| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Dep. of Environmental and Resource Sciences, Univ. of Nevada, Reno, NV 89557
* Corresponding author (qualls{at}unr.edu)
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONREFERENCES |
|---|
Abbreviations: t1/2, half-decay time • DOC, dissolved organic carbon
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONREFERENCES |
|---|
The conditions related to the tendency of soil C to become stabilized or sequestered have been classified as recalcitrance, interactions between molecules, and physical accessibility (Sollins et al., 1996). Recalcitrance refers to molecular-level characteristics such as elemental composition, functional groups, and molecular conformation that influence the inherent biodegradability of the molecule. Interactions refers to the intermolecular interactions with other inorganic or organic molecules that influence biodegradability—for example, complexation with Al. Physical accessibility refers to the location of the molecule with respect to enzymes or cells. For example, substrates may be located in submicrometer pores too small for microbial cells to enter (Sollins et al., 1996). These conditions do not necessarily occur exclusively.
The inherent recalcitrance of humic substances might be one reason that C tends to be stored for long periods in the soil. Humic substances are complex macromolecules modified from plant compounds or newly synthesized during decomposition (Stevenson, 1994). Because they accumulate in soil (Stevenson, 1994), they are believed to be resistant to microbial decomposition, but there is little direct experimental evidence for their inherent recalcitrance. One widely cited form of evidence of the recalcitrance of humic substances comes from the 14C dating of fractions of soil. In a Mollisol, the mean residence time of the humic acid fraction was 1235 yr (Campbell et al., 1967). However, the mean residence times of the unfractionated soil, fulvic acid fraction and humin fraction, were 870, 495, and 1140 yr, respectively. Radio-labeled fungal melanins that resemble humic acids have been prepared from fungal cultures and were found to be resistant to decomposition when incubated in soil (Linhares and Martin, 1978). Jandl and Sletten (1999) found that <5% of the C in the hydrophobic acid fraction of a water extract of litter was mineralized in a solution incubated 110 d.
Other studies suggest that humic substances are not so resistant to decomposition. An alkaline extract of a forest soil was incubated in liquid culture containing growth media, and wood-degrading basidiomycetes were able to reduce the color of the solution by 57% in 21 d (Gramss et al., 1998). Likewise, a humic acid extracted from forest floor material was incubated in liquid culture with glucose-containing growth media with manganese, and a litter-degrading basidiomycete was able to reduce the color of the solution by 75% in 42 d (Steffen et al., 2002). Up to 27% of aquatic humic substances were removed from solution when incubated in a glucose-containing nutrient broth, but <7% were removed without the full glucose broth (Hertkorn et al., 2002). A humic acid fraction from isolated lake sediments incubated in lake water lost >60% of C from the solution phase in 95 d (da Cunha-Santino and Bianchini, 2002). Beyond these observations, there are few comparisons of the decomposition of humic acid and other fractions of soil organic matter.
Our general objective was to compare the microbial mineralization rates of partially decomposed leaf litter that had been radiolabeled to that of the constituent fractions extracted from the litter: the humic acid, fulvic acid, hydrophilic acid, phenolic, and hydrophilic neutral fractions. Our specific hypothesis were: (i) the humic acid fraction would be mineralized with t1/2 on the order of years, (ii) the humic acid fraction would be mineralized most slowly of all fractions, (iii) the fulvic acid fraction would be mineralized more rapidly than the humic acid fraction, and (iv) that a weighted-average percentage mineralization of all constituent fractions would be similar to that of the litter from which they were extracted.
To test these hypotheses, tree seedlings were labeled with 14C and the leaf litter was allowed to decompose for 6 mo during which time it lost 57% of its mass. Then the partially decomposed litter was fractionated and the individual fractions were added to soil cores so that the mineralization rates of the individual fractions could be measured in an intact soil environment. The rationale for using partially decomposed leaf litter was to allow the process of humification to produce humic acids characteristic of decomposed organic matter. In a study of the formation and loss of humic substances during decomposition of pine (Pinus strobus L.) litter during 13 yr in the field, most of the increase in humic acid content occurred during the first year, in which 37% of the C was lost (Qualls et al., 2003).
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONREFERENCES |
|---|
The air-dried decomposed litter was ground in a ball mill and then extracted by the procedure outlined in Fig. 1 (Qualls et al., 2003). This procedure represents a slightly modified version of the procedure for analysis of humic substances found in Methods of Soil Analysis (Swift, 1996) with three additional steps from Leenheer's (1981) and Leenheer and Noyes' (1984) more detailed fractionation procedure to further separate out phenolic, hydrophilic acid, and hydrophilic neutral fractions. Two samples, each 0.664 g dry wt. (the equivalent of 0.300 g C) were extracted with 30 mL of 0.1 M NaOH using two sequential extractions under anoxic conditions in a N2–purged bag. After centrifugation at 3000 g, the solutions were quickly neutralized to pH 7.0 with about 0.5 mL of 6 M HCl, allowed to sit for 24 h, and then centrifuged again. The supernatants from both sequential extractions of both samples were combined and diluted to 667 mL. This extract was fractionated into hydrophilic and hydrophobic fractions, which were further fractionated into acidic, basic, and neutral components as follows. The diluted extract was pumped through 10 mL of XAD-8 resin. The weak hydrophobic acid (Leenheer and Noyes, 1984) fraction (i.e., the phenolic fraction) was eluted with 0.1 M NaOH. The effluent was acidified to pH 1, allowed to sit 24 h, and then centrifuged to recover the humic acid fraction. The supernatant was again pumped through XAD-8 resin at pH 1 and the fulvic acid fraction was then eluted with 0.1 NaOH. The effluent was then pumped through cation (AG-MP-50, Bio-Rad, Hercules, CA) and anion exchange columns (Duolite A-7, Diamond Shamrock Corp., Houston, TX) as described by Leenheer (1981). The hydrophilic acid fraction was then eluted with 1 M NaOH. Subsamples were taken for dissolved organic carbon (DOC) analysis (TOC 5050A, Shimadzu Corp., Columbia, MD) after each step of the procedure.
|
Subsamples of each fraction were analyzed for DOC, DON (Koroleff, 1983), and radioactivity within hours after isolation. Immediately after isolation, samples of the fractions were frozen until added to the cores 2 d later. Subsamples of the litter and the insoluble residue were wrapped in tin capsules and dropped into the combustion tube of a Shimadzu TOC 5050 at 680°C while the combusted CO2 was collected from the effluent gas stream bubbling through three test tubes, linked in series, containing 1 M NaOH. Radioactivity in the NaOH solutions was measured as described later. Total C and N in subsamples of the litter and insoluble residue was measured with a PerkinElmer 2400 CHN Analyzer (PerkinElmer Instruments, Shelton, CT).
Soil Core Preparation and Incubation
The soil used in the study was collected from the oldest soil of the Mt. Shasta mudflow chronosequence, one that has served as a classic example of a soil weathering chronosequence (Dickson and Crocker, 1953; Jenny, 1980; Sollins et al., 1983; Lilienfein et al., 2003). This soil was chosen because it has been widely used as an example of the rate of accumulation of forest floor and total soil C during soil development (Dickson and Crocker, 1953; Jenny, 1980; Sollins et al., 1983; Lilienfein et al., 2003). The soil was classified as a Humic Haploxerand. The 0- to 10-cm depth increment had a pH of 5.4, a bulk density of 0.93 ± 0.13, and contained 45 g kg–1 organic C, 21 g kg–1 allophane, and 1.8 g kg–1 oxalate-extractable Fe (Lilienfein et al., 2003). Twenty-four intact soil cores were taken from the upper 10 cm of the A horizon, excluding the O horizon, with 3.2-cm i.d. plastic tubes (80.4-cm3 volume). In addition, 25 plastic tubes were filled with a mixture of combusted, acid-washed, and neutralized sand containing 10% micrometer-sized silica dust (Soilmoisture Equipment Corp., Santa Barbara, CA). The silica dust was added to better mimic the soil texture and increase water holding capacity. After returning to the laboratory I determined the mass of the cores when all soil and sand cores were adjusted to a soil matric water potential of –20 kPa so that the microflora in all soils would be subjected to the same matric water potential. The soil and sand in the cores was left inside the plastic tubes, brought to near saturation, and immediately placed on a bed of diatomaceous earth on a membrane filter apparatus with a regulated vacuum of –20 kPa until the mass equilibrated (Qualls and Haines, 1992a). After weighing the cores to determine the target mass, I allowed them to dry from the top until at least 1 mL of water had been lost so that the additions of solution would not make the soil wetter than the target moisture content.
Aliquots of each of the following fractions were thawed and added to three soil and three sand cores: (i) ground litter from which the other components were extracted, (ii) the NaOH insoluble residue (humin), (iii) humic acid, (iv) the phenolic fraction, (v) fulvic acid, (vi) the hydrophilic acid fraction, and (vii) the hydrophilic neutral fraction. Solutions were drawn into the soil within seconds by the matric potential gradient, so it did not appear that the aliquot remained in a highly concentrated area at the top of the core. The purpose of additions to the sand cores was to evaluate the mineralization in the absence of adsorption interactions that may influence mineralization rates in soil. Because the sand cores were unable to supply exogenous nutrients, aliquots of the extracts or the solid material were mixed with a nutrient solution and a microbial inoculum. The nutrient stock was mixed like that of Stanford and Smith (1972), with both N and P and micronutrients, and then added at a concentration such that the C/N ratio of the amended extract was 8, similar to microbial biomass. The microbial inoculum was prepared as in Qualls and Haines (1992b). A mixture of freshly gathered forest floor and A horizon soil, from the site where the cores were taken, was suspended in water and chopped in a blender. The suspension was filtered sequentially through a 37-µm sieve and 0.2-µm membrane filter and the particles deposited on the membrane filter were washed with water, resuspended in water, and 100 µL was added to the aliquots of extract and solid samples. The ground litter and insoluble residue samples were mixed into the upper 2 cm of the soil or sand to prevent drying on the surface.
One sand core to which no radioactive substrate was added was used as a blank for determining background radioactivity. Cores were placed in sealed Mason jars and the respired 14CO2 was collected in two vials, each containing 5 mL of 1 M NaOH. Water was added to each core whenever it was below the target weight to maintain constant soil matric water potential. Cores were incubated for 365 d at 25 ± 0.5°C.
The NaOH solutions were removed on Days 1, 2, 7, 14, 21, 35, 63, 91, 119, 147, 175, 203, 231, 259, 287, 315, 343, and 364. The two 5-mL NaOH solutions were combined and 1 mL was added to 10 mL of Ecolite scintillation fluid (ICN, Inc., Costa Mesa, CA), which was counted for 10 min in a Beckman LS60001C liquid scintillation counter (Beckman Coulter, Inc., Fullerton, CA). Only disintegrations with energy from 18 to 160 KeV were counted to exclude chemiluminescence, which interferes in counting 14C in aqueous alkaline solutions. Personnel at Beckman Coulter determined counting efficiency using this wavelength distribution. This counting efficiency correction was necessary to convert raw counts to actual disintegrations per unit time (Beckman Coulter, 1995). The radioactivity of the blank sand core was subtracted from all measurements.
To compare the mineralization of the natural substrates with a very easily mineralized substrate, a tracer-level concentration of uniformly labeled 14C labeled D-glucose was added to three soil and three sand cores to which no other substrates were added. One milliliter of a 2.04 x 10–7 M solution of the uniformly labeled 14C D-glucose (specific activity 9000 GBq mol–1; Moravek Biochemicals, Brea, CA) was added to each of the cores, which were then treated like the other cores.
Analysis of Mineralization Curves
The curves for the cumulative C mineralized across time were fit to a two-phase (i.e., labile and refractory), first-order model for the accumulation of the product of two simultaneous first-order reactions (Molina et al., 1980):
 | [1] |
Multiple comparisons of the means for cumulative percentage of 14CO2 evolved after 1 yr of incubation were made using Tukey's HSD test after transformation of the values by the arcsin of the square root of the value expressed as a proportion using SPSS (2000). Transformation of the values was necessary because standard deviations tended to be larger for larger proportions.
|
RESULTS AND DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONREFERENCES |
|---|
|
|
|
The fulvic acid fraction mineralized over twice as fast as the humic acid fraction, but less than one-third was mineralized. This is consistent with the lower 14C mean residence times that have been observed in fulvic acid in soil (Campbell et al., 1967). In soil, one explanation for the younger mean age of fulvic acid could be that it is continually leached from freshly deposited plant litter into the soil. The water-soluble fraction of litter can be comprised of 37% fulvic acid shortly after senescence (Qualls et al., 1991). This study suggests that its more rapid decomposition rate is another reason for its lower mean residence time in the soil.
The phenolic fraction was defined as the component of the NaOH-soluble C that was hydrophilic at alkaline pH, but became hydrophobic at pH 7 (Fig. 1). Polyphenols, and perhaps some phenolic fragments of lignin dissolved by NaOH, would be expected to occur in this fraction (Qualls and Haines, 1991). Only about 36% of this fraction was mineralized in 1 yr, suggesting an intermediate degree of biodegradability. Using 14C ring-labeled pyrocatechol as a model phenolic compound, Martin and Haider (1979) found a mineralization rate of 24% in a soil after 84 d, which was very similar to our phenolic fraction which mineralized 26% in the same amount of time (Fig. 2).
The relatively easily mineralized hydrophilic acid fraction was a very small fraction of the total NaOH-extractable C, and may contain low-molecular-weight organic acids and perhaps fulvic-acid-like molecules with a high charge-to-size ratio. The solution containing the hydrophilic acid fraction contained NaCl, but this seems unlikely to have inhibited microbial mineralization because the NaCl would have been diluted by the much greater volume of pore water in the core and the initial mineralization rate was relatively rapid.
The relatively easily mineralized hydrophilic neutral fraction may contain uncharged labile compounds such as carbohydrates and other uncharged molecules originating from plant material and microbial biomass. The cumulative percentage of C mineralized from this fraction after 1 yr was, however, significantly less than that of glucose (57.6 vs. 66.2%; P = 0.025, with a two-sample t test).
The NaOH-insoluble residue of decomposed leaf litter is likely to be very different from the mineral-associated humin residue isolated from mineral soil. It was more easily mineralized than any of the NaOH-soluble fractions. The insoluble residue likely included a large proportion of cellulose, lignin, and lipids. I assume that the NaOH did not peptize and dissolve the lipids since a hydrophobic neutral fraction was not detectable in the extract. The humin fraction in mineral soil has a mean residence time comparable to the humic acid fraction (Campbell et al., 1967) and is believed to be comprised of bound humic acids and very stable alkyl components formed over very long periods of decomposition in the presence of mineral soil (Rice, 2001).
The mineralization rate of uniformly labeled 14C-D-glucose is shown in Fig. 3 for comparison with mineralization of the litter extract fractions. Glucose was mineralized to the greatest extent of all the substrates in soil. The mineralization curve for glucose did not fit the two-component model (Eq. [1]) well, so a modeled line is not shown in Fig. 3. The fitted curve tended to underestimate the rate in the early phases and overestimate the rate in the late phases. While glucose can be taken up directly and easily metabolized by cells, it has been shown in many studies that the complete mineralization of a substantial proportion of the C is delayed for periods approaching 1 yr. Amato and Ladd (1992) added 14C glucose to 23 soils and at the end of 308 d, 14.8 to 28.9% of the radioactivity remained in the soil, with 2.1 to 15.3% remaining in living microbial biomass. There have generally been three explanations for the persistence of the C originating from glucose: (i) that a proportion of the glucose C is used to form cellular components that are inherently difficult to decompose after death of the cell, (ii) that these components are otherwise stabilized by humification or physical protection, or (iii) that a fraction remains in living cells. Comparison of the mineralization curves of glucose with the most rapidly mineralized fractions of the decomposed leaf material, such as the insoluble material and hydrophilic neutral fractions, suggests that a large proportion of the C may undergo the initial stages of decomposition, enzymatic hydrolysis and cellular uptake, rapidly. However, complete mineralization of a portion of the glucose C is delayed by its transformation into more refractory cellular components.
|
|
|
Why the mineralization rates in sand tended to be lower than in soil is unknown. There was no indication that the lack of an initial microbial community was limiting because there was no obvious lag phase in the initial rates of mineralization in the sand (data not shown). Despite the addition of a nutrient amendment sufficient to provide a substrate C/exogenous N ratio of 8, there might still have been some degree of nutrient limitation. In fact, the mineralization of glucose was slowed to a greater degree than other substrates that presumably contained endogenous organic N (Table 3). Another difference between soil cores and the sand was that the soil cores contained excised and decomposing roots while the sand did not. It is possible that the decomposition of roots might have stimulated microbial activity in the soil and influenced the decomposition rate of the added substrates through cometabolism.
Mineralization of Litter compared with that of Constituent Fractions
One way to evaluate whether the method of separation, or the inherent separation of the components, affected the mineralization of the individual components is to compare a weighted average of the sum of its constituent fractions to that of the litter (Fig. 2, Table 1). The percentage of each constituent fraction in the litter (from Table 1) was multiplied by the percentage mineralized (from Fig. 2), and the product was summed for all fractions. This "weighted average mineralization rate" was 44% (±weighted SE of 2.5%) compared with the 39 ± 0.5% of the litter that was mineralized. A t test showed no significant difference between the means (P = 0.23). Thus, there was a reasonable correspondence between this reconstruction of the mineralization rate of the litter and the rates of the separated components.
A chemical separation of components might have three possible artifacts with respect to their properties in the natural material: (i) treatment with 0.1 M NaOH (and 1 M HCl in the case of the humic acid) may produce chemical modifications that affect biodegradability; (ii) intermolecular interactions in the whole material that affect biodegradability are disrupted by separation; and (iii) two fractions were solid phase while the others were liquid phase. Chemical treatment with NaOH might be expected to increase biodegradability if it resulted in extensive hydrolysis. However, the mineralization rate of the humic acid fraction was very slow. The removal of lignin or humic acid that might have inhibitory effects on the mineralization of other associated components might also be expected to increase mineralization rates of the other isolated components. For example, it has been hypothesized that encrustation with lignin physically protects a portion of the cellulose (Blanchette, 1991). Because the insoluble residue was in solid phase, it might be suspected that the more limited surface-to-volume ratio of the particles might tend to slow decomposition. In fact, the solid phase insoluble residue mineralized most rapidly and without evidence of a substantial lag phase. After the humic acid fraction was adjusted to pH 7, it was a colloidal suspension that was dispersed enough that it did not appear turbid. Consequently, it was more comparable to the other liquid fractions in terms of physical accessibility. In summary, the correspondence between the mineralization rate of the litter and the weighed average mineralization rate of the constituent fractions suggests that the separation did not introduce major artifacts.
Advantages and Limitations of the Method
The method of adding 14C-labeled substrates to evaluate the relative rates of mineralization of different fractions had several advantages: (i) the substrates could be added to intact soil cores (with the exception of the solid substrates mixed into the upper 2 cm), with a natural microbial community; (ii) the added substrates could be distinguished from native C; and (iii) the method yielded information on the entire decomposition curve up to 1 yr, not just the initial rates. The fact that the cores were intact was important because the physical soil structure was not disturbed, a factor that might be important in protecting C in interstices. In addition, the microbial community, including fungal mycelia, was intact. It was considered particularly important that fungal mycelia were not disturbed, since they are important in hydrolysis of lignin and humic substances (Blanchette, 1991; Steffen et al., 2002). One limitation of this study was that it was not possible to compare organic matter from a wide variety of sources because it requires labeling during a long period of growth, something much more difficult for perennial plants like trees. Humic and fulvic acids from different sources may differ in their biodegradability. For example, aquatic humic substances from groundwater, lake water, and river water showed different rates of removal from solution when incubated in liquid cultures (Hertkorn et al., 2002). However, in another study, humic acids from alkaline extraction of lake sediment, and humic acids extracted with water from decomposing macrophyte litter exhibited relatively similar rates of removal from solution during aqueous incubation (da Cunha-Santino and Bianchini, 2002).
|
ACKNOWLEDGMENTS |
|---|
Received for publication July 11, 2003.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  F. Adani and M. Spagnol Humic Acid Formation in Artificial Soils Amended with Compost at Different Stages of Organic Matter Evolution J. Environ. Qual., June 23, 2008; 37(4): 1608 - 1616. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| The SCI Journals | Agronomy Journal | Crop Science | |||
| Journal of Natural Resources and Life Sciences Education |
Vadose Zone Journal | ||||
| Journal of Plant Registrations | Journal of Environmental Quality |
The Plant Genome | |||
0. All overall models, with the r2 indicated in the table, were significant.
