View all publications

Appendix 1: Cyanotoxins and their distribution in New Zealand

Cyanotoxins are a diverse assemblage of natural toxins that have a very broad range of toxicity mechanisms, ranging from hepatotoxicity (toxic to the liver) and neurotoxicity (toxic to nerves or nerve tissue), to dermatotoxicity (affects the skin). Some cyanotoxins have also been shown to promote liver tumour growth when ingested in low doses over extended periods (Falconer and Humpage, 1996; Kuiper-Goodman et al, 1999). Based on their chemical structure, cyanotoxins can be divided into the following three groups: cyclic peptides (microcystins and nodularins), alkaloids (cylindrospermopsins, anatoxins and saxitoxins) and lipopolysaccharides (LPS).


Globally, microcystins are the most frequently found cyanotoxin (Chorus and Bartram, 1999). Microcystins are cyclic peptides, and more than 80 microcystin variants have been isolated and characterised (Spoof, 2004). Each variant differs with respect to the methyl groups and two amino acids within the ring (see Figure A1.1). This results in pronounced differences in toxicity among the variants. The amino acid ADDA (see Figure A1.1) is unique to microcystins and nodularins, and is required for their biological activity (Sivonen and Jones, 1999).

Figure A1.1: General structure of microcystins (X and Y are the common variable amino acids)

Figure A1.1: General structure of microcystins (X and Y are the common variable amino acids)

Skeletal formula for the chemical structure of microcystins, showing the part of the structure that is the amino acid ADDA.

Microcystins are hepatotoxins that block protein phosphatase 1 and 2a in affected organisms (MacKintosh et al, 1990). This binding is inhibitory, highly specific and irreversible. The chief pathway into cells for microcystins is the bile acid carrier, which is found in liver cells and, to a lesser extent, in intestinal epithelia (Falconer, 1983; Runnegar et al, 1993). Most microcystins are highly toxic, with intra-peritoneal (ip) mouse toxicities ranging between 50 and 300 µg/kg body weight (mouse) (Sivonen and Jones, 1999). In vertebrates, a lethal dose of microcystin causes death by liver necrosis (premature death of cells) within hours to a few days. In addition, Fitzgeorge et al, (1994) published evidence for disruption of nasal tissues by the common hydrophilic variant microcystin-LR. Toxicity by oral uptake is generally at least an order of magnitude lower than toxicity by intra-peritoneal injection. However, intra-nasal application in these experiments was as toxic as intra-peritoneal injection, and membrane damage by microcystin enhanced the toxicity of anatoxin-a. This uptake route may be relevant for water sports such as waterskiing that lead to inhalation of spray and droplets.

Fitzgeorge et al, (1994) demonstrated that microcystin toxicity is cumulative. A single oral dose showed no increase in liver weight (a measure of liver damage), whereas the same dose applied daily over seven days caused an increase in liver weight of 84 per cent and thus had the same effect as a single oral dose 16 times as large. This may be explained by the irreversible covalent bond between microcystin and the protein phosphatases, which leads to subsequent damage to cell structure (Goldberg et al, 1995; Maynes et al, 2006). Sub-acute liver injury is likely to go unnoticed for two reasons:

  • liver injury shows externally noticeable symptoms only when it is severe
  • acute dose-response curves for microcystins are steep, so little acute liver damage may be observed up to levels close to severe acute toxicity.

The two potential mechanisms for chronic microcystin damage to the liver are progressive active liver injury (as described above) and the promotion of tumour growth. The tumour-promoting activity of microcystins is well documented in animals, although microcystins alone have not been shown to be carcinogenic. Epidemiological evidence from China (Ueno et al, 1996) has linked the continual consumption of low doses of microcystins in drinking water to primary liver cancer. The International Agency for Research on Cancer (IARC) has recently classified microcystin-LR as a group 2B carcinogen (Grosse et al, 2006). Group 2B compounds are considered possible carcinogens to humans.

Numerous incidents of animal and human poisonings have been attributed to microcystins (Ressom et al, 1994). One of the most severe cases occurred in Brazil in 1996, when processes at a water treatment plant failed and manual addition of chlorine to tanker loads of water supplying a hospital was insufficient to remove microcystins. This resulted in over 50 fatalities at the dialysis treatment clinic (Azevedo et al, 2002). In a further case from Brazil, the death of 88 people, mostly children, was associated with drinking water from a newly constructed reservoir, which contained a bloom of Microcystis spp. (Teixeira et al, 1993).

Microcystins in New Zealand

Microcystins are the most commonly found cyanotoxin in planktonic cyanobacteria in New Zealand. They have been identified in over 60 water bodies (Christoffersen and Burns, 2000; Wilding, 2000; Hamill, 2001; Stirling and Quilliam, 2001; Wood, Briggs et al, 2006; Wood, Stirling et al, 2006). The highest concentrations recorded (36.5 mg/L) were found in cyanobacterial scum material from Lake Horowhenua (Wood, Stirling et al, 2006).

Of New Zealand planktonic species, Microcystis spp. are the only confirmed microcystin producers. The production of microcystins by Anabaena spp. and Planktothrix spp. is suspected due to the detection of microcystins in environmental samples dominated by these species and/or via the detection of genus-specific genes involved in microcystin production from environmental samples (Christoffersen and Burns, 2000; Wood, Stirling et al, 2006; Wood, unpublished data).

Analysis of environmental samples indicated that microcystins are also produced by benthic species (Hamill, 2001; Wood, Stirling et al, 2006). The production of microcystins by a benthic species (Planktothrix sp.) has recently been confirmed (Wood, Heath, McGregor et al, in press). In March 2003 the eastern shore of Lake Taupo was lined with thick gelatinous mats of Nostoc commune (Appendix 7) that contained high levels of microcystins (708 mg/kg). Gelatinous colonies accumulated along the shoreline following a storm event in which waves had dislodged the Nostoc commune from rocks. A water sample collected close to the shoreline at Lake Taupo also contained microcystins and demonstrated that some of the microcystins were being released from these mats back into the water body (Wood, Stirling et al, 2006).


Nodularins have a very similar structure to microcystins and are produced by Nodulariaspumigena, which is primarily a brackish-water species. Nodularin is also a potent inhibitor of protein phosphatases 1 and 2a (Honkanen et al, 1991; Maynes et al, 2006). Nodularin has an ip LD50 of 60 µg/kg body weight (mouse) (Carmichael et al, 1988).

Nodularins in New Zealand

Nodularin has been identified from blooms of Nodularia spumigena in Lake Ellesmere (Carmichael et al, 1988) and Lake Forsyth. There is long history of stock deaths around both of these lakes (Connor, 1977). Nodulariaspumigena is known to occur in other brackish lakes around New Zealand (Etheredge and Pridmore, 1987), although the occurrence of nodularin has not been investigated.


Cylindrospermopsin causes extensive damage to the liver and kidney and is a potent inhibitor of protein synthesis (Terao et al, 1994; Falconer et al, 1999; Froscio et al, 2003). Clinical symptoms may appear several days after exposure, so it is often difficult to determine a cause–effect relationship. Falconer and Humpage (2001) suggest that cylindrospermopsin may also act directly as a tumour initiator, which has implications for both short- and long-term exposure. Crude extracts of Cylindrospermopsis raciborskii (a common cylindrospermopsin producer) injected or given orally to mice also induce pathological symptoms in the kidneys, spleen, thymus and heart (Seawright et al, 2000). Two variants of cylindrospermopsin exist (see Figure A1.2): 7-epicylindrospermopsin, a toxic minor metabolite of the cyanobacterium Aphanizomenon ovalisporum (Banker et al, 2000), and deoxy-cylindrospermopsin (Norris et al, 1999), which is now thought to be as toxic as cylindrospermopsin. Pure CYN and 7-epi-CYN exhibits an LD50 of 2.0 mg/kg body weight (mouse) after 24 hours, but 0.2 mg/kg body weight after five days (Ohtani et al, 1992).

Figure A1.2: General structure of cylindrospermopsin

Figure A1.2: General structure of cylindrospermopsin

Skeletal formula for the chemical structure of cylindrospermopsin, showing variation between 7-epicylindrospermopsin and deoxy-cylindrospermopsin at C7.

Cylindrospermopsis raciborskii was implicated in one of the most significant cases of human poisoning from exposure to a cyanobacterial toxin. In 1979, 148 people required hospitalisation with symptoms of gastro-enteritis after a local water supply on Palm Island (Australia) was dosed with copper sulphate to control a dense algal bloom (Byth 1980; Bourke et al, 1983). The copper sulphate caused the cells to break apart and resulted in the release of cyanotoxins into the water supply (Hawkins et al, 1985). Recent cattle deaths in Queensland (Australia) have been attributed to cylindrospermopsin (Saker et al, 1999).

Cylindrospermopsins in New Zealand

Cylindrospermopsin was first identified in New Zealand in Lake Waitawa (Otaki) in 1999, although the species responsible for its production was not confirmed (Stirling and Quilliam, 2001). Cylindrospermopsis raciborskii was identified for the first time in a bloom in Lake Waahi (in March 2003), during which liquid chromatography-mass spectrometry (LC-MS) confirmed the presence of cylindrospermopsin and deoxy-cylindrospermopsin (Wood and Stirling, 2003). These two incidents remain the only detections of cylindrospermopsin in New Zealand. Multiple samples with high concentrations of C. raciborskii have been analysed, but with no cylindrospermopsin detected (Wood, unpublished data), indicating that not all strains of C. raciborskii in New Zealand produce this toxin.

Anatoxin-a and homoanatoxin-a

Anatoxin-a and homoanatoxin-a are neurotoxic poisons. They are powerful depolarising neuromuscular blocking agents acting through the nicotinic acetylcholine receptor (Carmichael et al, 1979). Because of their small size they are rapidly absorbed when ingested orally. In affected animals these toxins can cause convulsions, coma, rigors, cyanosis, limb twitching, hypersalivation and/or death. Anatoxin-a is often linked with animal and wildfowl poisonings (Ressom et al, 1994), but there have been no reported human fatalities from anatoxin-a. Anatoxin-a and homoanatoxin-a have ip LD50 of 200−250 µg/kg body weight (mouse) (Devlin et al, 1977; Skulberg et al, 1992).

Anatoxin-a and homoanatoxin-a in New Zealand

Following the rapid deaths of dogs near the Waikanae River (Lower North Island) in 1998, the toxicity of a benthic mat of Oscillatoria sp. mat was investigated using a mouse bioassay and high-performance liquid chromatography with fluorescence detection (HPLC-FLD). The presence of natural degradation products of anatoxin-a was confirmed (Hamill, 2001). Further sudden deaths of dogs were reported at the Mataura River (Lower South Island) in 1999 and 2000. Benthic Oscillatoria-likesp. mats were collected and their toxicity confirmed (Hamill, 2001). Wood, Stirling et al (2006) identified anatoxin-a in three planktonic samples collected from Lake Rotoehu (Rotorua), Lake Henley (Masterton) and Lower Karori Reservoir (Wellington); all three samples were dominated by Anabaena spp.

Aphanizomenon issatschenkoi was identified for the first time in New Zealand in 2003, and LC-MS analysis of a strain isolated from Lake Hakanoa (Waikato) confirmed it was producing anatoxin-a (Wood, Rasmussen et al, 2007). Interestingly, despite the absence of cylindrospermopsin production, genes implicated in the biosynthesis of cylindrospermopsin were successfully amplified from the Aph. issatschenkoi strain. Aph. issatschenkoi is now a common bloom-forming species in the Waikato region, and blooms have been reported elsewhere in the North Island (Wood, unpublished data).

In November 2005 at least five dogs died rapidly after contact with water from the Hutt River (Wellington). Extensive mats of benthic material were present in the river at the time of the poisonings. Subsequent LC-MS analysis identified anatoxin-a, homoanatoxin-a and their degradation products, dihydro-anatoxin-a and dihydro-homoanatoxin-a (Wood, Selwood et al, 2007). The causative species was identified as Phormidium autumnale (Wood, Selwood et al, 2007). Since this incident, mats of P. autumnale have commonly been linked to dog poisoning events in other parts of New Zealand (eg, Canterbury, Bay of Plenty) and anatoxin-a and homoanatoxin-a have been detected on multiple occasions (Wood, unpublished data). Examination of stomach contents from dead dogs has revealed copious amounts of ‘algal’ material, suggesting the dogs had ingested cyanobacterial material rather than being exposed directly to toxins that are free in the water column (Wood, Selwood et al, 2007). It is unknown whether dogs are more susceptible to anatoxin poisoning than other organisms.


Anatoxin-a(S) is structurally different and up to 10 times more potent (towards mice) than anatoxin-a. It is a cholinesterase inhibitor that induces hypersalivation, diarrhoea, shaking and nasal mucus discharge in mammals (Carmichael, 1992; Mahmood and Carmichael, 1987). It is thought to be produced only by Anabaena lemmermannii (Henriksen et al, 1997) and A. flos-aquae (Mahmood and Carmichael, 1987). Anatoxin-a(S) has an ip LD50 of 20 µg/kg body weight (mouse) (Carmichael, 1992). Anatoxin-a(S) has not been detected in New Zealand.


Saxitoxins are fast-acting neurotoxins that inhibit nerve conduction by blocking sodium channels (Adelman et al, 1982). Saxitoxins are also produced by various marine dinoflagellates under the name of paralytic shellfish poisons (PSPs), and the human health effects caused by saxitoxins are well described from numerous reports of human toxicity associated with the consumption of shellfish containing relatively high concentrations of PSPs. No PSP-like illnesses have been reported in humans from the consumption of drinking water or contact with recreational water containing saxitoxins (Chorus and Bartram, 1999). More than 30 saxitoxin variants have been isolated and characterised. Saxitoxin has an ip LD50 of 10 µg/kg body weight (mouse). Other analogues are mostly less toxic than saxitoxin.

Saxitoxins have caused sheep mortalities in Australia (Negri et al, 1995) and were identified in an extensive bloom of A. circinalis in 1990 on the Murray Darling River (Australia), which resulted in the death of over 1600 stock (Bowling and Baker, 1996).

Saxitoxins in New Zealand

A cyanobacterial bloom (predominantly Anabaena planktonica) in the Waikato River in 2003 caused taste and odour problems in the drinking water supplied to the city of Hamilton and other towns along the length of the Waikato River. Saxitoxins were detected (via ELISA and neuroblastoma assay) in water samples taken from the water treatment intake and throughout the water treatment process (Kouzminov et al, 2007), but levels were well below the provisional maximum acceptable values set out in the Drinking-water Standards for New Zealand 2005 (Ministry of Health, 2005b). The saxitoxin-producing organism/s were not identified.

Using an ELISA and neuroblastoma assay, Wood, Stirling et al (2006) detected low levels of saxitoxins in 38 different water bodies. Although only low levels of saxitoxins were detected, the results imply that saxitoxins may be more prevalent in New Zealand water bodies than previously assumed. Further investigation and chemical analysis are required to confirm which species are responsible for saxitoxin production and to determine the variants produced.


Lipopolysaccharides (LPS) are an integral component of the cell wall of all gram-negative bacteria, including cyanobacteria. Found in the outer cell membrane, LPS form complexes with proteins and phospholipids (Chorus and Bartram, 1999). Lipopolysaccharides can elicit irritant and allergenic responses in humans and animals tissue (Torokne et al, 2001; Pilotto et al, 1997). Cyanobacterial LPS are considerably less potent than LPS from pathogenic gram-negative bacteria such as Salmonella (Keleti and Sykora, 1982). Although comparatively poorly studied, LPS from cyanobacteria have been implicated in human health problems associated with exposure to cyanobacteria (Ressom et al, 1994). Recent studies (eg, Pilotto et al, 2004) showed no correlation between dermatological reactions experienced by exposed individuals and the presence/absence of other cyanotoxins in the samples tested.

Lipopolysaccharides in New Zealand

Lipopolysaccharides almost certainly occur in New Zealand cyanobacteria, although no attempts to detect or quantify them have been made. There have been numerous reports of humans experiencing skin rashes after contact with water containing cyanobacteria, particularly from Waikato and Rotorua lakes (Wilding, 2000; M Bloxham, Environment Bay of Plenty, Whakatane, personal communication; D Hood, Waikato District Health, Hamilton, personal communication).

β-N-methylamino-l-alanine (BMAA)

β-N-methylamino-l-alanine (BMAA) is a non-protein amino acid produced by cyanobacteria. BMAA is considered a possible cause of the amyotrophic lateral sclerosis/Parkinsonism–dementia complex that has an extremely high rate of incidence among the Chamorro people of Guam. It has been suggested that BMAA biomagnifies through the food web. In the Chamorro case, a root symbiont of the genus Nostoc is found on cycad trees. The Chamorro eat fruit bats, which feed on cycad seeds (all of which contain BMAA; Cox and Sacks, 2002). There is current debate on the occurrence of BMAA in other cyanobacterial genera (eg, Cox et al, 2005; Rumsby et al, 2008). Current international research should provide further information on sources of BMAA and guidance on acceptable levels.

Cyanotoxin and toxicity testing

For health assessments in recreational use water bodies it is recommended that total (ie, combined intracellular and intracellular) toxin content and/or toxicity concentrations in samples are measured.

A range of methods has been developed to detect and identify cyanotoxins and their toxicity (see Lawton et al, 1999 for an in-depth review). In New Zealand, five methods are currently commercially available for cyanotoxin or toxicity analysis (Appendix 8).

Enzyme-linked immuno sorbent assay (ELISA)

An ELISA is available for detecting total ADDA containing microcystins and nodularins (Fischer et al, 2001). It uses antibodies raised against ADDA (an amino acid unique to microcystins/nodularin; see Figure A1.1) and should detect over 80 per cent of all known microcystin variants and nodularin. ‘Free’ ADDA may also be detected in some instances, potentially overestimating total microcystin load in a sample. This method cannot distinguish between microcystins and nodularin. However, nodularin is only produced by Nodularia spumigena, a brackish water species, and so this is unlikely to be problematic. The ELISA offers a rapid and cost-effective method of determining the total microcystin/nodularin content of a sample. A ‘dip-stick’ method based on the same technology is available commercially ( This enables users to rapidly screen samples for microcystins or nodularin and requires no additional equipment. An ELISA kit is also available through ABRAXIS for cylindrospermopsin (

Liquid chromatography mass spectrometry (LC-MS)

Liquid chromatography mass spectrometry methods are available for microcystins, nodularin, anatoxin-a, homoanatoxin-a and cylindrospermopsins in New Zealand. This method detects the specific mass of individual toxins in a sample and thus provides information on which variants are present. This is particularly relevant for microcystins, where over 80 variants exist, each varying in its toxicity. Routine LC-MS screens may miss unusual microcystin variants and therefore underestimate the total microcystin load. Recent research in New Zealand has demonstrated a high correlation between the ADDA-ELISA and LC-MS for microcystin detection (Mountfort et al, 2004; Wood, Mountfort et al, 2008).


This is an ELISA-based technique that detects all saxitoxin variants to varying degrees ( The test determines the presence or absence of saxitoxins. It is not truly quantitative, nor does it provide information on which variants are present.

Acetylcholinesterase assay

The biochemical activity of anatoxin-a(S) can be exploited in an enzyme-based assay to detect the inhibition of acetylcholinesterase (AChE), thereby providing an indication of the presence of this toxin. Although this assay is available, it has not yet been validated for the detection of anatoxin-a(S) in New Zealand.

Toxicity tests

In the strict sense, toxicity refers only to animal-testing data and is expressed as the amount of cyanobacteria lethal to an animal (usually normalised per kilogram of body weight). Although not routinely used in New Zealand, mouse bioassays are available. A mouse bioassay may be used when animal or human poisonings indicate the presence of toxic substances but results for known toxins are negative. However, a positive mouse test does not definitively demonstrate that the cyanobacteria or toxin being tested is also acutely toxic to humans, although a positive result does provide strong evidence that an active toxin is present.