|Home | About | Journals | Submit | Contact Us | Français|
The first plant glycine-rich proteins (GRPs) have been isolated more than 20 years ago based on their specific expression pattern and/or modulation by several biotic and abiotic factors. This superfamily is characterized by the presence of a glycine-rich domain arranged in (Gly)n-X repeats. The presence of additional motifs, as well as the nature of the glycine repeats, groups them in different classes. The diversity in structure as well as in expression pattern, modulation and sub cellular localization have always indicated that these proteins, although classified as members of the same superfamily, would perform different functions in planta. Only now, two decades later, with the first functional characterizations of plant GRPs their involvement in diverse biological and biochemical processes are being uncovered. Here, we review the so far ascribed functions of plant GRPs.
In plants, glycine-rich proteins (GRPs) are characterized by the presence of semi-repetitive glycine-rich motifs. In general, those genes present developmentally regulated and tissue-specific expression pattern. In several plant genera, their expression is also modulated by biotic and abiotic factors.1 Plant GRPs are classified based on their general structure, taking into consideration the arrangement of the glycine repeats as well as the presence of conserved motifs1–3 (Fig. 1). Class I GRPs may contain a signal peptide followed by a high glycine-content region with (GGX)n repeats. Class II GRPs may also contain a signal peptide and present a characteristic cysteine-rich C-terminal. Class III GRPs may present a signal peptide and contain a lower glycine content compared to other classes. Among those proteins, the oleosin domain is the signature motif for a sub-group of this class. Class IV GRPs are also known as RNA-binding GRPs. Besides the glycine-rich domain, they present either a RNA-recognition motif (RRM) or a cold-shock domain (CSD). They may also present CCHC zinc-fingers in their structure. Based on the diversity of domain arrangements, Class IV GRPS are subdivided into IVa (which contain one RRM motif besides the glycine-rich domain), IVb (one RRM and a CCHC zincfinger), IVc (a cold-shock domain and two or more zinc-fingers) and IVd (two RRMs). The identification of a different set of GRPs with a high glycine content but with mixed patterns of repeats from eucalyptus,3 also present in Arabidopsis and rice genomes (Galvão, Cardeal, Mangeon and Sachetto-Martins unpublished results), led to the proposition of a new class of GRPs (Class V).
In the literature, several names were given to different plant GRPs over the years (Table 1). In some cases, the same name was attributed to different GRPs and some symbols ascribed to GRPs were previously used in the literature to designate genes from different families. In order to facilitate the reference of GRPs, we propose here a unified nomenclature (Table 1). We adopted the three-letter system with the indication of the species from which the gene is from. Duplicated names, as well as symbols already used in the literature to refer to proteins other than GRPs, were disregarded as a way of avoiding misleading nomenclature. Hereafter, we refer to plant GRPs with the proposed nomenclature for a matter of clarity.
In the past few years, data towards the functional characterization of plant GRPs have been amassed. Several biological processes appear to have GRPs as players. Some GRPs roles could be deduced by previous modulation data, expression pattern or localization, but some are quite surprising. In this review, we go over the current literature to summarize the recent advances in GRP function in plants. Since past reviews already focus in the tissue expression pattern, subcellular localization and modulation of gene expression,1,4 here we only explore works regarding functional characterization of plant GRPs.
A structural function in cell walls has been proposed for plant GRPs since the isolation of the first GRP genes.5,6 In analogy to HRGPs, those proteins have been proposed to act as scaffold or agglutinating agents for deposition of cell wall constituents.7 Indeed functional results point to different GRPs acting as cell wall constituents. The French bean PvGRP1.8, a Class I GRP, has been extensively characterized biochemically. It is localized to unlignified primary cell wall8 and it is implicated in protoxylem development.9 The amassed data suggest that PvGRP1.8 performs a structural role in cell wall as a part of a repair system of the protoxylem10 and that it connects lignin rings leading to cell wall fortification.11
Recently, a reverse genetics approach was used to characterize a glycine-rich protein gene from Arabidopsis that according to microarray results could be involved in secondary cell wall formation. The analysis indicated that this GRP is implicated in the maintenance of protoxylem structure.12
Yeast two-hybrid experiments identified the interaction between AtGRP9 and a cinnamyl alcohol dehydrogenase (AtCAD5), an enzyme from the lignin biosynthesis pathway. These results suggest the involvement of this Class I GRP in lignin biosynthesis and/or deposition.13
Interestingly, another GRP that belongs to a different class is also involved in cell wall structure. The Class II GRP from tobacco NtCIG1 presents a structural role. This GRP, present in the vascular tissue, enhances callose deposition in the cell wall.14 A non-GRP called GrIP interacts with NtCIG1 in order to increase its levels of protein accumulation, representing then sequential acting players in the callose deposition biochemical pathway.15
The recurrent observation of GRP genes modulation by pathogens led to the proposition of the involvement of GRPs in plant defense.1 NtCIG1 gene was isolated from a subtractive screening for genes induced specifically by low concentrations of cadmium. Exposure of tobacco plants to this condition was known to block the systemic spread of the turnip vein-clearing tobamovirus (TVCV). Through transgenic analysis, it was demonstrated that the altered levels of NtCIG1 affected the virus movement in the plant. This containement of the virus was due to the callose deposition caused by NtCIG1.14 In this case, the structural role of a GRP (see above) has implications in plant defense. This finding is quite remarkable once it describes the mechanism through which the resistance is conferred.
Another example of the involvement of GRPs in plant defense comes from studies from the interaction between Arabidopsis and Pseudomonas syringae. This bacterium uses a type III protein secretion system to inject in plants effector proteins that cause disease. This type III efector is a mono-ADP-ribosyltransferase (HopU1-His) that has RNA-binding proteins presenting RRM motifs as its substrate. Biochemical analysis indicates that two Class IVa GRPs—AtRBG7 and AtRBG8—are indeed ribosylated by this type III efector in vitro. Detailed studies with AtRBG7 have shown that two arginine residues in the RRM motif are the target of HopU1-His and none of the arginines present in the glycine-rich domain were ribosylated. Further analyses using Atrbg7 mutant have shown that they are more susceptible to P. syringae than wild-type plants, suggesting that AtRBG7 plays a role in innate immunity.16
In search for peptides with antimicrobial activity, two glycine-rich peptides presenting GGH repeats were isolated from roots of Capsella bursa-pastoris. The two peptides named Sheperin I and II presented activity against several bacteria and fungi. Both peptides are encoded by a single gene, here referred as CbGRS1. This gene is expressed exclusively in roots.17
A GRP presenting an oleosin domain on its structure—here AtOGB3—is required for pollen hydration. The analysis of the pollen from Atogb3 mutant has shown that the observed delay on pollen hydration was not caused by a disturbed ability to absorb water. Therefore, it was derived from a failure to interact with stigma. All other aspects of pollen were normal and subsequent steps of pollination occurred properly, although with a delay caused from the pollen hydration step. This delay rendered a disavantage to mutant pollen on competition assays. This disavantage was overcome if the mutant pollen was developed in the presence of wild-type pollen, such as in heterozygous plants for the mutation, suggesting that a secreted substance from wild-type plants, probably AtOGB3, can complement the mutant pollen phenotype.18 The functional characterization of other Oleosin-GRPs could clarify if this function is conserved throughout this group of GRPs.
The implication of a GRP in signal transduction events has been suggested from protein interaction experiments. A yeast twohybrid screen identified two GRPs—AtGRP3 and AtGRP3S—as interactors of the extracelluar domain of the cell wall associated protein kinase WAK1. Co-immunoprecipitation experiments from Arabidopsis tissues have shown that, in planta, a complex of AtGRP3, WAK1 and KAPP is present.19 KAPP, a kinase-associated protein phosphatase is also known to interact with a number of receptor-like kinases.20 In vitro truncation experiments indicate that the cysteine-rich C-terminus of AtGRP3 is necessary and sufficient for its interaction with WAK1. AtGRP3 was completely absent from Arabidopsis protoplasts, suggesting an extracellular localization for this protein. In experiments using these protoplasts, exogenous application of AtGRP3 induced its own expression as well as WAK1 expression, indicating a positive feedback loop that enhances AtGRP3-WAK1 signaling. Both genes are induced by salicylic acid. AtGRP3 also induced the expression of PR1 gene from the salicylic acid-induced pathogen response pathway, indicating that AtGRP3-WAK1 signaling may be involved in plant-pathogen interaction responses.19
Several class IV GRPs are modulated under osmotic stress conditions and by the phytohormone abscisic acid (ABA). The RNA-binding activity of those proteins has been biochemically demonstrated, suggesting that they may be involved in RNA stabilization, processing and transport. For some of those proteins an RNA-chaperone activity has been demonstrated. In vitro binding experiments have shown that AtRBG4 protein is able to bind RNA, ssDNA and dsDNA in a non sequence-specific manner.21
In Arabidopsis, two GRPs here named AtCSG1 and AtCSG2 were shown to bind homoribopolymers, ssDNA and dsDNA.22,23 AtCSG1 binds to poly(G) and poly(U) preferentially.22 It was shown that the N-terminus of the protein binds specifically to poly(U), while the C-terminus binds specifically to poly(G). The same pattern was also observed for the N- and C-terminus halves of AtCSG2.24 AtCSG2 binds to poly(U), poly(A) and poly(G).23,24 Experiments using molecular beacons indicate that AtCSG1, but not AtCSG2, presents DNA melting activity, suggesting that AtCSG1 can act as a RNA chaperone.22 Contradicting results suggest that AtCSG2 also presents DNA melting activity.25
AtRBG7 was shown to preferentially bind G-rich RNA sequences and presented DNA melting activity, which implies that it can work as a RNA chaperone.22 More detailed experiments have shown that AtRBG7 binds to both ssDNA and ssRNA and that the minimal binding sequence 5′UUC UGG3′ is required for AtRBG7 recognition. AtRBG7 preferentially binds to ssRNA if it is fully stretched rather than to molecules presenting a stable secondary structure.26 AtRBG7 binds to its own pre-mRNA generating an alternativelly spliced form that is rapidly degraded. A mutation on a conserved arginine of the RRM motif of AtRBG7 prevents this autoregulation.27 AtRZL1a binds both ssDNA and dsDNA, although it presents a higher affinity for the former. Regarding homoribopolymers, it binds preferentially poly(U) and poly(G) presenting a very tight binding to the latter.28
A more detailed functional analysis of those proteins has been performed using mutants and transgenic plants. Plants carrying mutant alleles of AtRBG2 presented lower germination rate under salt stress compared to the control plants, while plants overexpressing AtRBG2 presented a higher germination rate. Interestingly, AtRBG2 presents transcription anti-termination activity, suggesting that it can act as a RNA chaperone.29
On the other hand, plants carrying mutant alleles of AtRZL1a30 or AtRBG7,31 present resistance to dehydration and high salinity in germination assays. Moreover, plants overexpressing AtRGB4,21 AtRGB7,31 AtRZL1a30 and AtCSG1,24 display increased sensitivity to osmotic stress. Surprisingly, an AtCSG1- related gene—AtCSG2—promotes increased resistance to osmotic stress when overexpressed,24 indicating that these two related proteins present opposite effects in plants under osmotic stress.
Osmotic stress has been shown to generate reactive oxygen species. It was verified that Atrzl1a mutants present lower levels of H2O2 when grown under high salt conditions, and the opposite phenotype was found for plants overexpressing AtRZL1a that also presented higher H2O2 levels under dehydration conditions. Indeed, AtRZL1a was shown to modulate the levels of several proteins involved in ROS homeostasis.30 Therefore, it would be interesting to test if the altered levels of H2O2 account for the difference in the germination rates of mutants and overexpression lines of AtRZL1a under osmotic stress conditions.
AtRBG7 is highly expressed in guard cells and it was demonstrated to regulate the stomata aperture in an ABA-independent manner. Atrbg7 mutants displayed smaller stomata aperture, while overexpression plants presented stomata with greater aperture under osmotic stress and high salt treatments when compared to wild-type plants.31 Therefore, both drought and high salt resistance of Atrbg7 mutants and sensitivity of overexpression lines could be explained by the control of water deficit through regulation of stomata aperture.
Class IVc RNA-binding GRPs are characterized by the presence of a cold shock domain. The homology with known cold shock proteins of prokaryotes suggests a role in plant cold acclimation. Indeed, cold shock domain containing GRPs display functions similar to those cold shock proteins of bacteria, such as RNA chaperone and transcription anti-termination activity.22,25,32 In monocots, Class IVc GRPs from wheat and rice were analyzed regarding their RNA binding properties. The wheat GRP TaCSG1 was shown to bind both ssDNA and dsDNA. A truncated version of the protein not containing the C-terminal zinc-fingers binds preferentially to ssDNA, implicating the zinc-fingers and the interspersed glycine-rich domain, at least partially, in the binding to dsDNA. Both versions of the protein were able to bind RNA, with a preference for poly(U) and poly(G) homoribopolymers.33 Point mutations in the RNA-binding sequences RNP1 and RNP2 present in the cold-shock domain of TaCSG1 indicate that RNP1 and RNP2 are necessary for ssDNA-binding. The transcription anti-termination activity and the nucleic acid melting activity of TaCSG1 suggest that it can function as a RNA chaperone in wheat.32 In rice, two TaCSG1-related GRPs—here, OsCSG1 and OsCSG2—are able to bind ssDNA.34
Heterologous expression of AtCSG1,22 OsCSG1,34 OsCSG2,34 or TaCSG1,32 complement the E. coli CSP mutant strain BX04 lack of growth in cold conditions. Although no BX04 complementation by AtCSG2 was observed by Kim and coworkers,22 weak complementation was demonstrated by another group.25 Experiments with truncated TaCSG1 proteins suggest that complementation of BX04 phenotype requires the presence of at least one zinc-finger domain.32
Interestingly, other RNA-binding GRPs that lack cold shock domain also play a role in plant cold tolerance. Heterologous expression of AtRBG2,29 AtRBG7,22 or AtRZL1a35 can complement the growth phenotype of E. coli strain BX04, which suggests that those proteins can also act as RNA chaperones. Indeed, a proteomic analysis was able to identify some potential targets that are differently regulated in plants overexpressing AtRBG2 compared to wild-type during cold stress.29 Proteome analysis using plants overexpressing AtRZL1a during cold stress revealed the modulation of several targets. Interestingly, the transcript levels of those targets were not affected in the transgenic plant, suggesting that AtRZL1a acts as an RNA chaperone during cold stress.35
The overexpression of AtRZL1a,28 AtRGB2,29 or AtRGB7,31 in Arabidopsis promotes freezing tolerance. Freezing response mediated by AtRBG7 may be conferred by stomata aperture regulation.31 Additionally, overexpression of AtRZL1a28 or AtRGB2,29 promotes germination and seedling growth under cold conditions. Moreover, corroborating these results, lesions on AtRZL1a28 or AtRGB2,29 loci confer opposite phenotypes as those obtained with the respective overexpressing lines. Puzzlingly, despite their opposite effects during osmotic stress in wild-type plants (see above), both AtCSG1 or AtCSG2 overexpression promotes freezing tolerance of Atrgb7 mutants,24 which may indicate functional redundancy between different subclasses of Class IV RNA-binding GRPs.
The involvement of GRP in plant development has been proposed based on the developmentally regulated expression pattern of some of those genes.1 However, the confirmation of this hypothesis has only recently been demonstrated through experimental data. The RNA-binding GRP gene AtCSG2 acts to repress flowering. AtCSG2 silenced transgenic plants show early flowering, reduced stamen number and abnormal embryo development.23
Two other RNA binding GRPs, from a different subclass have been also implicated in flowering time control. AtRBG7 and AtRBG8 function in the autonomous pathway to promote flowering transition. Both Atrgb7 mutants and transgenic plants with reduced AtRBG7 or AtRBG8 mRNA levels show upregulation of the flowering repressor FLC and late flowering phenotype. Similar to other autonomous pathway mutants, the late flowering phenotype is independent of day length conditions, sensitive to vernalization and mediated by FLC. In addition, overexpression of AtRBG7 causes early flowering and plants show reduced FLC mRNA levels.36
As observed for AtRBG7, AtRBG8 also binds to its own pre-mRNA causing an alternative spliced form that decays rapidly. Both genes are circadian-regulated and their auto-regulation is responsible for their oscillation. Interestingly, they also regulate each other. AtRBG7 promotes the generation of the alternative spliced form of AtRGB8 and vice versa, causing an interlocked negative feedback loop involving these two proteins.37
A GRP from Class I, without any characterized conserved motif besides a signal peptide, has been recently implicated in plant development. AtGRP5 was shown to be involved in cell elongation processes. The analysis of knockout mutant and transgenic plants overexpressing or reducing the expression of AtGRP5 presented altered root and inflorescence axis length and leaf size. Confocal analysis indicates the presence of AtGRP5-GFP fusion in the vacuole, suggesting that the phenotypes observed can be, at least, partially explained by perturbance of cell elongation processes. In fact, measurements of root cell length and hypocotyl elongation experiments support this hypothesis, implicating AtGRP5 in cell elongation.38 Additional analyses are necessary in order to characterize how AtGRP5 influences cell elongation.
The diversity in structure, expression pattern, modulation and subcellular localization of plant GRPs was a very strong indication that these proteins would perform very distinct functions in planta.1 Only now, when the first functional data of plant GRPs are available, this indication was confirmed. Some of the questions that remain unanswered are what is the function of the glycine-rich domain present in those proteins and how they are involved in their function.
For some proteins of specific classes, additional domains seem to be of more importance for function than the glycine-rich domain does. This is the case for the cysteine-rich domain of AtGRP3 responsible for interaction with AtWAK1,19 and the RNA-binding sequences of Class IV GRPs.21–28,32–34 On the other hand, Class I GRPs that present no additional characterized domain besides a signal peptide presented functional properties. PvGRP1.8 and AtGRP5 are composed of nothing but a glycine-rich domain suggesting that this domain is responsible for their function.9,38
One possibility is that some GRPs may act as part of a multi-component complex and that the glycine-rich domain is responsible for their interaction with their complex partners. It is also possible that, even in GRPs presenting other functional domains, the glycine-rich region is necessary for conformational purposes that affect protein activity. As additional GRP functions gets uncovered, it is also of great interest to understand how the glycine-rich region affects GRP function and if GRPs from different classes are evolutionary related or just happen to share a similar domain.
Previously published online: www.landesbioscience.com/journals/psb/article/10336