Report

Abnormal Lignin in a Loblolly Pine Mutant

See allHide authors and affiliations

Science  11 Jul 1997:
Vol. 277, Issue 5323, pp. 235-239
DOI: 10.1126/science.277.5323.235

Abstract

Novel lignin is formed in a mutant loblolly pine (Pinus taeda L.) severely depleted in cinnamyl alcohol dehydrogenase (E.C. 1.1.1.195), which converts coniferaldehyde to coniferyl alcohol, the primary lignin precursor in pines. Dihydroconiferyl alcohol, a monomer not normally associated with the lignin biosynthetic pathway, is the major component of the mutant's lignin, accounting for ∼30 percent (versus ∼3 percent in normal pine) of the units. The level of aldehydes, including new 2-methoxybenzaldehydes, is also increased. The mutant pines grew normally indicating that, even within a species, extensive variations in lignin composition need not disrupt the essential functions of lignin.

Lignins are complex phenolic plant polymers essential for mechanical support, defense, and water transport in vascular terrestrial plants (1, 2). They are usually derived from three hydroxycinnamyl alcohol precursors2a through c in varying proportions (Fig.1). In gymnosperms—for example, pine and other conifers—lignin is polymerized from only two of the three monomers, p-coumaryl alcohol 2a and coniferyl alcohol 2b, with coniferyl alcohol being predominant (∼90%). p-Coumaryl alcohol– derived subunit levels are increased in compression wood, which forms during mechanical or gravitational stress and in wood knots (3). In woody angiosperms, lignin is derived from coniferyl alcohol 2b and sinapyl alcohol 2c in roughly equal proportions. Precursors and derivatives of hydroxycinnamyl alcohols also contribute to the lignin structure. For example, acetylated monolignols (hydroxycinnamyl acetates) have been implicated in kenaf (Hibiscus cannabinus) (4) and woody angiosperms (5), and p-coumarate esters are found in all grass lignins, implicating hydroxycinnamyl p-coumarates as precursors (6, 7). Low levels (∼5%) of cinnamaldehydes and benzaldehydes are present in all isolated lignins and are responsible for the bright crimson staining of lignified tissues by phloroglucinol-HCl (8).

Figure 1

Some precursors and products in the lignin biosynthetic pathway. The normal lignin monomers are thep-hydroxycinnamyl alcohols 2;p-coumaryl alcohol 2a, coniferyl alcohol2b, and sinapyl alcohol 2c. Coniferaldehyde1b is normally reduced regioselectively to produce coniferyl alcohol 2b. When CAD activity is depressed, coniferaldehyde 2b accumulates and could polymerize or copolymerize into lignin. Dihydroconiferyl alcohol 4b, observed previously only as a minor component of softwood lignins, is presumed to derive from coniferaldehyde 1b via a 1,4-reduction followed by a 1,2-reduction. However, no mechanism for this conversion has been reported. p-Coumaryl alcohol2a is readily derived from its aldehyde 1a in the mutant, implying that different CAD enzymes are involved for1a2a versus1b2b.

Removal of lignin from wood and plant fibers is the basis of chemical pulping to produce diverse pulp and paper products. Genetic engineering of the lignin biosynthetic pathway to lower lignin concentration or construct lignins more amenable to extraction is an active area of current research (9). However, several mutations have been identified and characterized that affect the lignin biosynthetic pathway (10). In maize (Zea mays) and related grasses, mutants characterized by a brown midrib (bm orbmr) have modified lignin (11). The bmphenotype can result from changes affecting cinnamyl alcohol dehydrogenase (CAD) (for example, bm1 of maize) (12, 13), O-methyl transferase (OMT) (for example, bm3 of maize) (12, 13), or both CAD and OMT (for example, bmr6 of sorghum,Sorghum bicolor) (14). Mutations in two other maize genes also result in brown midrib phenotypes, but the products of these genes remain unknown. A mutation in the gene encoding ferulate-5-hydroxylase has been identified in Arabidopsis thaliana, but it does not result in a brown midrib phenotype (15). No lignin mutants have been previously identified in woody plants.

CAD catalyzes the last step of the lignin precursor biosynthetic pathway (Fig. 1), reduction of hydroxycinnamaldehydes 1 to hydroxycinnamyl alcohols 2 (the conventional lignin monomers or monolignols) (16). A reduction in CAD activity might lead to accumulation of hydroxycinnamaldehydes 1 that could copolymerize with normal lignin monomers. Transgenic plants, suppressed in the synthesis of CAD (9, 17), sometimes have red-brown xylem tissue resembling that of grass brown midrib mutants. Such plants have increased aldehyde levels, although little of the aldehyde may actually be incorporated into the lignin (9, 17). The molecular basis for the color has not been established, but higher order polymers of coniferaldehyde1b, synthesized in vitro, have a wine-red color (18).

Here we report that a viable loblolly pine, homozygous for the mutantcad-n1 allele (19), incorporates novel monomers into its lignin in response to a CAD deficiency. The lignin structural changes were extensive and not predicted by the current view of the lignin biosynthetic pathway. The wood of this mutant is brown-red (Fig.2), similar to the color of the xylem inbrown midrib mutants (11) and transgenic plants suppressed in lignin biosynthetic enzyme activity (9,17). The cad-n1 allele is inherited as a Mendelian recessive gene that maps to the same genomic region as thecad locus. The cad-n1 allele was identified in a well-characterized heterozygous loblolly pine genotype (clone 7-56). In homozygous cad-n1 plants, CAD activity is 1% or less of wild type, and relative abundance of cad mRNA transcript is greatly decreased. In mutant plants, free coniferaldehyde 1b(the CAD substrate) accumulates to a high level. Unlike transgenic plants suppressed in CAD, cad-n1 mutant seedlings have decreased lignin content (19).

Figure 2

(Left) Wood chips from normal wood and a homozygous mutant with reduced CAD activity showing the brown wood phenotype. The mutant wood was obtained from a field test containing progeny from a cross between two half-sib loblolly pines each heterozygous for the mutant cad gene. (Right) Immediately after debarking, 2-year-old trees. The cad-n1mutant is readily identified by the red-brown color of its wood.

Milled wood lignins (20) were isolated for nuclear magnetic resonance (NMR) analysis from the wood of a 12-year-old CAD-deficient mutant and a normal sibling from the same cross (Fig. 2). An estimate of the subunit composition of this unusual lignin fraction, on the basis of quantitative NMR and other analytical data, is given in Table1.

Table 1

Estimates of subunit compositions [from quantitative13C-NMR and DFRC (29) data] of the normal (cad) and mutant pine (cad-n1) isolated lignins. 2a, p-coumaryl alcohol units;2b, coniferyl alcohol units; 1, cinnamaldehyde units (p-hydroxycinnamaldehyde and coniferaldehyde units are not distinguished); 3, benzaldehyde units (p-hydroxybenzaldehyde and vanillin units are not distinguished); (*) aldehydes at ∼188 ppm in13C-NMR spectra are identified as 2-methoxybenzaldehydes6 (22) from unknown sources—the other substituents on the aromatic ring are unknown; 4b, dihydroconiferyl alcohol (+ traces of dihydro-p-coumaryl alcohol 4a), the major component of the cad-n1mutant lignin.

View this table:

NMR spectra show that both coniferaldehyde 7 and vanillin5 (21) end groups are present in the lignin of the pine mutant, as may be expected from the suppression of CAD. Wood from the mutant also had a higher extractable aldehydes content (19). The HMQC-TOCSY (heteronuclear multiple-quantum coherence–total correlation spectroscopy) spectrum reveals the side-chain coupling network in which protons 7, 8, and 9 correlate with the aldehyde carbonyl carbon, C-9 in 7 and the simple 1-bond correlation between C-7 and H-7 in 5 (Fig.3). However, such components are also present in milled wood lignins from normal loblolly pine (Fig. 3). From quantitative NMR, these aldehydes each account for ∼15% of mutant lignin units and ∼7% in the normal pine lignin (Table 1). More striking are 2-methoxybenzaldehyde components 6(22), the peaks at ∼188 parts per million (ppm), that are greatly enhanced in the mutant (Fig. 3). The source of these previously unreported 2-methoxybenzaldehydes in lignins is unknown. Lignins from both the normal and mutant trees contained higher than normal concentrations of p-coumaryl alcohol units because of the preponderance of knots.

Figure 3

Aldehyde carbonyl group correlations from the HMQC-TOCSY (23) spectrum of lignin from thecad-n1 mutant plant showing the presence of cinnamaldehyde (with correlations to the three side-chain protons) and benzaldehyde (single correlations) units in the lignin. The normal pine lignin carbon section is shown to the left. The 13C spectra shown on projections to the left of the figure are normalized to the same methoxyl level; cinnamaldehyde 7 and benzaldehyde5 signals are about twice as abundant in the mutant. The higher field aldehydes markedly increased in the mutant have now been identified as 2-methoxybenzaldehydes 6 (22); their source is unknown.

Dihydroconiferyl alcohol units 8 are present and predominant. The HMQC-TOCSY experiment (23) (Fig.4) identified the coupling network for the aryl propanol side chain (red and orange contours) that is consistent with model compound data (24). Products8a,b (red) representing hetero-coupling of dihydroconiferyl alcohol with a conventional lignin monomer or oligomer as well as dibenzodioxocins 8c (orange) from initial 5-5-homo-coupling of dihydroconiferyl alcohol monomers are present in roughly equal amounts, reinforcing the claim that dihydroconiferyl alcohol is a major monomer during lignification.

Figure 4

Regions of the HMQC-TOCSY spectra (23) of milled wood lignins from (A) the pine cad-n1mutant, (B) from a cad normal wood, and (C) from a synthetic lignin (26). Structure assignments are most easily seen in spectrum (C) from the synthetic lignin that derived from coniferyl alcohol 2b containing ∼2% dihydroconiferyl alcohol 4b (26). Although synthetic lignins of this type have quite different substructure ratios from plant lignins, they contain all of the structural units and are valuable for spectral assignment. Thus, in (C), β-aryl ether units 10, α,β-diaryl ethers11 (scarce in plant lignins), phenylcoumarans12, and resinols 13 are readily identified, along with coniferyl alcohol end groups 9 and the dihydroconiferyl alcohol units 8 (and the aldehyde units5 and 7 in Fig. 3). Red- and orange-colored contours show the unambiguously identified components 8arising from dihydroconiferyl alcohol monomers 4. NMR provides a convenient distinction between products of hetero-coupling of 4 with conventional lignin monomers or oligomers to give8a,b (red) and those from initial 5-5-homo-coupling of dihydroconiferyl alcohol monomers to give8c (orange). Both are equally represented in thecad-n1 mutant, whereas the normal pine has only the higher field component, and the synthetic lignin has a trace of the lower field component. NMR data from cross-coupled dimeric models for 4-O-β structures 8a and 5-β/4-O-α (phenylcoumaran) structures 8band the dibenzodioxocin 8c coincide with the lignin data observed here (24). In the CAD mutant, dihydroconiferyl units are dominant, displacing much of the intensity from the normal coniferyl alcohol–derived region. Some of the minor units can be seen in the pine samples when looking at lower contour levels (not shown). The normally predominant β-aryl ether (blue) and phenylcoumaran (green) components (B) are severely reduced in the cad-n1mutant, with only some β-ether peaks being observable at comparable contour levels—these may also arise from p-coumaryl alcohol (in addition to coniferyl alcohol). Gray contours are from intense methoxyl signals, carbohydrate impurities, and other lignin structures not discussed in this paper.

Dihydroconiferyl alcohol products are seen in synthetic lignins that are prepared from impure monomers. The best syntheses of hydroxycinnamyl alcohols 2 (25) from hydroxycinnamate esters or hydroxycinnamaldehydes 1 still produce small amounts of 1,4-reduction products 4. Purification of coniferyl alcohol is difficult because dihydroconiferyl alcohol cocrystalizes with it. A synthetic lignin prepared (26) from coniferyl alcohol 2b containing a few percent dihydroconiferyl alcohol 4b provides a convenient model for the lignin in the pine mutant. Its HMQC-TOCSY spectrum (Fig.4C) shows the same dihydroconiferyl alcohol side-chain signals as in the pine lignins. A parallel between the lignins isolated from the mutant tree and hydride reduction product synthetic lignins is apparent. A small amount of the initial dihydroconiferyl alcohol homo-coupling product 8c (orange) is seen in the spectrum of the synthetic lignin (Fig. 4C)—the saturated compound is quickly and efficiently dimerized via radical processes. Normal softwood lignins contain small amounts of dihydroconiferyl alcohol units (Fig. 4B). The source of these subunits is unknown. Although dihydroconiferyl alcohol and its glucoside have been detected in young plant tissues including pine (27) and may function as growth factors (28), they are not considered part of the normal lignin biosynthetic pathway.

CAD normally effects a regioselective “1,2-reduction” (at C-9) of coniferaldehyde 1b to produce coniferyl alcohol2b. Our results suggest that the loss of CAD activity activates or up-regulates pathways on the basis of “1-4 reduction” (at C-7) and subsequent 1,2-reduction during lignin formation to produce the dihydroconiferyl alcohol monomer 4b (Fig. 1). Analogously, synthetic preparation of coniferyl alcohol versus dihydroconiferyl alcohol can be selected by hydride reactions with 1,2- versus 1,4-regiochemistry (Fig. 1) (25). An alternative possibility is that a small structural change in the enzyme (for example, a disulfide bridge) affecting the active site of the CAD enzyme might be enough to provide the “hydride” equivalent to the 7-carbon site. This possibility is unlikely, and an alternative enzymatic activity is probably required because the relative abundance of steady-state cad mRNA transcripts is greatly decreased in the mutant, and the amount of CAD enzyme activity is reduced to ≤1% of wild type (19). If the biochemical reduction is not totally regioselective, the small amounts of 4b producing the dihydroconiferyl units 8 seen in normal pine lignins could be explained, but this rationale would not allow production of4b in such major proportions without a significant shift in enzyme activity or without enhanced activity of an alternate enzyme. At least one new enzyme would be required to explain these results. It is also possible that coniferaldehyde is not the precursor to dihydroconiferyl alcohol, and that its synthesis is up-regulated from other sources in the plant.

The amount of subunits derived from p-coumaryl alcohol2a in the mutant is unchanged (Table 1), whereas the amount of coniferyl alcohol subunits 2b is greatly reduced (29). These results imply that the formation ofp-coumaryl alcohol 2a uses an independent mechanism such as an additional enzyme with “1,2-reductase activity” specific for p-hydroxycinnamaldehyde1a. Furthermore, few dihydro-p-coumaryl alcohol4a units were detected (29). The 1,4-reductase activity proposed for the formation of dihydroconiferyl alcohol is therefore equally specific for coniferaldehyde 1b.

Incorporation of novel monomers into lignin is inconsistent with the high level of enzymatic specificity recently extended to lignin formation from observations of specificity in lignin biosynthesis (30). Independence from rigid enzymatic control is further supported by other examples of incorporation of nontraditional monomers into lignins: (i) ferulates and diferulates actively incorporate into lignins of grasses, effecting significant lignin-polysaccharide cross-linking (31); (ii) acylated monomers are implicated in a variety of species (4-7); (iii) 5-hydroxyconiferyl alcohol 2d-derived subunits are readily assimilated into a lignin polymer in OMT-deficient plants that have a reduced ability to produce sinapyl alcohol (32).

Well-characterized differences in lignin subunit composition have long been known between major taxonomic groups of higher plants—for example, between lignins of hardwood and softwood trees (33). However, the narrow range of variation in lignin compositions within groups (10) has suggested structural constraints imposed for vascular function and support. The ability of this pine mutant to produce a functional lignin polymer from unexpected subunits extends the limit of “metabolic plasticity” for the formation of lignin, within an individual species. Concepts of lignin function based on the previous range of lignin compositions must now be reexamined in view of the unusual structure and composition of lignin in this mutant pine. A greater understanding of these processes should increase our opportunities to modify lignin content or composition through genetic engineering.

  • * To whom correspondence should be addressed at U.S. Dairy Forage Research Center, USDA-ARS, 1925 Linden Drive West, Madison, WI 53706–1108, USA. E-mail: jralph{at}facstaff.wisc.edu

REFERENCES AND NOTES

View Abstract

Navigate This Article