The relationship between rates of lignin and cellulose decay in aboveground forest litter

Abstract

We examined the rates of lignin decay (dC₃/dt) and holocellulose decay (dC₂/dt) in aboveground leaf litter of predominately northern conifer forests to test our hypothesis that the rate of lignin decay is a linear function of the lignocellulose index (LCI = lignin/[lignin + holocellulose]). We proposed that lignin decays only when LCI > 0.4 and that LCI cannot exceed 0.7, approximating humus. These constraints suggest that holocellulose content of litter is a control on lignin decay. We evaluated this hypothesis with a set of 112 field studies in two ways, by: (1) comparing simulations from a decomposition model to observations across the full range of litter decay and (2) analyzing the relative rate of lignin decay ([dC₃/dt]/[dC₂/dt + dC₃/dt]) as a function of litter LCI. Simulated dynamics of LCI in decaying litter were highly correlated to observed patterns, particularly when litter fractions of water and ethanol solubles from model output were pooled with holocellulose fractions (mean R² = 0.87 ± 0.02, P < 0.01). More detailed analyses of 64 of these studies yielded variable relationships between lignin decay rate and litter LCI; a regression based on pooled data (N = 385; total number of observations) from these studies produced a slope and an intercept that were not significantly different from predicted (slope = 2.33, intercept = −0.93). Both site and litter characteristics had significant effects on the proposed LCI threshold for lignin decay (LCI ∼ 0.4), but no effects on slope or intercept, suggesting that the proposed lignocellulose control hypothesis is relatively robust across a range of litter and forest types used in this study.

Introduction

Decomposition is a complex ecological process that is strongly influenced by litter chemistry (Berg and McClaugherty, 2008, Berg et al., 1997). Simple models often calculate an empirical decay rate coefficient for overall litter decomposition that is inversely related to lignin content, with high lignin content litters decaying slower than low lignin litters (Meentemeyer, 1978, Melillo et al., 1989, Taylor et al., 1989, Aber et al., 1990, Aerts, 1997). More complex models variously divide litter into distinct chemical constituents and simulate the decay of each pool through time (e.g., Moorhead et al., 1999). In many of these more complicated models, the lignin content or overall lignocellulose index (LCI = lignin/[lignin + holocellulose]; Melillo et al., 1989) of the decaying litter influences the rates at which various compounds decay (Parton et al., 1987, Rastetter et al., 1991, Currie and Aber, 1997). However, other litter constituents seldom exert any direct mechanistic control on lignin decay.

In contrast to the more common modeling approaches, Moorhead and Sinsabaugh (2006) recently developed a theoretical model of litter decay that couples changing litter chemistry to specific microbial activity (the Guild Decomposition Model, GDM). A central hypothesis of GDM is that relative amounts of lignin and holocellulose (cellulose + hemicelluloses) define a threshold for alternative controls on decay. When holocellulose is relatively abundant and lignocellulose index (LCI) is low, few data suggest substantial degradation of lignin despite high rates of holocellulose decay (Berg and McClaugherty, 2008). Because LCI values increase during decomposition to a constant value of about 0.7 (Melillo et al., 1989, Hirobe et al., 2004, Osono and Takeda, 2005), Moorhead and Sinsabaugh (2006) argued that access to holocellulose likely reduces microbial degradation of lignin at low values of LCI simply because holocellulose is a much more energy-rich resource, but that as LCI increases an increasing proportion of the holocellulose is shielded by lignin and therefore lignin must be decomposed before more holocellulose will be accessible. This represents a new mechanism linking lignin to holocellulose decay in models of decomposing litter.

Although Moorhead and Sinsabaugh (2006) describe GDM in detail, in brief, concentrations of lignin (C₃) increase over time in decaying litter, which in turn, increasingly retards holocellulose (C₂) decay (Meentemeyer, 1978, Melillo et al., 1989, Taylor et al., 1989, Aber et al., 1990). For this reason, Moorhead and Sinsabaugh (2006) calculated the rate of holocellulose decay in two ways. First, they used a Michaelis–Menten function of substrate availability when LCI = 0:

$$d{C}_2/dt=V_{\max }{C}_2/\left(K_2+C_2\right)$$

where V_max is the maximum rate of decay and K₂ is the half-saturation coefficient. Next, they calculated holocellulose decay as a fixed fraction of lignin decay when LCI = 0.7:

$$d{C}_2/dt=\left(3/7\right)\left(d{C}_3/dt\right)$$

because Melillo et al. (1989) noted that decomposing humus maintained a constant LCI = 0.7 (i.e., C₂:C₃ = 3:7).

Moorhead and Sinsabaugh (2006) also calculated the rate of lignin decay with a Michaelis–Menten function when LCI = 0.7, arguing that lignin degradation was probably the rate-limiting step at this stage of lignocellulose decomposition (Melillo et al., 1989):

$$d{C}_3/dt=V_{\max }{C}_3/\left(K_3+C_3\right)$$

However, the decay rate for lignin was also defined in terms of dC₂/dt when LCI = 0.7 (see also Eq. (2)):

$$d{C}_3/dt=V_{\max }\left(7/3C_2\right)/\left[K_3+\left(7/3C_2\right)\right]$$

Lignin may inhibit holocellulose decay by interfering with the affinity of extracellular enzymes to bind with substrate (Sinsabaugh and Linkins, 1989, Berg and McClaugherty, 2008), thus increasing the half-saturation coefficient of the Michaelis–Menten reaction (Eq. (1)). Moorhead and Sinsabaugh (2006) simulated this effect by replacing dC₂/dt on the left side of Eq. (2) with the right side of Eq. (1), and replacing dC₃/dt on the right side of Eq. (2) with the right side of Eq. (4):

$$V_{\max }{C}_2/\left(K_2^{\prime }+C_2\right)=3/7V_{\max }\left(7/3C_2\right)/\left[K_3+\left(7/3C_2\right)\right]$$

where $K_2^{\prime }$ is the effective half-saturation coefficient at LCI = 0.7, and thus:

$$K_2^{\prime }=K_3+4/3C_2$$

An exponential function was then used to calculate $K_2^{\prime }$ for values of LCI ≤ 0.7:

$$K_2^{\prime }=K_2\exp [\text{LCI}\left(\ln \left(K_3+4/3C_2\right)-\ln \left(K_2\right)\right)/0.7]$$

where K₂ is the half-saturation value when LCI = 0.

Although many studies have variously reported the inhibitory effects of lignin on litter decay (see above), Berg and McClaugherty (2008) were among the first to note that lignin did not appear to decay when litter contained high concentrations of holocellulose (i.e., LCI was low) and suggested that microbiota preferentially utilize easily degraded compounds because it is more energetically efficient. Moorhead and Sinsabaugh (2006) suggested that the probable mechanism controlling this relationship is at the level of enzyme production by microorganisms, rather than a direct effect on enzyme activity, and described this control on lignin degradation as a linear function of LCI:

$$\left(d{C}_3/dt\right)/\left(d{C}_2/dt+d{C}_3/dt\right)=m\text{LCI}+b$$

where m and b are the slope and intercept, respectively. Initial estimates of these parameters were derived from detailed field studies reported by Aber et al. (1984). In brief, the absolute amount of lignin in foliage litter of six different tree species showed no sign of decay (i.e., dC₃/dt < 0) when LCI < 0.40, thus defining a minimum threshold for lignin decay:

$$\left(d{C}_3/dt\right)/\left(d{C}_2/dt+d{C}_3/dt\right)=0.0$$

Moreover, a constant LCI = 0.7 in decaying humus (Melillo et al., 1989) provided a maximum rate of lignin decay, relative to holocellulose decay, when LCI = 0.7.

$$\left(d{C}_3/dt\right)/\left(d{C}_2/dt+d{C}_3/dt\right)=0.7$$

Eqs. (9), (10) yield estimates for m = 7/3 and b = −0.9333 and permit calculating the rate of lignin decay from Eq. (8):

$$d{C}_3/dt=-1d{C}_2/dt\left(7.0\text{E}9\times \text{LCI}-2.8\text{E}9\right)/\left(-5.6\text{E}9+7.0\text{E}9\times \text{LCI}\right)$$

Very few data sets of sufficient detail exist to validate or test this relationship. Aside from the data used by Moorhead and Sinsabaugh (2006) to develop their model (Aber et al., 1984), perhaps the best known data are from a long-term series of decomposition studies performed in Europe (Berg et al., 1991, Berg and Johansson, 1998). In particular, these studies included detailed descriptions of the surface decomposition of leaf litter from tree species under field conditions in north temperate climates. Herein we use these data to evaluate the interactive lignocellulose control hypothesis of GDM as well as to perform an independent validation of model behavior.