Straying off the Highway: Trafficking of Secreted Plant Proteins and Complexity in the Plant Cell Wall Proteome

From simply skimming through the abstract lists of more or less any collection of both basic and applied plant-related journals, it is immediately apparent that the plant cell wall represents a nexus of many fields of research: growth and development, plant-pathogen interactions, abiotic stress, self- and interorganismal recognition, signaling systems, numerous primary and specialized metabolic processes, biomaterials and bioproducts, and many others. That said, and to deal with an issue of semantics, while the term cell wall can refer specifically to the structural matrix that surrounds all plant cells, for the purposes of this Update it is used more broadly, also to include the apoplast, or extracellular environment. Given its multifunctional nature then, it is not surprising that the apoplast houses a dynamic and complex proteome, and the compendium of cell wall proteins continues to grow as researchers from disparate disciplines discover new roles for extracellular proteins. In addition, however, as cell wall proteomics projects develop and the subcellular localizations of an ever-growing list of plant proteins are determined, a number of surprises have been thrown up, both in terms of the identity of secreted proteins and the trafficking pathways that they follow. The purpose of this Update is to give some examples of previously unsuspected aspects of plant cell wall protein trafficking that are challenging long-held assumptions, rather than to provide an exhaustive review, and to highlight some questions that can be categorized into the “who, how, where, and when” of the cell wall proteome. Developing a comprehensive catalog of the cell wall proteome is generally far more challenging than for most intracellular organelles, which can be isolated in highly purified fractions, relatively free from nonspecific protein contamination. Cell wall proteins are not only spread throughout the apoplastic milieu, but also show a wide range of affinities for the extracellular matrix itself, from highly mobile with no apparent interaction, to covalently bound. In addition, when cell walls extracts are prepared, by tissue or cell homogenization followed by centrifugation, substantial amounts of intracellular protein inevitably associate with the wall pellet, while proteins and peptides that were not bound to the wall in vivo are lost from the extract. There are certainly other technical challenges, such as the fact that most secreted proteins are glycosylated, which complicates separation and identification, but the major confounding factors are still contamination and incomplete capture. Analyses of protein populations from the walls of many plant organs and tissues have been reported (for review, see Lee et al., 2004; Jamet et al., 2006, 2008a), together with detailed methodologies to optimize extraction (Feiz et al., 2006; Watson and Sumner, 2007; Jamet et al., 2008b). These have resulted in catalogs of proteins whose identity matches known functions associated with wall-related processes, such as polysaccharide modification, defense, and signaling, as well as many with unknown functions. However, a subset is consistently detected whose localization in the wall is surprising, based on annotated or even experimentally established intracellular localization. These are often simply dismissed as contamination, and for many this is certainly the case; however, lists of secreted proteins from extracellular fluids that were collected under supposedly nondestructive conditions that would avoid major cell lysis, or from suspension cell media, also often include proteins that have an established intracellular role (Isaacson and Rose, 2006). While contamination can be a major contributing factor there are other explanations that should be considered, at least in some cases. First, there are a growing number of examples of known intracellular molecules that are also secreted to, or synthesized in, the apoplast under certain conditions, or in specific tissues. These include extracellular ATP, which likely plays a signaling role and is required for maintaining cell viability (Chivasa et al., 2005; Clark and Roux, 2009), and polymeric DNA that enhances resistances to fungal infection when secreted at the root tip (Wen et al., 2009). Similarly, a recent article described an apoplastic extracellular γ-glutamyl transferase, suggesting the existence of an extracellular system to salvage glutathione derived from the plant or external sources (Ferretti et al., 2008). Prior to such reports, the identification of enzymes associated with the biosynthesis or metabolism of these compounds in cell wall protein extracts would likely have been explained as contamination, but clearly now their presence could be seen in a new light, and the same will apply as other extracellular metabolites/substrates are detected. Other explanations for the presence of proteins or peptides in the apoplast with predicted intracellular functions are that they simply have high homology to known intracellular proteins but distinct functions, or that they have more than one function and are localized in multiple cellular compartments. Given that the cell wall represents the interface with the environment, there are undoubtedly intense selective pressures for existing extracellular proteins to evolve new functions, or to recruit new proteins to the apoplast through gene duplication and retargeting. Indeed, there are already many known examples of secreted proteins or peptides with more than one function (e.g. Kuwabara and Imai, 2009; Stotz et al., 2009), almost all of which have some relationship to abiotic or biotic stress. As an aside, it should be noted that some proteins with apparent multiple functions, based on the presence of both a detectable biological activity and sequence homology to a known enzyme, may not necessarily be truly dual function, or moonlighting, proteins. An example is the family of glucanase inhibitor proteins that are secreted by species of the oomycete Phytophthora, which appear to be derived from an ancestral Ser protease, and yet have lost the catalytic triad (Rose et al., 2002). These proteins bind and inhibit the activity of secreted plant endo-β-1,3-glucanases and show evidence of evolutionary selection on specific residues from both protein partners at the docking interface (Damasceno et al., 2008). Such studies hint at the adoption of new roles in response to evolutionary pressure, in this case on the part of the pathogen-secreted protein, but the principal applies equally to plant wall proteins. Another category of unexpected cell wall-localized proteins has been suggested by the identification of secreted proteins through individual protein localization studies, or functional screens (Zhu et al., 1994; Lee et al., 2004, 2006; Kim et al., 2008; Cheng et al., 2009), which do not apparently have targeting sequences that would direct to them to the classical secretory pathway. In the next section we discuss the implications for such observations for compiling a more complete compendium of the cell wall proteome. The archetypal pathway for secretion of a eukaryotic protein (Fig. 1) involves an N-terminal region, termed the signal peptide or leader sequence, which directs the nascent protein to the endoplasmic reticulum (ER), whereupon the remainder of the protein is cotranslationally translocated through the Sec61 complex into the ER lumen (Rapoport, 2007). At this point the protein can also undergo N-linked glycosylation. It then passes through the endomembrane system, or secretory pathway, comprising the Golgi apparatus and trans-Golgi network, where it is packaged into vesicles that migrate to, and fuse with, the plasma membrane, releasing the protein cargo into the cell wall. Many proteins are also retained in the ER or Golgi, which can include phases of retrograde trafficking, or are targeted to the vacuole or other post-Golgi compartments, as described in an excellent review by Foresti and Denecke (2008). Schematic diagram of the plant classical secretory pathway, highlighting protein cotranslational translocation from ER-bound ribosomes in the ER lumen and the subsequent major trafficking pathways. Soluble and membrane-anchored proteins are shown in red and blue, respectively. N, Nucleus; Chl, chloroplast; TGN, trans-Golgi network; PVC, prevacuolar compartment; PM, plasma membrane; CW, cell wall, Vac, vacuole. Adapted from Foresti and Denecke (2008). Several features of the secretory system appear to be distinctly different in plants and a number of aspects are still not well understood, such as the identity of the transport mechanisms that might convey secreted proteins from the Golgi, trans-Golgi network, or prevacuolar compartment to the plasma membrane, whether there exist post-Golgi compartments, or if there are multiple targeting pathways, as has been suggested in yeast (Saccharomyces cerevisiae; Foresti and Denecke, 2008; Hwang and Robinson, 2009; Richter et al., 2009). Additionally, there is evidence that some proteins can also traffic to the chloroplast from the Golgi (Villarejo et al., 2005; Kitajima et al., 2009) or ER (J.K.C. Rose and S.-J. Lee, unpublished data). The plastid-targeting machinery and protein structural elements that direct such targeting are not known, although multiple surface regions of the mature proteins appear to be important (Kitajima et al., 2009), nor is it known what regulates whether the proteins are directed to the plastid or to the wall. These observations suggest a previously unsuspected complexity in interorganellar targeting and provide further indications of multifunctionality of secreted proteins. Another major enigma is the growing evidence that plant proteins can be secreted to the apoplast via routes that are independent of the ER-Golgi pathway. Such proteins do not have canonical signal peptides and treatment with chemical inhibitors, such as brefeldin A, which disrupt vesicle trafficking between the ER and Golgi, does not perturb secretion. Nonclassical or leaderless secretion was first described 20 years ago for two mammalian proteins (Cooper and Barondes, 1990; Rubartelli et al., 1990) and since then considerable progress has been made in characterizing a number of different pathways (Fig. 2) in mammalian and yeast cells, which will not be described in detail here (for review, see Nombela et al., 2006; Nickel and Seedorf, 2008; Prudovsky et al., 2008; Nickel and Rabouille, 2009). It therefore seems likely that nonclassical secretion is common to all eukaryotes, including plants, although this field is essentially unexplored. Several algorithms have been developed to assist with computational prediction of nonclassically secreted proteins (e.g. Bendtsen et al., 2004); however, these typically use methods that involve identifying structural features that are conserved or frequently observed in secreted mammalian proteins, rather than identifying sequence motifs. Given the radical differences in the plant and mammalian extracellular environments, it is likely that the respective protein populations also collectively exhibit very different structural and compositional properties, which would limit predictive value of such computational tools with plant proteins. Indeed, in our hands, candidate proteins that we have identified through functional screens as being secreted by nonclassical pathways are not predicted to do so using existing algorithms (data not shown). Schematic diagram of a mammalian cell showing the major classes of nonclassical secretory pathways that have been identified or proposed. These pathways are resistant to the inhibitory effects of brefeldin A, which disrupts ER-Golgi vesicular traffic. N, Nucleus; BFA, brefeldin A; MVB, multivesicular bodies; PM, plasma membrane. Adapted from Nickel and Rabouille (2009). It is tempting to ask what sorts of plant proteins are likely to be secreted by unconventional routes. Some hints may come from studies of mammalian proteins, which have been classified into two classes: those that are mostly secreted and those that usually have an intracellular function but that are secreted following specific signals and in specific tissues (Nickel and Seedorf, 2008). It has also been suggested that such secretion can be rapid and activated by external stresses (Keller et al., 2008). It is possible that speed of secretion might provide a selective advantage, although other explanations for alternative secretory routes are the possibility that protein folding in the environment of the ER and posttranslational modification may not be desirable for some proteins. Moreover, it may be important in some cases to separate some enzymes from their substrates, or inhibitory and degradative factors, if they reside in the same secretory compartments. The super highway for plant cell wall-localized proteins is certainly the classical ER-Golgi route, and only a small subset is likely to be secreted by nonclassical pathways; however, this remains an exciting and entirely uncharacterized area. In addition to the questions surrounding the composition of the cell wall proteome and the nature of the associated trafficking pathways, two other important and poorly understood issues are the timing and spatial regulation of protein secretion. A recent review (ZÁrský et al., 2009) called into question the preconception that the plant and yeast secretory pathways are constitutive, while that of animal cells is highly regulated. There is indeed growing evidence of complex and highly coordinated spatiotemporal protein secretion in plants. Several cell wall proteins show evidence of retention in the Golgi or other compartments of the secretory pathway, prior to targeting to the apoplast by some currently undefined signal, although the mechanism of release may involve posttranslational proteolytic processing of a proregion of the protein (e.g. Dal Degan et al., 2001; Wolf et al., 2009). Interestingly, both these particular proteins act to degrade pectic polysaccharides and so it may be that the accumulation of inactive preproteins prior to the secretion of the processed, and thus enzymatically activated, mature polypeptides prevents precocious attack on the cell wall polymers. Polarized or spatially regulated secretion also provides an additional layer of complexity to the overall picture. It is known that plasma membrane-localized proteins, such as the PIN family of auxin efflux carriers, can show clear polar trafficking to specific cell faces (Richter et al., 2009; ZÁrský et al., 2009), and another remarkable example is the spatially coordinated targeting of secretory vesicles to specific regions of the plasma membrane adjacent to the sites of infection during microbial challenge (Hückelhoven, 2007). These are particularly well-studied examples, but it may well be that targeting of secretory vesicles to plasma membrane domains is more common than is currently thought. The purpose of this Update is to highlight just some of the burning questions surrounding the plant cell wall proteome, focusing particularly on the challenges of determining which proteins truly reside in the wall and the pathways by which they arrive. There are certainly many other fascinating, and generally neglected, aspects of cell wall protein research, including the potential for phosphorylation, the identification of extracellular protein complexes, and postsecretion proteolytic and glycanolytic processing, but these are beyond the scope of this review. We would like to acknowledge the many researchers who have made important contributions to the field, but whose work we unfortunately cannot present in this Update due to length limitations.