Mad moves of the building blocks – nucleotide sugars find unexpected paths into cell walls

This article comments on: Zhao X, Liu N, Shang N, et al. 2018. Three UDP-xylose transporters (UXTs) participate in xylan biosynthesis by conveying cytosolic UDP-xylose into the Golgi lumen in Arabidopsis. Journal of Experimental Botany 69, 1125–1134..

A raft of recent studies, including a new paper by Zhao et al. (2018), has shed new light on the importance of cellular topology in controlling the flux of nucleotide sugars to their destinations. It is an exciting time, as the rapidly progressing elucidation of the convoluted flux of carbon bodies through nucleotide sugar metabolism and inter-organellar transport means we are approaching an understanding of the control principles of plant cell wall polysaccharide biosynthesis.
Plant science is approaching the vision of cell walls constructed 'fit for purpose' (Johnson et al., 2017). The assembly of carbohydrates from monosaccharides encompasses the stereospecific transfer of specific (di)phosphonucleotide-activated sugars onto specific acceptor molecules by glycosyltransferases estimated to be encoded by well over 500 Arabidopsis loci (Hansen et al., 2012). At least 16 nucleotide sugars are generated in the nucleotide sugar interconversion network and many of the enzymes involved have now been cloned and characterized (Reiter and Vanzin, 2001;Seifert, 2004;Reiter, 2008;Bar-Peled and O'Neill, 2011). More recently, Golgi-localized nucleotide sugar transporters have been identified (Temple et al., 2016). But how does nucleotide sugar metabolism and transport align with carbohydrate biosynthesis in the living plant?

Regulation of substrate flux
Most nucleotide sugars are made from UDP-D-glucose (UDP-Glc) in the cytosol (Box 1). This central carbohydrate precursor is converted to UDP-L-rhamnose, UDP-galactose and UDP-D-glucuronic acid (UDP-GlcA), the latter being converted to UDP-D-apiose and UDP-D-xylose (UDP-Xyl). Only the enzyme that generates UDP-D-galacturonic acid (GalA) from UDP-GlcA is exclusively Golgi localized. By contrast, two UDP-sugar interconverting enzymes exist in two forms localized in topologically separated compartmentsone form in the cytosol and another one in the Golgi. The reason for this plant-specific duality is not immediately apparent.
The first example is the biosynthesis of UDP-Xyl, a precursor for xylans (York and O'Neill, 2008), xyloglucan, rhamnogalacturonan II (RG II) and xylogalacturonan (Caffall and Mohnen, 2009). UDP-Xyl synthase (UXS) catalyses the decarboxylation of UDP-GlcA to UDP-Xyl with the UXS3, -5 and -6 genes encoding cytosolic UXS and the UXS1, -2 and -4 loci encoding the Golgi-localized isoforms. Because UDP-Xyl is exclusively required inside the Golgi, the three cytosolic UXS isoforms appear redundant. More crucially, to give the cytosolic pool of UDP-Xyl access to the site of carbohydrate biosynthesis, a Golgi-localized UDP-Xyl transporter is needed. Indeed, three UDP-Xyl transporters (UXT1-3) were ingeniously identified using a biochemical assay and the uxt1 mutant showed a defect in glucuronoxylan (GX) structure and abundance. This phenotype suggested that polysaccharide biosynthesis somehow depended on cytosolic UDP-Xyl despite the presence of functional UXS in the Golgi. However, there was only a relatively minor reduction of total cell wall xylose in uxt1 and no carbohydrate other than GX was affected in the mutant. Moreover, the uxt2 uxt3 double mutant was phenotypically normal (Ebert et al., 2015).
This key paper by Ebert et al. (2015) opened important questions for understanding how substrate flux is regulated in planta. Does the GX-specific defect in uxt1 reflect substrate channelling from UXT1 to GX-specific xylosyltransferases? And vice versa, is the relatively subtle overall phenotype of uxt1 due to genetic redundancy in UDP-xylose transport or the action of Golgilocalized UXS? Zhao et al. (2018) address these questions, showing that in the uxt1 uxt2 uxt3 triple mutant, plant growth and secondary cell wall thickening are severely affected and, in addition to GAX, xyloglucan is also defective. Although not contradicting the substrate channelling hypothesis their paper conclusively demonstrates the crucial importance of the cytosol to Golgi transport of UDP-Xyl for multiple xylose-containing carbohydrates. This further corroborates previous triple-mutant studies comparing the two forms of UXS, which showed that cytosolic UXS, but not the Golgi-localized enzyme, affected the incorporation of xylose into cell wall polysaccharides (Kuang et al., 2016;Zhong et al., 2017).

Mechanistic basis of separation
What might be the mechanistic basis for this apparent separation of metabolite fluxes (Box 1)? A common feature of both UXS and UDP-Glc dehydrogenase (UGD), which generates UDP-GlcA, is that both enzymes are inhibited by UDP-Xyl (Harper and Bar-Peled, 2002;Pattathil et al., 2005;Klinghammer and Tenhaken, 2007). Therefore, developmental up-regulation of cytosolic UXS at the onset of secondary cell wall formation would increase cytosolic UDP-Xyl, thereby feedback-inhibiting UGD. The inhibition of UGD would down-regulate the biosynthesis of both UDP-GalA and, via Golgi-localized UDP-Xyl biosynthesis, UDP-Larabinose (UDP-Ara), which are mainly needed during primary cell wall biosynthesis. But if xylose-containing carbohydrates receive UDP-Xyl from cytosolic UXS, what is the biological function of the Golgi form? Two genetic studies failed to establish a clear role for UXS1, -2 and -4 (Kuang et al., 2016;Zhong et al., 2017). However, as the loci are primarily expressed in tissues that undergo primary cell wall formation (Kuang et al., 2016), it would be interesting to further investigate the uxs1 uxs2 uxs4 carbohydrate-chemical phenotype in actively expanding tissues.
Indirect evidence for the importance of Golgi-localized UXS in providing the substrate for UXE and ultimately for arabinosyltransferases came from the finding that a transporter for UDP-GlcA, called UUAT1, was important for the arabinose content of cell walls (Saez-Aguayo et al., 2017). Once inside the Golgi, UDP-GlcA is either converted to UDP-GalA by UDP-glucuronic acid 4-epimerase (GAE) or to UDP-Xyl by UXS. From UDP-Xyl, UDP-Xyl 4-epimerase (UXE) generates UDP-Ara and, crucially, the uxe1 mutant showed a dramatic reduction in cell wall arabinose content (Burget and Reiter, 1999;Burget et al., 2003).
polymers. However, a mutase is required for efficient conversion of the product of UXE UDP-Arap into UDP-Araf. This mutase is a protein complex encoded by the REVERSIBLY GLYCOSYLATED POLYPEPTIDE1, loci and is partially bound to the cytosolic face of the Golgi. Knockdown of RGP1 and RGP2 resulted in massive growth defects and a near-complete loss of cell wall arabinose (Rautengarten et al., 2011). The recently identified UDP-arabinofuranose transporter UAfT  channels UDP-Araf back into the Golgi where it is the substrate for multiple arabinosyltransferases. However, the potential metabolite pathways into cell wall-linked Araf do not end here.
UDP-Arap exported from the Golgi to the cytosol is expected to mix with UDP-Arap generated from free L-arabinose during carbohydrate breakdown using arabinokinase and UDP-sugar pyrophosphorylase (Geserick and Tenhaken, 2013;Behmüller et al., 2016). This salvage pathway is the mechanistic explanation for the reversion of the uxe1 phenotype by experimentally applied L-Ara (Burget and Reiter, 1999). And yet another pathway potentially leads to UDP-Araf. The cytosolic bifunctional UXE/UDP-glucose 4-epimerase (the UGE1 and UGE3 isoforms) (Kotake et al., 2009) should in fact generate cytosolic UDP-Arap from the plentiful supply of UDP-Xyl by cytosolic UXS. This potential option, however, is contradicted by the strong arabinose deficiency in cell walls in the uxe1 mutant (Burget and Reiter, 1999) and the absence of a phenotype in the uge1 uge3 double mutant (Rösti et al., 2007). One possible explanation might be that precise co-expression at the level of both mRNA and protein is required for normal metabolite flux during growth and development. Another possibility to explain the observed specificity is substrate channelling between adjacent enzymes and transporters assembled in a complex. At any rate, while the metabolic pathways of nucleotide sugars appear capricious, they undoubtedly represent a flexible and efficient delivery system hardened by billions of years of evolution.