Caffeoyl Shikimate Esterase (CSE) Is an Enzyme in the Lignin Biosynthetic Pathway in Arabidopsis

See allHide authors and affiliations

Science  06 Sep 2013:
Vol. 341, Issue 6150, pp. 1103-1106
DOI: 10.1126/science.1241602

Lignin Biosynthesis Complications

Lignin is a polymer that lends its sturdy properties to wood and makes plant cell walls tougher, which creates problems for chemists converting cellulosic plant biomass into biofuels. Vanholme et al. (p. 1103, published online 15 August; see the cover) have identified a new step in the biosynthetic pathway of lignin in Arabidopsis in which caffeoyl shikimate esterase catalyzes synthesis of caffeate. Cellulose from mutant plants, which had reduced amounts of lignin, was more efficiently processed into glucose.


Lignin is a major component of plant secondary cell walls. Here we describe caffeoyl shikimate esterase (CSE) as an enzyme central to the lignin biosynthetic pathway. Arabidopsis thaliana cse mutants deposit less lignin than do wild-type plants, and the remaining lignin is enriched in p-hydroxyphenyl units. Phenolic metabolite profiling identified accumulation of the lignin pathway intermediate caffeoyl shikimate in cse mutants as compared to caffeoyl shikimate levels in the wild type, suggesting caffeoyl shikimate as a substrate for CSE. Accordingly, recombinant CSE hydrolyzed caffeoyl shikimate into caffeate. Associated with the changes in lignin, the conversion of cellulose to glucose in cse mutants increased up to fourfold as compared to that in the wild type upon saccharification without pretreatment. Collectively, these data necessitate the revision of currently accepted models of the lignin biosynthetic pathway.

The evolutionary emergence of lignin, a phenolic polymer deposited in the secondary cell wall, allowed the development of vascular land plants. The hydrophobic and strengthening nature of lignin enables conducting xylem vessels to transport water and nutrients from the roots to photosynthetic organs while withstanding the negative pressure caused by transpiration (1, 2). The strengthening of fiber cells by lignification allows vascular plants to grow tall and stand upright (1, 3, 4). However, these very same physicochemical properties of lignin are a barrier to the isolation of cellulose fibers by chemical pulping and the enzymatic hydrolysis of cell wall polysaccharides in biorefining. Biomass feedstocks with less lignin or with more-degradable lignin would reduce the high processing costs and carbon footprint of paper, biofuels, and chemicals (5).

The lignin biosynthetic pathway has been extensively studied (610). In dicotyledonous plants, lignin is mainly synthesized from two monomers or monolignols, coniferyl alcohol and sinapyl alcohol (6, 11, 12), that upon incorporation into lignin give rise to guaiacyl (G) and syringyl (S) units, respectively. p-Coumaryl alcohol gives rise to the less abundant p-hydroxyphenyl (H) lignin units. 3-Hydroxylation of the aromatic ring, catalyzed by p-coumarate 3-hydroxylase (C3H, Fig. 1), diverts flux away from H lignin and toward G and S lignin. The discovery that C3H accepts p-coumaroyl shikimate and quinate esters as substrates (1315) led to the identification of hydroxycinnamoyl-CoA:shikimate/quinate hydroxycinnamoyltransferase (HCT) as the enzyme catalyzing the preceding step, the production of p-coumarate esters from p-coumaroyl–coenzyme A (CoA) (16) (Fig. 1). The suggestion that HCT also catalyzes a second reaction in the lignin pathway, converting the resulting caffeate esters into caffeoyl-CoA (16), was attractive because it brought the interpreted pathway back to the next expected intermediate, caffeoyl-CoA. Here we describe caffeoyl shikimate esterase (CSE, encoded by At1g52760) as an enzyme in the lignin biosynthetic pathway that, together with 4-coumarate:CoA ligase (4CL), bypasses the second HCT reaction.

Fig. 1 The lignin biosynthetic pathway incorporating the CSE-dependent reaction established in this study.

CSE was first identified through a potential function in phospholipid repair upon oxidative stress (17). However, we identified CSE as a candidate for involvement in lignification, based on analyses designed to identify genes coexpressed with known components of the lignin biosynthetic pathway. Of 13 genes identified by each of three different coexpression software tools and publicly available data sets, 9 were established lignin pathway genes, but three, including CSE, had no known role in lignin biosynthesis (fig. S1 and table S1). Of these three, only CSE was also identified as being coexpressed with lignin pathway genes in a set of lignin mutants (7). Consistent with the coexpression analysis, CSE-reporter fusion proteins expressed from transgenes (composed of the native CSE promoter, exons, and introns, fused to a reporter gene) were detected in lignifying vascular tissue of primary transformants (fig. S2).

To investigate a role for CSE in lignification, we studied two transfer DNA insertion mutants: cse-1, a knockdown mutant with an insertion in the promoter, and cse-2, a knockout mutant with an insertion in the second exon (fig. S3). Although cse-1 did not show developmental abnormalities, the inflorescence stems of cse-2 mutants were 37% smaller and 42% lighter at senescence than were those of the wild type (Fig. 2 and fig. S4). CSE transcript levels analyzed by quantitative reverse transcription polymerase chain reaction were 6.3% of that in the wild type in cse-1 and undetectable in cse-2, the weak and strong mutant alleles, respectively (fig. S3).

Fig. 2 Phenotype and lignin characteristics of cse mutants.

(A) Fully grown plants after cultivation for 8 weeks in short day photoperiods and for 5 weeks in long day photoperiods. Height and weight measurements are in fig. S5. (B) Transverse stem sections of cse-2 mutants and wild-type plants. Wiesner staining and additional images of cse-1, cse-2 CSE 1, and cse-2 CSE 2 sections are in fig. S5. Collapsed vessel elements are indicated by arrowheads. Scale bar, 100 μm. (C) Lignin levels determined by the acetyl bromide method. See table S2 for lignin data of cse-1. ***0.001 > P; unpaired two-sided t test (D) Relative H:G:S lignin composition as determined by thioacidolysis (left) and from whole–cell wall NMR integrals (right). See table S2 for full thioacidolysis details. (E) Partial short-range 13C-1H [heteronuclear single-quantum coherence (HSQC)] spectra (aromatic region) of the cell walls of cse-2 mutants and wild-type plants. For the side-chain region of the HSQC spectra, see fig. S6. Error bars indicate ± SEM.

Analysis of transverse sections of cse-1 and cse-2 mutant stems revealed reduced autofluorescence and less intense Wiesner and Mäule staining in vessels and fibers of cse-2, indicative of reduced lignin content and fewer lignin S units (Fig. 2 and fig. S5). In addition, cse-2 mutants had collapsed vessel elements (Fig. 2), a phenotype typical of plants with weakened secondary cell walls (18). The mutant phenotype of cse-2 was complemented in stable transgenic lines in which expression of CSE was driven by the cauliflower mosaic virus 35S promoter (Fig. 2 and figs. S4 and S5), verifying that the mutation of CSE is the cause of the observed phenotypes.

In order to examine the connection between CSE and lignification in more detail, senesced inflorescence stems of cse-1 and cse-2 were subjected to compositional analyses. The fraction of the dry weight made up by cell wall polymers after soluble molecules had been extracted (i.e., the cell wall residue) was significantly reduced from 79.8% in the wild type to 72.9% in the cse-2 mutant (0.05 > P > 0.01) (table S2). The fraction of acetyl bromide–soluble lignin released from this cell wall residue was reduced by 17 and 36% in the cse-1 and cse-2 mutants as compared to their corresponding wild-type controls (Fig. 2 and table S2). The lignin composition of cse-2, determined by thioacidolysis and nuclear magnetic resonance (NMR), revealed that the relative proportion of H units increased over 30-fold (Fig. 2, fig. S6 and table S2). Milder compositional shifts were apparent in the lignin of cse-1 (table S2). The increase in the proportion of H units in both mutants suggests that CSE is active in the general phenylpropanoid pathway after the branch leading to H units but before the pathways for G and S units diverge. At this part of the pathway, HCT, C3H, and caffeoyl-CoA O-methyltransferase (CCoAOMT) are also active (Fig. 1) and, accordingly, plants with reduced HCT and C3H activity also have lignin enriched in H units (14, 19, 20).

Because lignin composition was altered in both cse mutant alleles, we expected to find a shift in phenolic metabolism. We analyzed methanol-soluble phenolics of stem extracts of both cse mutant alleles by liquid chromatography mass spectrometry (LC-MS) (5). The abundance of two compounds, both oligolignols containing G and S units, was reduced in each mutant allele (compounds 28 and 29; figs. S7 and S8 and table S3). These findings are consistent with the reduced biosynthesis of lignin in the cse mutants, because the abundance of oligolignols during Arabidopsis stem development is correlated with the amount of lignin (7). Of the 27 compounds with increased abundance in the cse mutants as compared to the wild type, 21 could be identified. Nineteen were caffeate- and ferulate-derived products (figs. S7 and S8), of which caffeoyl shikimate was most abundant (Fig. 3). In addition, (hexosylated) H unit–containing oligolignols accumulated in the cse mutants, as well as a neolignan containing an H unit and ferulate (figs. S7 and S8). These data also support the hypothesis that CSE functions after the branch in the lignin pathway where G and S unit biosynthesis diverges from that of H units.

Fig. 3 Caffeoyl shikimate accumulates in cse mutants and is hydrolyzed by CSE.

(A) Representative LC-MS chromatograms of wild type, cse-1, and cse-2 plants, showing the increased peak of caffeoyl shikimate in the cse mutants. See fig. S7 and table S3 for full details. m/z, mass-to-charge ratio. (B) Enzyme kinetics curve measured with the recombinant CSE, showing that CSE accepts caffeoyl shikimate as a substrate.

Some of the compounds that accumulate in cse mutants relative to the wild type might be substrates for CSE. In order to test this possibility, we incubated purified recombinant CSE enzyme with extracts of cse-1 mutants, after the complexity of the extract had been reduced by chromatographic separation into 96 fractions. Three compounds decreased in abundance upon treatment with CSE, and new compounds appeared in the same fractions (fig. S9 and tables S4 and S5). Caffeoyl shikimate, which was the compound that showed greatest abnormal accumulation in cse mutants, was almost completely hydrolyzed (97%) by recombinant CSE into caffeic acid (fig. S9). Modeling the structure of CSE revealed that caffeoyl shikimate could fit into the active site (fig. S10). Correspondingly, crude extracts from cse-2 lignifying tissues were less able to hydrolyze caffeoyl shikimate into caffeate than those of the wild type (fig. S11). We therefore suggest that caffeoyl shikimate is a substrate for CSE in vivo. Because caffeoyl shikimate is an accepted intermediate in lignin biosynthesis (Fig. 1) (13, 15, 21), this places CSE in the lignin biosynthetic pathway. The Michaelis-Menten constant (Km) and maximum reaction velocity (Vmax) values of CSE for caffeoyl shikimate were 96.5 μM and 9.3 picokatals (pkat) per μg of protein, respectively (Fig. 3). We also tested p-coumaroyl shikimate, which is structurally similar to caffeoyl shikimate and an intermediate in the phenylpropanoid pathway, as a potential substrate of CSE (13, 15). The Km and Vmax values of CSE for p-coumaroyl shikimate were 211 μM and 0.66 pkat per μg of protein, respectively. Thus, CSE showed a higher affinity for caffeoyl shikimate, with a Vmax/Km value that is 31 times greater than that for p-coumaroyl shikimate, suggesting that caffeoyl shikimate is the preferred CSE substrate.

Current lignin pathway models indicate that caffeoyl shikimate is converted to caffeoyl-CoA (Fig. 1) (1, 6, 16). When we tested whether this reaction could also be catalyzed by CSE with caffeoyl shikimate and CoA as substrates, only caffeate, but not caffeoyl-CoA, was produced (fig. S12). Our data suggest that current lignin biosynthetic pathway models should be revised to include the CSE-catalyzed conversion of caffeoyl shikimate into caffeate, although we cannot exclude the possibility that CSE can convert caffeoyl shikimate into other phenolic compounds in vivo.

Because the Arabidopsis 4CLs involved in lignification (4CL1 and 4CL2) have Km and Vmax values of the same order of magnitude for both p-coumarate and caffeate (7, 22, 23), caffeate might be used by 4CL to form caffeoyl-CoA in vivo (Fig. 1). Current pathway models indicate direct conversion by HCT of caffeoyl shikimate and CoA into caffeoyl-CoA (1, 6, 16). We confirmed that purified recombinant Arabidopsis HCT enzyme could indeed catalyze this reaction (fig. S13). These data show that caffeoyl shikimate may be a substrate for both CSE and HCT in vivo. A HCT-dependent route from caffeoyl shikimate to caffeoyl-CoA (Fig. 1) may explain how residual lignin in the cse-2 null mutant is synthesized. In addition, the C3H/C4H (cinnamate 4-hydroxylase) heteromeric complex may contribute to carbon flux toward lignin (Fig. 1), because homologs of these enzymes from poplar have been shown to convert p-coumarate to caffeate (24). However, these alternative routes to lignin biosynthesis do not fully compensate for a loss of CSE activity, because cse mutants are compromised in lignification and development. Likewise, the accumulation of caffeoyl shikimate that occurs in cse mutants suggests that HCT is relatively ineffective at metabolizing this substrate in vivo.

Lignin limits the processing of plant biomass to fermentable sugars (25, 26). Processing of cse mutant plants, which have reduced lignin content, might yield more sugars on saccharification. We compared cellulose-to-glucose conversion of senesced stems from both cse mutants and wild-type plants. Cell wall residues of senesced inflorescence stems of cse-1 have normal amounts of cellulose, whereas those of cse-2 have 73% of the normal amount of cellulose (table S2). The cellulose-to-glucose conversion of the unpretreated cell wall residue was monitored over a period of 48 hours (Fig. 4); when the plateau was reached, the conversion had increased from ~18% in the wild type to ~24% in cse-1 (i.e., a relative increase of 32%) and to ~78% (fourfold higher than in the wild type) in cse-2. Therefore, saccharification efficiency increases as lignin content decreases. On a plant basis, cse-2 mutants released 75% more glucose than the wild type. Saccharification efficiency from material derived from cse-2 plants is similar to that of ccr1-3, a mutant in the lignin pathway gene for cinnamoyl-CoA reductase that has the highest saccharification efficiency described so far (26).

Fig. 4 Cellulose-to-glucose conversion during saccharification of the senesced inflorescence stems of cse mutants.

h, hours. Error bars indicate ± SEM. *0.05 > P > 0.01, **0.01 > P > 0.001, ***0.001 > P; unpaired two-sided t test.

We found orthologs of CSE in a wide range of plant species (fig. S14), including biofuel feedstocks such as poplar, eucalyptus, and switchgrass. Consistent with a potential conserved role in lignification, CSE copurifies with lignin biosynthetic enzymes in extracts from poplar xylem (27). The characterization of CSE in other species will reveal how widely the revision of the lignin biosynthetic pathway we propose here applies and whether CSE could be a generally useful target for reducing cell wall recalcitrance to digestion or industrial processing in biomass crops.

Supplementary Materials

Materials and Methods

Figs. S1 to S14

Table S1 to S5

References (2855)

References and Notes

  1. Acknowledgments: The authors thank A. Bleys for help in preparing the manuscript and K. Graham for technical support. We gratefully acknowledge funding through the European Commission’s Directorate-General for Research within the 7th Framework Program (FP7/2007-2013) under grant agreements 211982 (RENEWALL) and 270089 (MULTIBIOPRO); Stanford University’s Global Climate and Energy Project (Towards New Degradable Lignin Types, Novel Mutants Optimized for Lignin, Growth and Biofuel Production via Re-Mutagenesis, and Efficient Biomass Conversion: Delineating the Best Lignin Monomer-Substitutes); the Hercules program of Ghent University for the Synapt Q-Tof (grant AUGE/014); the Bijzonder Onderzoeksfonds-Zware Apparatuur of Ghent University for the Fourier transform ion cyclotron resonance mass spectrometer (174PZA05); and the Multidisciplinary Research Partnership Biotechnology for a Sustainable Economy (01MRB510W) of Ghent University. R.V. is indebted to the Research Foundation-Flanders for a postdoctoral fellowship and L.S., I.C., and P.A. to the Agency for Innovation by Science and Technology (IWT), CAPES-Brazil (grant 201660/2010-5), and the CNPq-Brazil sandwich Ph.D. (grant 201998/2011-4), respectively, for predoctoral fellowships. J.R. and H.K. were funded in part by the U.S. Department of Energy’s Great Lakes Bioenergy Research Center (DOE Office of Science BER DE-FC02-07ER64494). G.G.S. was partially funded by the Scottish Government. W.B. is on the Science Advisory Board of the NSF-funded project Regulation and Modeling of Lignin Biosynthesis, DBI-0922391. A patent application, “Modified plants” PCT/GB2013/051206, on the modification of CSE expression to improve processes that require carbohydrate extraction, has been filed jointly by the University of Dundee and the Flanders Institute for Biotechnology (VIB).
View Abstract

Stay Connected to Science

Navigate This Article