amplified using oligonucleotides 555 + 557. The GTG start codon was changed to ATG using 194 overlapping PCRs and oligonucleotides 555, 556, 557 and 558. The information about TSS was taken from (22). The whole intergenic region alr3233-alr3234 plus 60 nucleotides within the coding region of alr3234, containing the predicted NsiR1 interaction sequence, was amplified using oligonucleotides 553 + 870. PCR products containing hetF::sfgfp and alr3234::sfgfp were digested with NsiI-NheI and cloned into NsiI-NheI-digested pXG10-SF vector, resulting in translational fusions of truncated HetF or Alr3234 to sfGFP (plasmids pMML2 and pMBA86, respectively, SI Appendix, Table S2). To express NsiR1.1 and NsiR1.4 in E.coli, the genes were amplified from genomic DNA using primers 559 (5’ phosphorylated) + 561 or 560 (5’ phosphorylated) + 562, respectively. These PCR products were digested with XbaI and fused to a plasmid backbone that was amplified from pZE-12luc with primers PlacOB and PLlacOD and digested with XbaI, rendering pMML3 and pMML4 (SI Appendix, Table S2). For the mutagenesis of 5’UTR of alr3234, a plasmid similar to pMBA86 was constructed using, in addition to the mentioned oligonucleotides, overlapping PCR with primers 871 + 872 containing the desired changes (pMBA87, SI Appendix, Table S2 and Table S3). The positions for the mutation of alr3234 5’UTR were selected based on hybridization energies predicted by IntaRNA (23). Different combinations of plasmids were introduced in E. coli DH5α and GFP fluorescence measurements were done by flow cytometry or with a microplate reader (Varioskan) using liquid cultures from individual colonies, as previously described (24). Sequences of inserts from plasmids used in the heterologous reporter system are shown in SI Appendix, Table S5. Fluorescence microscopy Fluorescence of Nostoc sp. PCC 7120 filaments carrying plasmid pMBA92, pMBA93 and pMBA94 (SI Appendix, Table S6) growing on top of solidified nitrogen-free medium, were analyzed and quantified as described (25) using a Leica HCX PLAN-APO 63× 1.4 NA oil immersion objective attached to a Leica TCS SP2 laser-scanning confocal microscope. Samples were excited at 488 nm by an argon ion laser and the fluorescent emission was monitored by collection across windows of 500–538 nm (GFP) and 630–700 nm (cyanobacterial autofluorescence). 195 Computational methods To detect nsiR1 homologs in the upstream region of hetF, the HetF sequence of Nostoc sp. PCC 7120 was used in a blastp search in IMG (26) against 100 cyanobacterial genomes, that were those used in (27), excluding those corresponding to the fast-evolving Prochlorococcus clade (SI Appendix, Figure S1). We took 4 kb upstream from the 50 hetF homologs that were identified and used one of the identical repeats of nsiR1 in Nostoc sp. PCC 7120 as input in a blastn search to predict putative nsiR1 homologs in sequences upstream the predicted hetF homologs. Sequences of nsiR1 homologs that were not in the upstream region of hetF were taken from (11).The interaction between NsiR1.1 and NsiR1.4 with alr3234 mRNA was analyzed using intaRNA (23). Quantification and statistical analysis The students’ t test was used to determine statistical significance. Number of biological samples can be found in the figure legends. 196 Figure S1. Co-occurrence of nsiR1 and hetF. Genomes are arranged and color-coded on the left as in (27) as follows: filamentous non-heterocystous in green, heterocystous ramified in blue, heterocystous non-ramified in pink, baeocystous in orange and unicellular in black. Presence or absence of a hetF homolog is indicated. Presence or absence of homologs of nsiR1, number of putative nsiR1 copies and genomic location of the nsiR1 arrays upstream to hetF (light green) or in a different genomic region (yellow) are also indicated. Nostoc sp. PCC 7120 is highlighted in red. 197 Figure S2. Genomic regions of nsiR1 repeats in 40 genomes. nsiR1 repeats are indicated by small arrows with solid colors (red for identical repeats, light blue for repeats with sequence divergence and green for repeats that might be antisense to hetF 5’-UTR). Genes are indicated by arrows with faded colors. Homologous genes are shown in the same color. hetF homologs are shown by dark grey arrows on the right side of the regions. The TSS of hetF in Nostoc sp. PCC 7120 is indicated by a bent arrow. The coherent clade of heterocyst-forming cyanobacteria is highlighted by a red frame. Nostoc sp. PCC 7120 is indicated in red. 198 Figure S3. Structure probing of NsiR1.1 and NsiR1.4. (A) Secondary structure model of NsiR1.1 and NsiR1.4. Black and white triangles indicate the positions sensitive to hydrolysis by RNase T1 or lead(II) acetate (Pb), respectively. (B) 5'-end labelled NsiR1.1 (left) or NsiR1.4 (right) were incubated with RNase T1 or lead(II) acetate and the resulting fragments analyzed on an 8% polyacrylamide sequencing gel. C, untreated control; OH, alkaline ladder; T1s, RNase T1 ladder. Nucleotide positions of NsiR1.1 or NsiR1.4 are shown on the left. 199 Figure S4. Accumulation of hetR transcripts in OE_C, OE_NsiR1.1, OE_NsiR1.4 and OE_as_NsiR1 strains. (A) hetR expression. Northern blots with RNA isolated from the indicated strains at different time points (indicated in hours) after removal of combined nitrogen were hybridized with probes for hetR (top panels) and rnpB (bottom panels), used as a loading control. Membranes are the same used in Fig 4D. Sizes (in nucleotides) are indicated on the left. (B) Quantification of hetR expression. The data are normalized to the amount of rnpB. 200 Figure S5. Heterocyst differentiation in strains with altered amounts of NsiR1. Bright field images of filaments of OE_C, OE_NsiR1.1, OE_NsiR1.4 and OE_as_NsiR1 strains growing in liquid media containing ammonium (NH4+) or growing in liquid media without combined nitrogen 24 h after nitrogen removal (N2). Filaments with enlarged irregular cells are pointed by black arrows. Scale bars, 25 µm. 201 Table S1. Strains Strain Description Reference Escherichia coli DH5α Used for routine transformation (28) Nostoc sp. PCC 7120 Wild type Pasteur Culture Collection 216 Non-functional HetR. S179N mutation. (1) OE_C SmRSpR, pMBA51 inserted in plasmid alpha. T1 terminator of E. coli rrnB gene expressed constitutively from the trc promoter. (5) OE_NsiR1.1 SmRSpR, pMBA77 inserted in plasmid alpha. NsiR1.1 transcribed constitutively from the trc promoter. This work OE_NsiR1.4 SmRSpR, pMBA78 inserted in plasmid alpha. NsiR1.4 expressed constitutively from trc promoter. This work OE_as_NsiR1 SmRSpR, pMBA42 inserted in plasmid alpha. Antisense to NsiR1 transcribed constitutively from the trc promoter. This work 202 Table S2. Plasmids Name Description Reference pJV300 Control plasmid expresses a ~50 nt transcript derived from the T1 terminator of E.coli rrnB gene. (29) pRL278 NmR, shuttle vector for conjugation in Nostoc with a Neomycin resistance cassette. (7) pMBA20 ApRSmRSpR, plasmid for the overexpression of transcripts from the rnpB promoter and followed by the T1 transcriptional terminator of E. coli rrnB gene. (5) pMBA37 ApRSmRSpR, plasmid for the overexpression of transcripts from the trc promoter and followed by the T1 terminator of E. coli rrnB gene, used as transcriptional terminator. (5) pMBA51 ApRSmRSpR, control plasmid, expresses a 56 nt transcript derived from the T1 terminator of E. coli rrnB gene from the trc promoter. (5) pXG0 CmR, control plasmid without GFP (20) pXG10-SF CmR, plasmid for construction of translational sfGFP fusions. (21) pZE12-luc ApR, plasmid used for the expression of NsiR1 under control of the PLlac-O promoter. (30) pMBA42 ApRSmRSpR, plasmid based on pMBA37 for the overexpression of an antisense to NsiR1.1. This work pMBA77 ApRSmRSpR, plasmid based on pMBA37 for the overexpression of NsiR1.1. This work pMBA78 ApRSmRSpR, plasmid based on pMBA37 for the overexpression of NsiR1.4. This work pMBA86 CmR, sfGFP reporter plasmid based on pXG10-SF containing the alr3234 5’UTR plus sequences encoding the first 20 amino acids of Alr3234. This work pMBA87 CmR, same as pMBA86 but with T4G and G9A substitutions in the alr3234 5’UTR (alr3234 MUT). This work pMBA91 ApRNmR, Kanamycin/ Neomycin resistance version of pMBA20. This work pMBA92 ApRNmR, plasmid based on pMBA91 expressing from rnpB promoter the gene gfpmut2 plus its 5’UTR. This work pMBA93 ApRNmR, plasmid based on pMBA91 expressing from the rnpB promoter a translational fusion of hetF 5’UTR plus sequences encoding the first 20 amino acids of HetF fused to GFP. This work pMBA94 ApRNmR, plasmid based on pMBA91 expressing from rnpB promoter a translational fusion of alr3234 5’UTR plus sequences encoding the first 20 amino acids of Alr3234 fused to GFP. This work pMML2 CmR, sfGFP reporter plasmid based on pXG10-SF containing the hetF 5’UTR plus sequences encoding the first 20 amino acids of HetF (with mutation in start codon GTG -> ATG). This work pMML3 ApR, plasmid based on pZE12-luc expressing NsiR1.1. This work pMML4 ApR, plasmid based on pZE12-luc expressing NsiR1.4. This work 203