Cyanobacteria andAzolla Chapter 7
Chapter 7
Cyanobacteria and Azolla as
Cyanobacteria
The cyanobacteria or blue-green algae are photosynthetic bacteria and some of them are able to fix N2(Chapter 2). They can be divided into two major groups based on growth habit: the unicellular forms and the filamentous forms. N2-fixing species from both groups are found in paddy fields but the predominant ones are the heterocystous filamentous forms (Table 7.1). Cyanobacteria are not restricted to permanently wet habitats as they are resistant to desiccation and hot temperatures, and can be abundant in upland soils (Roger and Reynaud, 1982). However, wet paddy soils and the overlying floodwaters provide an ideal environment for them to grow and fix N2.
Natural distribution
Free-living cyanobacteria can grow epiphytically on aquatic and emergent plants as well as in floodwater or on the soil surface. Early surveys indicated that cyanobacteria were only present in a small proportion of rice fields. Only 5% of over 900 soil samples from Asia and Africa (Watanabe, 1959; Watanabe and Yamayoto, 1971) and
Subdivision of the
Cyanobacteria Important genera and their morphology Unicellular group
Section I Anabaenagroup Section IV Nostocgroup Section IV Aulosiragroup Section IV Scytonemagroup Section IV Calothrixgroup Section IV
Gloeotrichiagroup Section IV
Fischerellagroup Section V
Unicellular strains (Aphanothece,Gloeothece)
Heterocystous strains with a thin sheath, without branching, do not form mucilaginous colonies of definite shape (Anabaena,Nodularia,Cylindrospermum,Anabaenopsis) Heterocystous strains with a thick sheath, without branching, forming mucilaginous colonies of definite shape (Nostoc) Heterocystous strains with a thick sheath, usually without branching, do not form diffuse colonies on agar medium (Aulosira)
Heterocystous strains with false branching, without polarity, forming velvet-like patches on agar medium (Scytonema) Heterocystous strains with false branching, with polarity, forming velvet-like patches on agar medium (Calothrix, Tolypothrix,Hassalia)
Heterocystous strains, with polarity, forming mucilaginous colonies of definite shape (Gloeotrichia,Rivularia) Heterocystous strains with true branching (Fischerella, Westiellopsis,Stigonema)
Table 7.1. The main taxa of N2-fixing cyanobacteria (indicating the Section of the Cyanobacteria to which the groups belong – see Table 2.3) found in rice soils in Southeast Asia. (From Rogeret al., 1987.)
147
Z:\Customer\CABI\A4042 - Giller - Nitrogen Fixation\A4042 - Giller + Watson - Nitrogen Fixation SET #L.vp 04 June 2001 10:56:57
only some 33% of more than 2200 samples of Indian rice soils (Venkataraman, 1975) were found to contain cyanobacteria. This reported infrequent occurrence was almost certainly due to only small samples of soil being taken from each field and also to the use of unsuitable methods for detection. Many other studies have found cyanobacteria in all of the soils sampled (Roger and Reynaud, 1982, and references therein). Although the relative abundance may vary widely, heterocystous genera generally account for about half of the cyanobacteria in rice fields (Whitton and Roger, 1989). In fact, in deepwater rice fields studied in Bangladesh virtually all of the cyanobacteria were heterocystous forms (Whittonet al., 1989).
Numbers of heterocystous cyanobacteria in rice soils expressed as colony- forming units (cfu) ranged from 10 to 107cfu g-1soil, with a mean value of 2.5´105cfu g-1soil or 8.3´104cfu cm-2in ten surveys in which more than 280 soils were sampled (Rogeret al., 1987). In most of these studies a most probable number (MPN) method for counting was used, in which serial dilutions are made and the population size is estimated on the basis of presence or absence of growth of cyanobacteria at the different dilutions. However, a modified direct-plating method using selective media gave an average value of 3.2´105cfu cm-2 over 102 soil samples, which was roughly four times greater than the mean number found previously (Roger et al., 1987). On average, heterocystous cyanobacteria formed less than 10% of the population of eukaryotic green algae and the abundance of cyanobacteria increased both with the amount of available phosphorus and with pH values over the range 4–6.5. Above pH 6.5 the numbers of cyanobacteria showed no obvious relationship with pH. These results agree with earlier observations that N2-fixing cyanobacteria are more abundant in phosphorus-rich soils of neutral to alkaline pH (Roger and Kulasooriya, 1980).
Amounts of N2fixed by cyanobacteria in rice production
An average value from 38 measurements of N2-fixation by cyanobacteria collated from the literature was 27 kg N ha-1per rice crop, with a maximum of 50–80 kg N ha-1 (Roger and Kulasooriya, 1980). However, most of these measurements were made using the acetylene reduction assay and are unlikely to be accurate, given the problems of calibration of the assay and the well-documented diurnal fluctuations in measured rates of N2-fixation (Chapter 4). In a detailed study of ARA due to cyanobacteria in 190 rice fields in the Philippines, the mean activity was 126mM C2H2 reduced m-2h-1, roughly equivalent to 12 kg N ha-1 fixed over a cropping season (Fig. 7.1). ARA estimates of N2-fixation indicated that greater amounts of N were fixed by cyanobacteria (7 kg N ha-1) on the wet soils before flooding than in the standing waters (2 kg N ha-1) of deepwater rice fields (Rotheret al., 1989). A bloom of cyanobacteria usually contains less than 10 kg N ha-1, though a dense bloom may contain up to 25 kg N ha-1(Rogeret al., 1986; Roger and Ladha, 1992).
Such blooms may exhibit high rates of N2-fixation and can persist for several weeks (Rotheret al., 1989). For a cyanobacterial bloom to be sufficiently large to
make a significant input of fixed N, it would have to be readily visible in the field (Rogeret al., 1986).
Of course, in the short term, the important measurement is not simply the amount of N2fixed but the amount acquired by the rice crop. Experiments using
15N-labelled cyanobacterial cells spread on the soil surface or incorporated into the soil showed that between 36 and 51% of the added N was recovered by rice in the first season (Wilsonet al., 1980). Similar pot and field experiments indicated that 23–28% of the N in15N-labelled cyanobacteria incorporated into the soil was recovered in the first rice crop whilst only 14–23% of the N was recovered if the cells were left on the soil surface (Tirolet al., 1982). Timing is also important: N2
fixed or released towards the end of the growing season will be too late to influence production of the current rice crop (Whitton and Roger, 1989).
Based on these data it can be assumed that one-quarter of the N2 fixed by cyanobacteria is utilized by the next rice crop. Then if 15–25 kg N ha-1are fixed during each crop, this would represent a benefit of some 4–6 kg N ha-1. Thus the amounts of N2 fixed by cyanobacteria are likely to be insufficient to sustain high yields of rice but will be important in the long-term maintenance of soil fertility in paddy fields.
Inoculation with cyanobacteria
The earlier reports of the sparse distribution of cyanobacteria in paddy fields have been used as a justification for an intensive research effort into technology for inoculation with cyanobacteria, or ‘algalization’, particularly in India. A method for production of algal inoculants was developed in India that was suitable for use by small-scale farmers (Venkataraman, 1981). An initial inoculum containing six spe- cies of cyanobacteria was provided to farmers by the ‘All India Coordinated Project
Fig. 7.1. Measurements of acetylene reduction activity (ARA) in flooded rice soils.
Each value is the mean of nine to 13 measurements during a crop cycle. Over a cropping season, 10mM C2H2m-2ha-1 is estimated to be roughly equivalent to 1 kg N2fixed ha-1. (Unpublished data of P.A. Roger.)
149
Z:\Customer\CABI\A4042 - Giller - Nitrogen Fixation\A4042 - Giller + Watson - Nitrogen Fixation SET #L.vp 04 June 2001 10:56:57
on Algae’. The farmers then multiplied this inoculum in shallow tanks in up to 15 cm of water to which some soil, phosphorus fertilizer and insecticide were added, together with some lime where necessary. The tanks were simple in design, either consisting of a shallow pit lined with plastic sheet, or larger inoculum production units were made by mounding up soil to make shallow bunds (ridges) in the field.
After a few weeks a mat of cyanobacteria and green algae developed; this was allowed to dry out and flakes of the inoculum were scraped up and stored for later use.
Sufficient flakes to inoculate 1 ha (8–10 kg) were produced from a single tank in 2–3 months. The density of total cyanobacterial and algal propagules in these inocula varied from 2´106 to 9´107cfu g-1soil but only 2–32% of these were hetero- cystous cyanobacteria, the type most important for N2-fixation (Rogeret al., 1987).
It has been suggested that use of a multi-strain starter inoculum and production of inoculum using local soil in the tanks would lead to selective growth of the strains best adapted to the local soil conditions. However, this was not borne out by the results of Rogeret al. (1987), who found that one or two strains, most commonly a Nostoc, were generally dominant among the N2-fixing cyanobacteria present in the inocula produced. The results of these workers, indicating that N2-fixing cyanobacteria are in fact abundant in rice fields, put the necessity for inoculation in question. Nevertheless, should inoculation be deemed necessary, Rogeret al. (1987) recommended that a better multi-strain inoculum would be produced by mixing single-strain inocula at the time of inoculation. Even so, as the inoculum is produced in local soil, it is still possible that strains present in the soil may dominate and so the original inoculum strains are lost before the inoculum even reaches the field (Whitton and Roger, 1989).
Results of inoculation experiments are not encouraging, even though many of these were conducted with inocula produced in the laboratory, i.e. in which local strains had not had a chance to outcompete inoculum strains prior to inoculation. An average response of 15% in yield of rice was found in experiments where inocula of cyanobacteria were applied in the field (Roger and Kulasooriya, 1980). In many experiments the effects of inoculation and N-fertilizer application on growth of rice were often additive and were attributed to production of plant growth-stimulating compounds by the cyanobacteria and not to N2-fixation. However, screening of 133 strains of cyanobacteria showed that 70% of the strains had an inhibitory effect on germination of rice and only 20% of the strains stimulated elongation of rice shoots (Pedurand and Reynaud, 1987). This suggests that hormonal effects of cyanobacteria are not the principal cause of improved rice growth when responses to inoculation are observed.
In many cases no benefits in yield were found after inoculation (Roger, 1995b).
This is perhaps not surprising as the recommended rate of inoculum application will provide on average less than one propagule for every 130 indigenous cyanobacteria already present (Roger and Kulasooriya, 1980). In one study the inoculated strains could not be detected even immediately after inoculation, presumably as they formed such a small proportion of the total algal population (Grantet al., 1985). In other experiments, inoculated strains did multiply but rarely dominated the population of cyanobacteria (Reddy and Roger, 1988). When beneficial effects on plant growth
due to inoculation have been found it is likely that establishment of large populations of cyanobacteria has been possible due to the large phosphorus content of propagules in the inocula, which will give the introduced cells a substantial growth advantage over indigenous strains (Roger et al., 1986). In any case, whether the inoculum strains succeed in becoming established or not, there is little evidence indicating that they can fix N2more effectively than indigenous cyanobacteria.
Manipulation of indigenous populations of cyanobacteria
Given the lack of success of inoculation, an alternative strategy to improve the inputs of N2-fixation is to enhance the growth of indigenous cyanobacteria (Roger, 1995a).
Low pH, low temperatures and phosphorus deficiency are all factors that are known to limit growth, but the ecology of cyanobacteria is poorly understood (Roger and Watanabe, 1986). It is apparent that addition of phosphorus fertilizers is likely to stimulate their growth and addition of lime to floodwaters will help in acid soils.
The most important management practice is to add phosphorus fertilizers to the floodwater, combined with ‘deep placement’ of the N fertilizer (Roger, 1995b).
When urea fertilizers are broadcast, substantial amounts of mineral N are present in the floodwater. As N is not limiting, green algae proliferate and bloom, effectively outcompeting the cyanobacteria. Dense blooms of green algae tend to cause the pH of the floodwater to rise, creating ideal conditions for loss of N through volatilization of ammonia. If N fertilizer is placed below the surface of the soil, the rice plants are able to take full advantage of the N without danger of gaseous loss of ammonia, and as N concentrations in the floodwater remain small, cyanobacteria are able to bloom and contribute N to their full potential. Unfortunately, this practice is rarely economic, due largely to the extra labour required to place the N fertilizer.
Another problem that restricts the size of populations of cyanobacteria is predation by invertebrates. Some cyanobacteria (e.g.Aphanothece,Gloeotrichiaand Nostoc) are able to form mucilaginous colonies, which renders them more resistant to grazing by invertebrates (Grantet al., 1985), but such strains generally contain little N. A standing crop of these cyanobacteria of 10 t ha-1 may contain as little as 3 kg N ha-1(Grantet al., 1985). Pesticides, including some natural products from plants such as neem (Azadirachta indica), can be used to reduce grazing pressure (Reddy and Roger, 1988). Research interest has been directed towards the isolation or production of pesticide-resistant strains but this work has not been extended outside the laboratory (Whitton and Roger, 1989).
Azolla as a Green Manure
The aquatic fern azolla is always found in an N2-fixing symbiosis with a cyanobacterial species calledAnabaena azollae, although the exact identity of the endosymbiont remains uncertain (as discussed in Chapter 2). Seven species ofAzolla are recognized in most taxonomic schemes.
151
Z:\Customer\CABI\A4042 - Giller - Nitrogen Fixation\A4042 - Giller + Watson - Nitrogen Fixation SET #L.vp 04 June 2001 10:56:58
Azolla can either be grown alone in rice fields before rice planting, if sufficient water is available to flood the fields, or it can be grown among the rice plants and periodically dug into the soil. A period of 2–3 weeks’ growth is required to accumulate sufficient biomass before this treatment can be carried out. In southern China, azolla is most commonly grown in the winter before transplanting of rice, whilst in the north it is either grown before rice, or intercropped with rice, or both (Lumpkin and Plucknett, 1982). In some areas rice is grown in densely planted strips with wide gaps containing azolla, which is incorporated into the soil several times during each crop. In Vietnam, a first crop of azolla is mounded up after 2–3 weeks’
growth and composted, allowing further multiplication and growth of the remaining fronds for 10 days. The compost is then turned into the soil with part of the living azolla when rice is transplanted, and the remainder of the living azolla is allowed to multiply between the transplanted rice plants so that more azolla can be turned into the soil later on, giving a total of approximately 80 kg of azolla N ha-1(Roger and Watanabe, 1986).
Production of ‘inoculum’
As many rice paddies are dry for much of the year, azolla must be reintroduced each growing season by inoculation. Azolla can be maintained in ponds, irrigation channels or slowly moving streams. The inoculum density required to ensure good establishment is high (2–5 t ha-1) to prevent the azolla from being overgrown by algae or weeds (Watanabe, 1982). The only way to achieve this practically is to multiply the initial inoculum within small areas of the fields where it is to be used (Van Hove, 1989). In Vietnam the initial inoculum, which is supplied from government farms, is multiplied in a part of the rice field until there is sufficient to be spread out over double the area. This process is then repeated, giving an exponential increase in the area covered.
Constraints to the growth of azolla
Under optimal conditions, azolla can double its mass every 2–3 days (Watanabe, 1982). In controlled environments, the growth rate ofA. mexicana,A. microphylla andA. pinnatawas greatest above 30°C butA. carolinianaandA. filiculoidesgrew better at temperatures below 25°C (Peterset al., 1980; Watanabe and Berha, 1983;
Kannaiyan and Somporn, 1989). However, total biomass production was greater in all species at temperatures below 25°C and differences in tolerance to higher tem- peratures were observed between strains (Watanabe and Berha, 1983).A. pinnata, A. microphyllaandA. mexicanawere also the species most tolerant to transient high temperatures (> 40°C) (Uhedaet al., 1999). Experiments in whichA. filiculoidesand A. microphyllawere grown with reciprocal crosses ofAnabaenaisolates have demon- strated that tolerance to high temperatures requires adaptation of both symbionts (Watanabeet al., 1989b).
In natural conditions various constraints, such as unfavourable temperatures, phosphorus availability, lack of water and insect pests, can limit azolla production.
Growth of azolla in the field is generally poor in the humid tropics when the average monthly temperature is greater than 27°C, but the highest rates of azolla production in the dry climate of Senegal (which were much greater than the rates of azolla production found for the same strains of azolla in the Philippines) occurred in months with a mean temperature above 29°C (de Waha Baillonvilleet al., 1991).
Shading by the rice crop can help in hot periods by reducing the temperature of both the floodwater and the air close to the surface. Conversely, shading by rice can severely limit growth of azolla, as it requires a high light-intensity for growth (Lumpkin, 1987). Temperatures that are too high in summer or too low in winter for azolla to survive obviously cause problems for the use of azolla in agriculture and many different methods have been devised in China to ensure survival of an inoculum for the next growing season (Lumpkin and Plucknett, 1982).
Growth of azolla is often limited by phosphorus availability. Azolla growing in the field commonly develops a reddish-purple coloration, which has often been considered to be characteristic of phosphorus deficiency, although intense sunlight and other physiological stresses can also cause azolla fronds to produce anthocyanins (Watanabe, 1982; Van Hove, 1989). If the water is shallow (< 3 cm deep) azolla can root into the soil, which may help in uptake of phosphorus, but generally addition of phosphorus fertilizers is necessary to ensure good growth. Watanabeet al. (1988) suggested that phosphorus fertilizers should be applied to nursery beds in split doses. Azolla enriched in P in this way can multiply several times before it becomes P-deficient, and the better growth can increase the N benefit to rice (Singh and Singh, 1995). There appear to be differences between azolla species in their tolerance to phosphorus limitation:A. pinnatagrew better under P-limiting conditions than A. microphyllaorA. mexicana(Kushari and Watanabe, 1991). At least some of the phosphorus added to azolla will be made available to a subsequent rice crop when the green manure decomposes.
If the paddy dries out, azolla can only survive for a few days and this can severely limit use of azolla in areas with unpredictable rainfall, unless the fields are irrigated (Lumpkin, 1987). A number of insect larvae attack azolla and damage tends to be particularly severe in hot climates (Watanabe, 1982). Damage by insects appears to be one of the most important factors that lead to poor performance of azolla in the field in Asia.
Amounts of N2fixed
Many of the early experiments looking at amounts of N2fixed by azolla were carried out using the acetylene reduction assay. Calibration of ARA measurements of N2-fixation showed that conversion ratios for C2H2: N2 ranged from 1.6 : 1 (theoretically impossible) to 7.9 : 1 and varied with theAzolla species under test, the length of the assay and the age of the culture (Eskew, 1987). Thus ARA measurements are unlikely to give reliable measurements of N2-fixation by azolla.
153
Z:\Customer\CABI\A4042 - Giller - Nitrogen Fixation\A4042 - Giller + Watson - Nitrogen Fixation SET #L.vp 04 June 2001 10:56:58