, Volume 6, Issue 3, pp 390–403

The Influence of Indigenous Food Procurement Techniques on Populations of Cyanobacteria in pre-European Australia: A Potential Small-scale Water Amelioration Tool


DOI: 10.1007/s10393-010-0276-3

Cite this article as:
Sadgrove, N.J. EcoHealth (2009) 6: 390. doi:10.1007/s10393-010-0276-3


During times of pre-European Australia, indigenous people utilized methods of food procurement that resulted in toxic phytochemicals from plants entering their waterholes. This paper focuses on three of these plants, namely the leaves of Acacia colei and Duboisia hopwoodii, which were used by hunters to poison water holes to stun fish or a drinking animal, and the seeds of Castanospermum australe, which were eaten following the leaching of toxins into a running stream. If consumed by humans, the main toxins from these plants—saponins/sesquiterpenes, nicotine/nornicotine, and australine/castanospermine—are fatal. However, it is undetermined whether populations of Cyanobacteria also can be affected. During this study, the previously mentioned plants were administered to populations of the species Anabaena circinalis, Microcystis aeruginosa, and Nodularia spumigena, while mimicking the traditional applications of these plants as closely as possible. Results varied with treatments and species; however, cell chlorosis manifested in nearly all treatments, concomitantly with thylakoid membrane disorganization. Cell dormancy typically manifested, along with destruction of populations at higher treatments. The results indicated that populations of Cyanobacteria could have been destroyed or inhibited by indigenous people during traditional applications of these plants. Findings presented herein indicate a more sophisticated and complex traditional Australian resource management scheme than is currently understood, contributing to the growing awareness of the plight of earlier indigenous Australians. The reintroduction of traditional water management techniques may have potential as a suitable small-scale water resource management strategy.


DuboisiaAcacia coleiCastanospermum australeCyanobacteriaAustralian Aboriginalnicotine


Cyanobacteria are autotrophic microscopic organisms existing in single, colonial, or filamentous structures in the water. In sufficient numbers, Cyanobacteria form a green or red slime, which reduces the quality of water bodies, both inland and coastal. When the concentration of cyanobacterial cells exceeds 20,000 cells per mL−1 they are capable of becoming a threat to humans (Whitton and Potts, 2000). When populations of Cyanobacteria reach threatening numbers the term “algal bloom” is applied. Today’s broad distribution and severity of cyanobacterial blooms is believed to be the result of land management changes that occurred during the industrial revolution and the colonisation of Australia by Europeans. It is now widely understood that these changes have benefited populations of Cyanobacteria by contributing additional nutrients to waterways (Codd et al., 1994).

Currently the cultures of the Australian indigenous people are receiving growing recognition for sustainable land management practices (Yibarbuk et al., 2001), which largely inhibited the formation of toxic cyanobacterial blooms. Despite this, indigenous people still did encounter blooms to some extent (Anonymous, 1880; Hayman, 1992, 1994). For some indigenous groups in Australia this posed a significant threat, particularly in areas underlain by sandstone where groundwater percolated further underground and populations relied heavily on small waterholes instead. The suggestion in this paper is that these dangers may have been alleviated to some extent, after the addition of phytochemicals to the water during food procurement.

The two plants that were most widely used this way in Australia were Acacia colei (or holosericea) and Duboisia hopwoodii, which are the quickest and most effective Australian animal poisons. Indigenous people used A. colei as a fish poison, which was released by placing the crushed bark and leaves of the plant in a dilly bag (a bag made of grass or palm leaves) and soaking in a waterhole. Approximately half an hour later the fish would float to the surface to be collected and eaten (Hiddins, 1999; Latz, 2004). After this procedure, the main compounds released into the water were saponins, tannins, and flavonoids (Cribb and Cribb, 1981; Isaacs, 2000). No literature has been found to determine exactly how fish were affected by these compounds; however, it is likely that the fish were suffocated when the saponins precipitated and clogged their gills. Alternatively, asphyxiation may have resulted from oxygen immobilization during saponin precipitation (Cribb and Cribb, 1981). Saponins themselves are conceded to be slightly toxic (Webb, 1969); however, toxic compounds from A. colei contribute less to its function as a fish “poison,” than with other native Australian plants used for the same purpose (e.g., Abrus precatorius, Acacia ditricha, Alphitonia sp., Barringtonia acutangula, Eucalyptus confertiflora, E. microtheca, E. polycarpa, Melaleuca argentea) (Hiddins, 1999; Isaacs, 2000; Lassak and McCarthy, 1983; Low, 1989). Due to the broad range of plants used in Australia as fish poisons, there is confusion in the non-Aboriginal community over the resultant toxicity of the waterhole, and the subsequent period of time before the water became drinkable again. In extreme cases, the waterhole required flushing by rain before it became drinkable again. However, after the use of A. colei it is possible that the water became drinkable in just a matter of days (Latz, 2004).

The Anangu people of central Australia (Pitjantjatjara and Yankunytjatjara) used Duboisia hopwoodii [Solanaceae] most frequently at Uluru, to stun emus during hunting (Latz, 2004; Robinson, 1980). The method involved soaking the crushed leaves in a waterhole, then hiding all other water sources with tree branches before waiting for an animal to approach and drink the poisonous water. This method was used to hunt all animals but was most affective on birds, such as the emu (Latz, 2004). The animal typically became “stupefied” or “drunk,” which made it easier to hunt.

The cholinergic compounds responsible for the toxicity of D. hopwoodii leaves are anabasine, nicotine, and nornicotine (Bottomley and White, 1951). Each of these has an affinity for the nicotinic acetylcholine receptor in mammals and works in the same manner as acetylcholine; however, the enzyme “acetylcholinesterase” is much less capable of closing nicotine and nornicotine gated ionic channels. In large dosages, these alkaloids can overstimulate the mammalian nervous system, potentially leading to respiratory failure and death (Lehninger and Nelson, 2004).

Again, there are divided opinions over the time-lapse required before a waterhole, previously poisoned by leaves from D. hopwoodii, detoxified and the water became drinkable again (Hiddins, 1999; Latz, 2004; Robinson, 1980). Latz (2004) suggested that, after the application of D. hopwoodii to hunt emus at Uluru, indigenous people reused the waterhole when flushed by the next rain. However, Latz (2004) also reported with uncertainty that the toxic nature of the waterhole diminished in a given timeframe despite flushing. This was confirmed during experimentation, when nornicotine evidently degraded into myosmine within a month and possibly a lot sooner, with approximately 30% efficiency. Myosmine is a nontoxic alkaloid (Zwickenpflug and Tyroller, 2006) that is present in a variety of grain foods and fruits. It is undetermined if a similar detoxification pathway would unfold after the use of other emu and fish poisons; however, it is worth further research.

Geographical chemical variations (chemotypes) of D. hopwoodii resulted in differences in its use by Aboriginal people across Australia. In regions other than the Northern Territory, where D. hopwoodii was much less toxic, it was used as a chewing tobacco, which was referred to as pituri by Europeans (Lassak and McCarthy, 1983). Extensive pituri trading routes existed to transport the chewable plant into regions where it grew naturally but was too poisonous for the locals to use for anything other than poison (Robinson, 1980). It is alleged that D. hopwoodii induced shamanistic effects when it was used as a chewing tobacco (Lassak and McCarthy, 1983). Although most of the usual effects were experienced, European settlers were not able to create this “shamanistic” experience when trying the substance for themselves (Robinson, 1980). The only report of shamanistic compounds, characterised from D. hopwoodii, focused on a chemotype growing in the Northern Territory near Alice Springs. Hyoscine, hyoscyamine, and the more commonly known scopolamine were found in the roots of this particular chemotype (Barnard, 1952; Launratana and Griffin, 1982). Despite the presence of these compounds, the roots of D. hopwoodii were never used by Australian Aboriginal people and this chemotype was not used for anything other than a poison. These compounds are more frequently found in the leaves of Duboisia myoporoides—plantations in Queensland currently supply the bulk of the world’s raw scopolamine (Foley, 2006).

This study also focused on Castanospermum australe, which produces a poisonous seed that was made edible by indigenous people after a lengthy method of preparation. The method involved scraping the seed into a vermicelli-like substance using a jagged mussel shell—some indigenous groups also elected to roast the seed at this point. The seed was then placed into a dilly bag and soaked in a running stream for 10 days, before it was cooked and eaten without toxic affects to the consumers (Cherikoff, 1993; Hiddins, 1999; Isaacs, 2000; Latz, 2004; Low, 1989; Robinson, 1980; Webb, 1969). During this procedure, the toxic alkaloids subsequently released into the running stream, and responsible for the seed’s toxicity, are australine and castanospermine (Molyneus et al., 1988). Despite the presence of these toxins in the seed, the streams remained nontoxic, due to the effect of the toxin’s dilution in the large bodies of water used in this method, conducted in tropical areas where water bodies were filled by abundant rain.

Currently it is undetermined if the toxins from the three previously mentioned plants, and if the use of these plants, resulted in an environment that was unsuitable for cyanobacterial growth. This paper seeks to answer these questions. Common Australian species of Cyanobacteria were exposed to the water-soluble toxins of these plants. Some of these plants are known to have been used by indigenous people for more than 1,000 years (Specht and Specht, 1999). Therefore, this paper may isolate a small-scale water management strategy that was inadvertently or deliberately used by indigenous people to control cyanobacterial blooms. The reintroduction of such practices may be a useful initiative for park and forest managers in modern times.

Materials and Methods

Obtaining Materials

Plant species were collected from various growing locations in Australia. Acacia colei (NJSadgrove21) leaves and seeds were donated by Alice Springs Desert Park, courtesy of Gary Dinham and staff. Castanospermum australe (NJSadgrove22) seeds were collected upon ripening, to maximize toxicity (McKenzie et al., 1988), by the author from various growing locations within the town of Lismore NSW. Duboisia hopwoodii (NJSadgrove23) was collected from Uluru (Ayers Rock) in Uluru Kata-Tjuta National Park in the Northern Territory, under the guidance of staff and local indigenous people (AU-COM2007003). All herbarium voucher specimens were lodged with the N.C.W Beadle Herbarium (NE) at the University of New England, Armidale, NSW in Australia.

Water solutions were prepared from each of the three plants and sent to the Centre for Phytochemistry and Pharmacology (Southern Cross University, Lismore NSW) to be analyzed. Duboisia hopwoodii was analyzed for nicotine and nornicotine using GC-MS, and A. colei and C. australe were profiled using LC-MS.

Cyanobacterial species were purchased from CSIRO Microalgae Research in Hobart Tasmania. The toxic strain of Anabaena circinalis used in this project was collected by CSIRO at Palm Island (CS-539), which is offshore from the Queensland coast, approximately 65 km northwest of Townsville. The nontoxic strain Microcystis aeruginosa was collected from Wyangala Australia (CS-600/01), and the toxic strain Nodularia spumigena was collected from Lake Alexandrina South Australia (CS-590/01). A culture medium was prepared to maintain the three species for use during the experiments, within an MLA medium (Bolch and Blackburn, 1996) and subcultured once every 3 weeks. Subculturing was achieved by dispensing the cyanobacterial culture into a fresh MLA medium at a volume ratio of 1 to 10 respectively. All cultures were situated under the light of a Phillip’s Daylight bulb, at an intensity of 80-μmol photons PAR m−2s−1, at 20°C with 12:12 light and dark cycles (Bolch and Blackburn, 1996). Light intensity was determined by using a Biospherical Optics QSL-100 light meter. Cyanobacterial numbers were propagated >106 cells ml−1 for use during the experiments.

Initial Experiments

Initial experiments were conducted to investigate the effect of the water-soluble plants toxins on the growth of Cyanobacteria. The experiments tested the effect of the three plant species individually, on each of the cyanobacterial species used in this project. The following is a description of one of these experiments, which was then replicated for each combination in Table 1. To commence this experiment, a culture medium (MLA) was prepared (Bolch and Blackburn, 1996) aseptically and then administered in 20-ml volumes to 15 fully autoclaved transparent vials. Bacteria were subcultured into the vials, using 2 ml of a medium containing populations at a cell density >106 cells ml−1, bringing the total volume in the vials to 22 ml, and a cell density of approximately ≥105 cells ml−1.
Table 1

Experiments conducted to determine the effect of the water-soluble plants toxins on the growth of cyanobacteria populations


No. vials





A. circinalis

A. colei



C. australe



D. hopwoodii



M. aeruginosa

A. colei



C. australe



D. hopwoodii



N. spumigena

A. colei



C. australe



D. hopwoodii

During the experiments that involved the use of A. colei or D. hopwoodii, plant materials were administered to the solution in solid form, after drying and grinding into a powdery consistency. A. colei and D. hopwoodii were administered in the quantities specified in Table 2, with three replicates (samples) per treatment, with four treatments and a control, bringing the total to 15 vials for each experiment.
Table 2

Volume or quantity of treatment administered to 22-ml vials during each experiment outlined in Table 1, with 3 replicates per treatment

Treatment no.





1 (Control)

Volume (ml)

C. australe






Quantity (g)

A. colei






D. hopwoodii






During the experiments that involved the use of C. australe, water-soluble toxins were administered in liquid form. The following is a description of the preparation of this solution. Thirty grams of shredded C. australe seed was soaked in 300 ml of distilled water for 12 h. The solution was strained to remove bulk seed, separated into the 5 volumes also outlined in Table 2, and then administered to the vials. Each vial was raised using fully autoclaved distilled water to make all volumes equal.

All experiments were situated in the light environment as described earlier. Experiments remained until visible growth had occurred (approximately 7–10 days), at which point cell counts were conducted to determine the overall results, using an hemocytometer. Most experiments were conducted in replicates of three, with two averaged cell counts per replicate. Results were analyzed statistically using t tests on the three replicate averages (significance was established at the 0.05 level, N = 3).

Additional Experiments

The previously described initial experiments were later replicated with variations to the methodology to solicit results to represent a broad range of potential factors. These variations are:
  • Experiments were replicated using nontoxic alternative plant material (Eucalyptus citriodora, Mallotus philippensis, and Pennisetum clandestinum), to determine whether toxins were/were not responsible for the results from the earlier experiments.

  • Experiments were replicated in the MLA medium, diluted to half, quarter, and one-eighth of its concentration.

  • Treatments of A. colei, C. australe, and D. hopwoodii were administered to pre-grown populations that were >106 cells ml−1, and had visible population density.

Additional tests were conducted to determine the extent the results were influenced by variations in light penetration, carbonate, CO2, phosphate, bioavailable nitrogen content, and pH in the MLA—resulting from the addition of organic matter (plant material) to the cultures.

Results of the Phytochemical Analysis of Each Plant Water-extract

Phytochemical analysis confirmed that the A. colei leaf solution contained tannins, flavonoids, an abundance of saponins and unknown sesquiterpenes. The presence of Castanospermine was confirmed in the seeds from C. australe; however, it is likely that seeds will exhibit variation in toxicity (Han et al., 1997).

Two water extracts from the leaves of D. hopwoodii were analyzed. The first sample was produced immediately before analysis, which incidentally contained nornicotine as the main toxic alkaloid. The second extract was produced in the same manner as the first; however, before analysis 5 ml of it was dispensed into 22 ml of an MLA medium and placed alongside the experiments for a month. The purpose was to determine if the quantity of nornicotine was reduced during experiments. This was confirmed—this sample contained no nornicotine. All nornicotine in this sample had degraded with approximately 30% conversion into myosmine, which is a nontoxic alkaloid commonly found in human foods (Zwickenpflug and Tyroller, 2006) and is slightly similar to nornicotine.


Effect of the Water-Soluble Plant Toxins on the Growth of Cyanobacterial Species, Within the MLA Growth Medium

Analysis confirmed that nearly all results from the experiments were statistically significant. Thus, any data selected for presentation or discussion herein are scientifically and statistically valid. The data presented in Figs. 1 and 2 exhibit an implicit variation between the controls for each treatment. This occurred because cyanobacterial populations were counted at slightly separate times for each leaf treatment (A. colei, C. australe, or D. hopwoodii). Thus, treatments are more accurately compared to the experimental control that corresponds with the treatment, which is outlined in the figure legend.
Figure 1

Summary of average cell counts from three experiments conducted on N. spumigena in an MLA medium, testing various treatments of A. colei, C. australe, and D. hopwoodii. Inconsistent results from C. australe treatments occurred after experimental error. This trend did not reemerge.
Figure 2

Summary of the results from experimenting with N. spumigena with treatments of Eucalyptus citriodora, Mallotus philipensis, and Pennisetum clandestinum. No standard error was determined because replicates were minimized during this experiment.

The results from the initial experiments indicated that population growths were severely inhibited by treatments from phytotoxic plants; inhibition was more severe as the treatment concentration increased. However, these phytotoxic plant treatments solicited results that were near identical to those from nonphytotoxic plants (compare Figs. 1 and 2 for an example of this). Thus, it is likely that the population deaths and dormancy observed in the results were due to a factor other than plant phytotoxins. It is possible that this effect resulted from interferences with the MLA growth medium. All initial experiments exhibited this trend.

Despite this, comparison of results from lesser concentrated treatments also yielded a trend. Once it was known that E. citriodora contains citronella (a potent bactericide) this trend emerged (Holliday, 2005). If results obtained from E. citriodora are ignored (Table 3), then the nonphytotoxic plant treatments at 0.003 g can be observed to have a stable if not boosting effect on cyanobacterial population growths. Thus, it is likely that less concentrated treatments, such as 0.003 g, have little or no influence over the integrity of the MLA medium. Despite this, phytotoxic plant treatments at 0.003 g still solicited a negative effect on population growths, indicating that phytotoxins were potentially responsible.
Table 3

Percentage decreases compared with control with treatments of 0.003-g or 0.5-ml concentrations, showing a comparison of phytotoxic and nonphytotoxic treatments




A. colei

C. australe

D. hopwoodii

E. citriodora

M. phillipensis

P. clandestinum

A. circinalis







M. aeruginosa







N. spumigena







Table 3 presents percentage differences between bacteria populations without and with treatment at 0.003 g or 0.5 ml. This table is intended to highlight the differences between phytotoxic and nonphytotoxic plant treatments. The implication of this data is that phytotoxins were detrimental to cyanobacterial population growths during the experiments. Two possibilities arise from this: (1) Cyanobacteria were killed, or (2) cyanobacterial cells became dormant after treatments and resumed growth after toxin degradation. The latter was supported by the observation of cell chlorosis, which manifested in nearly all treatments, concomitantly with thylakoid membrane disorganization. This was implicit during experiments with M. aeruginosa.

Summary of the Results from the “Additional Experiments”

Synergistic Effect of Treatment in Lower Nutrient Environments

Dilution experiments were conducted using the smallest possible treatment concentration required to produce an effect, which was determined during the initial experiments. Thus, most experiments were subject to 0.003-g or 0.5-ml concentrations.

Dilution experiments were conducted on growing populations of Cyanobacteria. The most interesting results were obtained from A. circinalis. In the absence of treatment, A. circinalis populations exhibit greater proliferation when the MLA is diluted. Depicted in Fig. 3 is how this trend wanes between quarter and eighth dilutions. However, also depicted is how the populations are affected in nearly the opposite way, by treatments as little as 0.001 g from A. colei and D. hopwoodii. These small treatments seem to benefit populations at half dilutions and are likely to benefit at full strength MLA. However, treatments became highly detrimental to populations when the MLA was diluted further, i.e., at quarter and eighth dilutions. Thus, depicted in Fig. 3 is a trend indicating that A. circinalis populations are highly affected by treatments in lower nutrient environments, which ordinarily benefit them. It is possible that nutrients in the MLA are responsible for degradation of toxins from the treatments.
Figure 3

Anabaena circinalis populations growing in dilutions of the MLA with treatments as low as 0.001 g from A. colei and D. hopwoodii. Treatments from C. australe as low as 0.5 ml (see Materials and methods).

M. aeruginosa and N. spumigena from Figs. 4 and 5 respectively, exhibit a dissimilar trend to that from A. circinalis. Both species exhibited reduced proliferation in diluted MLA and both also reduced proliferation in diluted MLA after treatments. Trends exhibited in populations after treatment were somewhat linear and somewhat in proportion to trends in untreated MLA. Thus, there was no synergistic interaction between treatment and lower nutrient environments.
Figure 4

Microcystis aeruginosa population growths in diluted MLA with treatments as high as 0.015 g from A. colei and D. hopwoodii or as high as 2 ml (see Materials and methods) from C. australe.
Figure 5

Nodularia spumigena populations growths in diluted MLA with treatments as little as 0.003 g from A. colei and D. hopwoodii, or 0.5 ml from C. australe. At a later stage populations treated with C. australe attenuated to near zero.

Effect of Water-Soluble Plant Toxins on Severe Cyanobacterial Blooms (>106 cells ml−1)

Cyanobacterial populations were first propagated to an exceptionally high density (>106 cells ml−1) and then exposed to treatments (A. colei and D. hopwoodii at 0.06 g, and C. australe at 2 ml). The experiments were then placed in both light and darkness for approximately 1 week before cell counts were conducted. Figure 6 conveys the trend obtained from species of A. circinalis and N. spumigena after treatment, and then 1 week in darkness. These data indicate that treatments resulted in the death of Cyanobacteria. Similar results were obtained when the experiment was conducted in light, with the exception of prolific growth in populations without treatment. Populations of M. aeruginosa exhibited a different trend: they were not killed after treatments. This indicates that M. aeruginosa populations were inhibited but not destroyed by treatments, in contrast to the response for A. circinalis and N. spumigena populations.
Figure 6

Cell counts of high population density Anabaena circinalis and Nodularia spumigena before and after exposure to 0.06 g of A. colei and D. hopwoodii or 2 ml of C. australe.

Miscellaneous Tests and Measurements

Light inhibition was measured during experiments using a worst-case scenario approach. The results are most applicable to A. colei treatments and lesser to D. hopwoodii and lesser still to C. australe treatments. It was confirmed that at high treatments during the experiments, the effect of light inhibition from the subsequent bioflavanoids, and possibly tannins, was significant—possibly accounting for up to 40% of the observed population growth inhibition in the most severe cases. This was further supported by the observation of red pigmentation or phycoerythrin production in N. spumigena (Riethman et al., 1988), contingent with high treatment concentrations. However, light inhibition was not a significant factor at lower treatments, which are taken to be the most relevant to this paper.

When measurements were taken to isolate alterations to the MLA after treatment, it was revealed that pH was substantially decreased in higher treatments but remained relatively stable at lower treatment concentrations. An exception to this was from treatments of C. australe, which caused severe pH decline in even the smallest treatments: from pH 8.1 to 7.3. CO2 estimates, taken from alkalinity and pH measurements (Hargreaves and Brunson, 1996) from small (0.5 ml) C. australe treatments were nearly eight times higher than other treatments and the control. In addition, plant available phosphate increased by approximately 6% and nitrates decreased by approximately 70% with approximately 25% conversion into NO2 and NH3. Small D. hopwoodii treatments (0.003 g) caused phosphate to decline by approximately 57% and a 15% decline in nitrates with approximately 20% conversion into NO2 and NH3. Small A. colei treatments (0.003 g) yielded no such results.


Mechanisms of Cyanobacterial Inhibition: Death or Reproductive Dormancy

The susceptibility of Cyanobacteria to treatments from the three plants varied between experiments and species. Microcystis aeruginosa exhibited a substantially higher survival rate throughout all experiments. The subsequent cell characteristics of M. aeruginosa, after experimental treatments, strongly indicated that the inhibition of population growths resulted from cell dormancy. Cyanobacterial cell dormancy is an adaptive mechanism, commonly used by M. aeruginosa (Bar et al., 2002). It also is exploited by many species to survive hardships, such as the drying of a wetland (Latour et al., 2004), or other varying extremes of the environment, subsequently leading to a slowing of metabolic processes, and a halt to cell divisions (Bar and Robinson, 1999). The chlorosis of cyanobacterial cells, in combination with the slowing of cell metabolism and degradation of chlorophyll, with a subsequent improvement in nutrient availability, ensures the long-term survival of dormant cells (Dagnino et al., 2006).

“Nutrient repletion” is one of the main causes of cell dormancy, particularly nitrogen deficiency (Dagnino et al., 2006). However, it is unlikely nitrogen deficiency contributed to the cell dormancy observed in populations of M. aerguginosa. Although nitrates decreased during experiments, this was restricted to treatments from C. australe and much less so from D. hopwoodii. Cell chlorosis, thylakoid membrane disorganization, and cell dormancy were observed in all experiments. In addition, cell chlorosis was observed in A. circinalis and N. spumigena populations. If this resulted from a nitrate deficiency, the number of heterocysts in the filamentous bacteria would increase to solicit nitrogen fixation (Dagnino et al., 2006). This was not observed in any of the experiments, suggesting that nitrate deficiency was not a significant factor. The cause of cell chlorosis during experiments remains to be determined.

One of the most significant alterations to the MLA, caused by the addition of plant materials, was a decline in pH, which occurred proportionally to the amount of plant material added to the medium. The pH of the MLA is typically buffered between 7.8–8 using sodium hydrogen carbonate and hydrochloric acid (Bolch and Blackburn, 1996). It is likely that this buffering capacity was compromised when plant materials were added to the growth medium. However, when treatments were administered at low quantities, the buffering capacity of the MLA was not breached—pH was maintained in the optimal range. The lower treatment quantities are those most focused on in this paper. Thus, pH is not a significant factor when interpreting them.

When examining the results more closely, treatments from plants containing toxins inhibited cyanobacterial populations at very low quantities, within the normal range of pH. This trend opposes that presented by nontoxic plant treatments. Therefore, it is possible that in treatments from plants containing toxins, the toxins themselves were able to inhibit growth of, or kill Cyanobacteria.

Cyanobacterial Dormancy and Death via Invertase Inhibition

The composition and variety of the cell-wall enzyme group “invertase” within Cyanobacteria may help to explain the variation of pH tolerances exhibited between cyanobacterial species, and the occurrence of cell chlorosis during the experiments. Invertases are enzymes that hydrolyze sucrose into more usable hexoses, such as glucose or fructose (Vargas et al., 2003). Invertases play an essential role in the utilization of energy stores after photosynthesis, and invertase inhibition subsequently leads to a built-up of starches and sucrose, with a significant decline in cell metabolism (Aloni et al., 2001), gene expression and wound- or pathogen-induced respiration in the apoplast (Greiner et al., 1998). Thus, inhibition of apoplasmic (cell wall) invertases may prevent cell mitosis or other forms of cell wall rejuvenation or repair, subsequently resulting in cell dormancy and chlorosis or death (Sturm and Chrispeels, 1990).

Studies have shown that invertases in the leaves of Nicotiana glauca, closely related to D. hopwoodii, are more heavily regulated by nonprotein, alkaloid invertase inhibitors (Projo et al., 1998), than by the currently accepted apoplasmic invertase inhibitor Nt-CIF (Hothorn et al., 2003). Thus, if invertase inhibition did occur during the experiments, it is likely that these alkaloids played the largest role.

The above-mentioned study (Projo et al., 1998) of the toxic alkaloids nicotine, nornicotine, and anabasine—from N. glauca and also highly potent in D. hopwoodii—concluded that they were effective inhibitors of the acid-soluble invertase in that plant. It was observed that, of all compounds tested, those yielding invertase-inhibiting abilities were those containing a quaternary nitrogen atom. During the experiments the alkaloids released by D. hopwoodii leaves were the same as those found in N. glauca. Thus, toxic alkaloids containing quaternary nitrogen atoms were present during the experiments until detoxification occurred. The capacity for these alkaloids to penetrate the cell wall of Cyanobacteria and inhibit invertases remains to be determined. If this can occur, then it is possible that the deaths and dormancy of Cyanobacteria after treatments from D. hopwoodii can be explained.

During treatments of D. hopwoodii, the toxic alkaloid nornicotine degraded with approximately 30% converting into myosmine—a nontoxic alkaloid located in cereals and fruits (Zwickenpflug and Tyroller, 2006). Myosmine does not contain a quaternary nitrogen atom, so it is unlikely to have the same invertase-inhibiting abilities as the other alkaloids mentioned earlier. Castanospermine also lacks a quaternary nitrogen atom. If responsible for inhibition of Cyanobacteria, castanospermine is not likely to have followed the same pathway.

During C. australe treatments, the toxic alkaloid castanospermine was present, along with a substantially raised NH3 concentration. At low treatments NH3 was three to four times richer than the recorded concentrations during the 1991 Barwon-Darling River bloom (Bowling and Baker, 1996). It is unclear whether this is beneficial to the bacteria; however, it is unlikely that the observed inhibition was caused by an overdose of ammonia. If it was caused by castanospermine, then a suggested pathway would be from its glucosidase inhibiting ability, which is currently the focal point of most anti-HIV investigations (Roja and Heble, 2006).

Comparing Experimental Results with a Conceptualization of Traditional Applications

During the experiments, cyanobacterial populations were propagated to numbers that were approximately 50–7,000 times higher than the minimum required to threaten human health (Whitton and Potts, 2000). Most severe cyanobacterial blooms to manifest in Australia reach populations ranging from 105 to 106 cells ml−1 (Baker et al., 2002; Beltran and Neilan, 2000; Burford and O’Donohue, 2006; Donnelly et al., 1997; Kemp and John, 2005; Robson and Hamilton, 2004; Van Buynder et al., 2001), which is approximately 10–1,000 times lower than population numbers reached during the experiments. Severe cyanobacterial blooms in Australia have been exacerbated by environmental changes, resulting from the introduction of European land management influences (Baird and Cann, 2005; Howard and McGregor, 2000). Thus, during times of complete indigenous occupation of Australia, it is likely that cyanobacterial blooms were substantially less severe than those occurring today, and a great deal less severe than those produced during the experiments.

Cyanobacterial blooms occurring outside the laboratory are nourished from passive sources and the subsequent composition of nutrients varies from case to case. In contrast, the MLA medium is a more consistent and nourishing nutrient source (Bolch and Blackburn, 1996), which possibly makes it a greater growth environment than most in situ cases. Therefore, after treatments from plant phytotoxins, such as those in the experiments, the subsequent population changes would dramatically differ between the MLA medium and in situ specimens. It is likely that in situ specimens would have a lower tolerance to the treatments. This was confirmed following experiments using the MLA at various dilutions.

Four additional observations to support this are as follows: (1) during experiments the degradation of toxic alkaloids occurred at a much faster rate in the MLA than in low or no nutrient solutions. Cyanobacterial blooms typically manifest in environments that are less supportive of growth than the MLA medium. Such environments are usually less nourishing and contain a different composition of nutrients, which is evident from various case studies (Burford and O’Donohue, 2006; Donnelly et al., 1997; Kononen et al., 1996). Thus, the degradation of plant-sourced alkaloids is likely to occur at a much slower rate in the environment, giving them a greater residence time in the water and a longer lasting effect on Cyanobacteria; (2) cyanobacterial populations in environments outside the laboratory are much less dense and less nourished (Bottomley and White, 1951; Burford and O’Donohue, 2006; Donnelly et al., 1997; Kononen et al., 1996). Thus, stressed populations are expected to recover at a slower rate; (3) during the experiments A. colei and D. hopwoodii leaves were first dried then crushed, before deposition into the cyanobacterial cultures during treatments. A study on the leaves of D. hopwoodii concluded that toxic alkaloids in the leaves were reduced to below a third after drying (Bottomley and White, 1951). Thus, by drying the leaves before treatments, toxic alkaloids from D. hopwoodii leaves were substantially reduced. Consequently the toxicity of D. hopwoodii treatments (and potentially A. colei treatments) was lower than the expected toxicity of traditional applications—in traditional applications D. hopwoodii treatments were three times more potent; (4) during A. colei and D. hopwoodii treatments leaves were dried, very finely crushed and deposited into the cultures. In traditional applications, leaves were moist, hammered into a paste with a rock, and soaked at the water’s edge, with care not to release leaf matter into the water (Hiddins, 1999; Latz, 2004). Thus, during traditional applications, the subsequent rise of nutrients and dissolved inorganic carbon was minimized as opposed to treatments during the experiments. More specifically, during traditional applications of these plants, subsequent favorable conditions for growth were minimized in contrast to the experiments with subsequent favorable conditions maximized.

Water Resource Management Implications from the Results

One of the more known causes of eutrophication in small waterholes is the decline in water volume after evaporation, and the cease-flow of creeks and rivers with subsequent stratification (Dodson, 2005). In such cases, fish populations decline as available resources are utilized. Deceased fish decompose and contribute a substantially large amount of nutrient into the water. In combination with worsening conditions and the development of high phytoplankton population densities, fish continue to die, degrade, and contribute nutrients to the water, resulting in a positive feedback loop that leads to a hypertrophic water resource (ASLO, 1985; Davis and Koop, 2006; Paerl et al., 2006; Steinberg and Hartmann, 1988).

During pre-European Australia, cyanobacterial blooms may have evolved after the evaporation of a water-body containing an abundance of fish. However, in traditional times no food resources were left unutilized. As a water-body decreased in size, it became increasingly possible to obtain and use the resources necessary to poison the water-body and remove the fish for food. Therefore, the utilization of fish poisons and removal of biomass from shrinking waterholes may aid the preservation of water quality, with the added benefit of phytoplankton inhibition. The reintroduction of such practices could benefit today’s land managers. In addition, such practices may be used to remove feral fish species from small- to medium-sized wetlands, followed by the reintroduction of native fish species.

Additionally, small water bodies with substantial cyanobacterial blooms can be ameliorated using fresh leaves from D. hopwoodii or A. colei, via direct leaching into the water. Toxins degrade naturally and it is likely that filamentous cyanobacterial populations also will be destroyed. Indigenous people used these methods to procure food for thousands of years in an ecologically sustainable way and caused no harm to the environment, which supports the argument in favor of this method.

Worthwhile Areas for Further Research

Three main research ideas arose from this work. The first was to determine if the pattern of cyanobacterial inhibition occurring in vitro, can manifest in situ, and be measured following contemporary executions of the food procurement methods described here. One way to do this would be to locate remote indigenous Australian communities still practicing these methods today. A survey of the waterholes under this kind of treatment, compared with those not, would reveal whether Cyanobacteria is less frequent in those waterholes currently under use. Further surveys may help to reveal if this trend is the result of other factors, such as a traditional knowledge of waterholes that are less prone to be toxic and thus, less susceptible to eutrophication. To combat this problem may require that the traditional use of these plants be executed in eutrophic waterholes—if eutrophication can be ameliorated by these measures than it also must be prevented, which is the true focus of this paper.

A second research idea is to determine whether water-soluble plant toxins can be used to inhibit the growth or survival of mosquito larvae, particularly the Anopheles mosquito, which is the carrier of various human pathogens causing malaria, such as Plasmodium falciparum. This species causes malaria to manifest in its most deadly form and is carried and transmitted by Anopheles gambiae (Gouagna et al., 2004). If the mosquito larvae from these species can be inhibited or destroyed by an environmentally friendly and safe method, similar to those described in this paper and used traditionally by local indigenous people, then the threat of malaria can be reduced. Additionally, drastic land management practices, such as filling or removing swamps or water marshes, will not need to be undertaken.

A third research idea is to explore the influence of nicotinic plants on anatoxin-a–producing Cyanobacteria. Anatoxin-a is a potent cholinergic agonist acting on the nicotinic acetylcholine receptor (Molloy, 1995), in a similar way to nicotine or nornicotine; however, both anatoxin-a and its homologue homoanatoxin-a, have an LD50 range of 200–250 μg kg−1 intraperitoneal (i.p.) mouse (Araoz et al., 2005), which is approximately 24 times more potent than nicotine, as determined by i.p. mouse bioassays (Karaèonji, 2005).

Currently biologists are intrigued by the geographic segregation of anatoxin-a and paralytic shellfish poison (PSP) producing strains of A. circinalis (Beltran and Neilan, 2000). Australian strains of A. circinalis do not produce anatoxin-a, nor do any such strains outside of Africa, America, Asia, and Europe. A similar pattern exists in species of Anabaena flos-aquae—the species most known for anatoxin-a production (Beltran and Neilan, 2000; Wood et al., 2007). Until recently, anatoxin-a production was not known in any cyanobacterial species in the southern hemisphere. It has since been identified in the benthic species Aphanizomenon issatschenkoi in New Zealand (Wood et al., 2007). Additionally, until recently anatoxin-a production was not found on any continent that was once a part of Gondwanaland; however, there are now two that are known: one is a benthic species in New Zealand (Wood et al., 2007), and the other is a thermophile from a hot spring in Kenya (Krientz et al., 2003).

If we observed the world-wide distribution of nicotinic plants, we may notice that it is strongly antagonistic to anatoxin-a–producing strains of Cyanobacteria, if a variety of facts are taken into consideration. Of all 64 naturally occurring Nicotiana species, 56% are native to the South American continent, 12% are native to the North American continent, 29% are native to the Australian continent, and the remaining 3% consists of a species native to Africa and the South Pacific Islands (Kunio, 1998). Of these species, those with the highest concentration of alkaloids are those occurring in South America and Australia (Australis Botanicals, unknown; Smith and Abashian, 1963). Widespread species in North America were introduced from South America during the mid to late Holocene, after the discovery by the indigenous North Americans of a wide range of uses, including tobacco (Gately, 2001). Thus, the long-term natural distribution of nicotine-producing plants is opposite to the world-wide distribution of anatoxin-a producing cyanobacterial strains. It could be argued that plants containing nicotine are a characteristic of Gondwanaland.

All Nicotiana species in Australia were used by indigenous people (Robinson, 1980). However, the plant most frequently used for its nicotine content was D. hopwoodii (Bottomley and White, 1951; Latz, 2004; Robinson, 1980)—with an alkaloid content that is approximately eight times higher than the average Australian Nicotiana species (Australis Botanicals, unknown). Frequently nicotine would cascade into a body of water, carried in the ashes from a recent bushfire by rain, no doubt lit by indigenous people practicing “fire-stick farming,” which was extremely common (Bowman, 1998; Yibarbuk et al., 2001). In addition to this, the widespread distribution of Nicotiana species and frequency of their use by indigenous people, plus the use of D. hopwoodii during water resource management, has resulted in the occurrence of nicotinic agonist alkaloids in Australian waters for a potentially long period of time—currently unknown. Cyanobacterial populations may have been influenced by these cholinergic compounds during this period. Thus, it would be interesting to know whether the occurrence of these compounds resulted in a downward evolutionary gradient, which prevented the appearance of, or removed anatoxin-a–producing strains of Cyanobacteria from continents of Gondwanaland origin. Is this antagonistic world-wide distribution, observed between anatoxin-a Cyanobacteria and nicotinic plants, a coincidence or is it meaningful?


The cyanobacterial species chosen for this project consisted of the three most common genera to produce severe blooms in Australian waterways. During experiments these species were exposed to extracts from three of the most widely used food procurement plants in traditional water resource management. Plant treatments were administered to populations of Cyanobacteria within the MLA medium in as similar a manner to traditional applications as practicable. The results show that the use of these plants during water resource management could prevent cyanobacterial blooms from manifesting. The most likely mechanism of inhibition, following the use of D. hopwoodii leaves, is the binding of alkaloids to thylakoid membrane invertases to prevent cell wall rejuvenation, repair, and mitosis and subsequently resulting in the observed cell dormancy and death. Another potential mechanism of inhibition is glucosidase inhibition, caused by the toxic alkaloid castanospermine after treatments from C. australe.

During traditional applications of the plants relevant to this paper, it is likely that water bodies were exposed to a quantity of plant extract that was similar to the quantities used in the smaller plant treatments during experiments presented. However, traditional uses of the plants utilized fresher plant materials, meaning the alkaloid content was higher. This further supports the hypothesis that during traditional food procurement, toxic alkaloids played a more active role in influencing the microbiology of the water body than the other factors discussed in this paper, such as pH variation, light inhibition, or variations in nutrient content.

It can be suggested that the long-term use of indigenous food procurement techniques, in water resource management, influenced the past/current distribution of species of Cyanobacteria. During the experiments, the three species were affected in different ways. For example, the results showed that A. circinalis populations were boosted by very small D. hopwoodii treatments in the MLA medium at its normal concentration and also when diluted to half. Despite this, D. hopwoodii treatments remained detrimental to M. aeruginosa and N. spumigena. Thus, D. hopwoodii treatments in environments that were richer in nutrients may have favored A. circinalis in the long term, causing it to become more frequent than other species.

The results reported here may have management implications for park and forest managers and may convey a need to encourage an awakening of the traditional Australian culture, particularly among Aboriginal people. The use of fish and emu poisons as a water resource management tool may help to prevent or ameliorate eutrophication in a number of ways, particularly to inhibit cyanobacterial populations directly, or indirectly when fish are removed from the waterhole and the biomass is subsequently reduced. The encouragement of these cultural practices also may help to strengthen the identities of these people and subsequently enrich Australian tourism.

The indigenous food procurement techniques investigated in this paper were just a sample of the numerous techniques practiced by indigenous people across Australia. These practices were common in desert, temperate, and tropical regions. They potentially influenced the microbiology of water bodies ranging from small to large during a period of up to tens of thousands of years. Thus, the influence of indigenous food procurement techniques on Australian aquatic microbiology may have been significant and is thus a worthwhile endeavor for further research.

Copyright information

© International Association for Ecology and Health 2010

Authors and Affiliations

  1. 1.Centre for Bioactive Discovery in Health and AgeingUniversity of New EnglandArmidaleAustralia