Nuts, oilseeds, and dried legumes

9 Nuts, oilseeds, and dried legumes

I Introduction

A Definitions Nuts are dry, one seeded fruit, which do not dehisce at maturity, and are usually enclosed by a rigid outer

casing or shell. Most nuts grow on large shrubs or trees and are known as tree nuts. Tree nuts include almonds (Prunus amygdalus), hazelnuts (Corylus avellana), pistachios (Pistachia vera), Brazil nuts (Bertholletia excelsa), pecans (Carya illinoensis), coconuts (Cocos mucifera), and macadamia nuts (Macadamia ternifolia). Although not covered strictly by the botanical definition, walnuts (Juglans regia) are usually considered to be nuts. The only major nut not from trees is the peanut (Arachis hypogaea), known as “groundnut” in some countries. Botanically, the peanut is a legume, a member of the pea family, but it will be treated as a nut in this chapter. Illipe nuts (Shorea aptera and related species) contain 50–70% fat and are exported from Southeast Asia as a cocoa butter substitute.

Peanut butter is the finely ground paste produced from (usually) roasted peanuts. In some countries other edible oils or ingredients are added to peanut butter, but with little effect on microbiology. Peanut sauces, known as satay sauces, are pastes made from dried shelled peanuts, spices and water, usually with other ingredients added. They are widely used in Asian countries, mostly consumed fresh, and are increasingly traded internationally, where they are normally subjected to sterilizing treatments as low-acid products.

Oilseeds are mostly small seeded crops that are grown primarily for oil production. They are drawn from a range of botanical families. Oilseeds include palm nuts (Elaeis guineensis, Ela. oleifera, and hybrids), rapeseed or canola (Brassica rapa, B. campestris), sesame (Sesamum indicum), sunflowers (Helianthus annuus), safflower (Carthamus tinctorius), cottonseed (Gossypium spp.), and cacao seeds (Theobroma cacao). The last named is treated in Chapter 10, along with cocoa butter. Oil from Zea mays,

known as maize oil or corn oil, is also an important commodity. Dried coconut is known as copra when sold in large pieces and as desiccated coconut when shredded. Comminuted coconut is known as coconut cream and is usually sold canned as a low-acid food. It is sometimes spray dried and sold as a powder.

Dried legumes are the seeds of leguminous plants, members of the family Leguminosae. Dried legumes that will be considered in this chapter include soybeans and the many other types of beans that are field dried. Fresh legumes are treated under vegetables in Chapter 5.

Coffee is a beverage made by brewing roasted beans of the coffee tree (Cafea arabica, C. canephora var. robusta, or hybrids). Coffee is consumed almost universally and the coffee growing, and manufac- turing industries are amongst the largest in international trade. Instant coffee is produced by freeze or spray drying brewed coffee.

B Important properties Nuts have very high nutritional and calorific values. The pH of all nuts and oilseeds is near neutral, in

theory rendering them susceptible to growth of all kinds of microorganisms during development and before natural drying at maturity. In practice, shells provide a highly effective barrier to the entry of bacteria during nut growth. After natural drying, the low a w of most nuts restricts bacterial spoilage or toxin production. However, contamination of nuts may sometimes occur post-harvest, for example,

441 with salmonellae, leading to concern both directly and with high a w products to which nuts are added,

NUTS, OILSEEDS, AND DRIED LEGUMES

e.g. dairy products. In the microbiological context, the most important property of both nuts and oilseeds is their high oil content. This provides a high susceptibility to attack by lipolytic bacteria and by spoilage fungi, with an exceptional potential for mycotoxin production.

Peanuts, with a unique growth habit in soil, are especially vulnerable to fungal invasion before harvest. Many kinds of fungi are found in peanuts, but the presence of Aspergillus flavus and the production of aflatoxins are of major concern.

Spoilage and mycotoxigenic fungi sometimes invade other nuts as well, but usually only as the result of insect or mechanical damage (tree nuts) or contamination during drying and processing (pistachios). Most dried legumes are rich in carbohydrates and low in oils, so they are microbiologically similar to cereals. However, soybeans are up to 20% oil on a moisture free basis (Waggle and Kolar, 1979), so they more resemble oilseeds in their microbiology. Soybeans also contain high levels of protein (40% or more, Waggle and Kolar, 1979), an attractive trait because of the high nutritional quality (Richert and Kolar, 1987).

Fresh coffee beans, known as cherries, are of relatively low nutritional value. The flesh of the fruit must be removed before the beans are fully dried, and this is accomplished by mechanical means or fermentation.

C Methods of processing Typical primary processing of all of the commodities considered here involves drying, which commonly

takes place in the field. Mechanical drying may be used for some crops in particular localities. Tree nuts. Tree nuts are almost always allowed to dry in situ, then harvested mechanically or by hand.

Some types are allowed to fall to the ground before collection, resulting in surface contamination by bacteria and fungi. Sometimes dehydration is used in adverse climates or seasons to complete the drying process.

Coconuts are harvested from tall trees either green (for consumption fresh) or at maturity. Nuts are pierced or broken, drained of water, the kernels cut into slices and sun dried to produce copra for oil production. A variety of other coconut products are made using both traditional and modern technologies (Hagenmaier, 1980).

Peanuts. Peanuts are harvested by pulling from the ground while high in moisture content and then drying. In developed countries pulling is mechanical. Pulling requires that the stems (pegs) supporting the nuts must be strong and not yet senescent, so the nuts are usually at least partially dried on the upturned bushes to weaken the pegs before mechanical threshing to remove the nuts from the bushes. Newer types of threshers permit mechanical removal of nuts immediately after pulling, followed by mechanical drying. In damp growing areas or seasons nuts are often mechanically dried after threshing

to ensure microbiological stability. Peanuts must be dried to about 8% moisture (below 0.70 a w ) to be microbiologically stable.

The field drying process is usually quite slow: in Australia field drying takes 6–10 days to complete even under good conditions (Pitt, 1989). In developing countries, harvesting is commonly carried out by hand. Pulling is aided by the use of hoes or forks, then the nuts are usually removed from the bushes by hand and dried in the sun on hessian or plastic sheets. If weather conditions are good, drying may require no more than 2–3 days. Much longer periods may be needed if conditions are adverse.

MICROORGANISMS IN FOODS 6

Oilseeds. All types of oilseeds are usually field dried. Dried legumes. Generally field drying is used, although in some localities finishing may be by me-

chanical drying in bins. Coffee. Coffee cherries are usually picked by hand, or mechanically on large farms, and are dried in

the sun or less commonly mechanically. The beans may be dried directly, and separated from the hull afterwards, or mechanically dehulled and dried, or dehulled by fermentation.

D Types of final products Large quantities of nuts are sold as snack foods, often without further processing or after shell removal.

Many nuts are also processed by roasting or frying, and again mostly consumed as snack foods. A proportion of nut production is incorporated into confectionery: chocolates, toffee bars, and muesli bars are three common types of products.

A major end use of peanuts is in peanut butter and related products. Dried peanuts are roasted and ground, usually with the addition of salt and sometimes other oils. In some Southeast Asian countries, peanuts are roasted and ground with added water to produce satay sauces, which may be heat processed as low-acid foods, especially for export. In some peanut producing countries, peanuts are boiled in salty water and consumed as a snack food: this product has a short shelf life. Sometimes boiled peanuts are subsequently roasted in the shell and then vacuum packed to produce a shelf-stable product.

Dried coconut may be sold in large pieces as copra, which is mainly used for oil production, or may

be shredded and sold as desiccated coconut for manufacturing or domestic use. Other products from coconut in international trade include canned coconut milk or cream, processed as a low-acid food in cans or pouches; spray-dried coconut cream, a powder made up of oil globules encapsulated in a protein film; desiccated pressed coconut, a by-product of coconut cream manufacture; coconut flour, made from fresh coconut or copra; and coconut oil (Hagenmaier, 1980).

Palm sugars and an alcoholic drink made from the sap of fruiting spath of coconut are popular in Southeast Asia and India. A variety of other coconut products, often made from young coconuts, are sold within the same regions. In Indonesia, a fermented coconut cake called “tempeh bongkrek” is made from coconut or coconut plus soybean fermented with Rhizopus oligosporus (Ko et al., 1979).

Oilseeds are normally further processed to provide oil. Residual meals are commonly used as animal feed ingredients. Dried legumes are usually consumed after boiling to soften them. They may also be used as soup ingredients, either domestically or commercially. Soybeans have become a very large commercial crop in the United States and Brazil, originally due to demand for soybean oil, a cheap commodity because soybean meal is a high protein, high-value feed ingredient. Since the 1950s, soybeans have been processed into food-grade products including soy flour, soy protein concentrate and isolated soy protein (Waggle and Kolar, 1979). Isolated soy protein, more than 90% pure, with high availability and digestibility, and with an excellent nutritional balance, has become important as a relatively low cost food-grade protein source (Kolar et al., 1985; Richert and Kolar, 1987).

Coffee beans are stored after drying (as “green” coffee) then graded and shipped to manufacturers. Beans are then roasted to varying degrees depending on type and extent of flavor desired. They may

be packed and distributed, usually under inert atmospheres, as beans or ground and extracted with hot water, then dried to make instant coffee.

NUTS, OILSEEDS, AND DRIED LEGUMES

Table 9.1 Fungi commonly invading peanuts before or immediately post harvest

After 2–13 days drying: Species

At harvest: average

average infection (%) Aspergillus flavus

infection (%)

Lasiodiplodia theobromae

Fusarium spp.

Macrophomina phaseolina

Penicillium funiculosum

Rhizoctonia solani

Rhizopus spp.

From McDonald (1970).

II Initial microflora

A Nuts Only a few systematic studies have been carried out on the microflora of nuts in the field. Little or

no information exists on the microflora of fresh tree nuts. However, the presence of shells provides

a very strong protective barrier against infection by both fungi and bacteria. Apart from possible systemic fungal infections from the trees themselves, tree nuts are essentially sterile before harvest. Studies on dried nuts have provided no definitive information about the time of infection of tree nuts by fungi.

The mycoflora of fresh peanuts has been surveyed only rarely. Data from one such study in Nigeria are given in Table 9.1 (McDonald, 1970). Apart from Asp. flavus, which was present at low levels, the fungi found were field fungi, with Fusarium species, Lasiodiplodia theobromae and Macrophomina phaseolina dominant.

B Oilseeds Little information has been published on the pre-harvest microflora of oilseeds. It is to be expected

that the bacterial flora will be similar to cereals (Chapter 8), and that some specific field fungi will be present, together with the types commonly found on cereals.

C Legumes Fresh legumes are readily contaminated externally on the pod with bacteria and fungi. However, the pod

provides some protection from entry of microorganisms during the growth phase. Important diseases of fresh legumes include black blight due to Alternaria alternata, anthracnoses due to Colletotrichum species, dry or soft pod rots due to Ascochyta species, and soft rots due to Pseudomonas and Xanthomonas species (Snowdon, 1991).

D Coffee Few studies have been carried out on the microflora of fresh coffee cherries. While cherries are fresh

and intact few microorganisms are present except as contaminants. Aureobasidium pullulans, Fusarium stilboides, and Penicillium brevicompactum are the most common mould species and Candida edax and Cryptococcus album the most common yeasts associated with fresh coffee cherries (Frank, 2001). If cherries dry on the tree, to produce what is known as “boia”, field fungi such as Alternaria and Cladosporium are commonly present (M.H. Taniwaki, unpublished).

MICROORGANISMS IN FOODS 6

III Primary processing

A Effects of processing on microorganisms Primary processing of nuts, oilseeds, and legumes usually involves drying under natural conditions.

Effective sun drying may reduce the initial microflora. However, if drying occurs under adverse condi- tions, mycoflora may increase both in kind and in numbers, and mycotoxin production may be initiated. The bacterial flora of these raw commodities reflects the environment in which they are grown and har- vested. The numbers and types of bacteria present result from contamination from soil and equipment and to other environmental factors.

Tree nuts Pistachios. Pistachios were contaminated with up to 14 Aspergillus species (Doster and

Michailides, 1994). The majority of the isolates came from nuts which had split or been damaged by insects in the orchard. The most commonly encountered species was Aspergillus niger, present in 30% of such kernels. Aspergillus flavus and Asp. parasiticus were also found, with the potential to form aflatoxins, and in addition Asp. ochraceus and Asp. melleus, potential producers of ochratoxin A (Doster and Michailides, 1994). Mould counts on 143 samples of freshly harvested pistachios in Turkey

ranged from 10 3 –10 4 per gram and after storage 10 5 –10 6 per gram (Heperkan et al., 1994). A significant proportion of kernels (6–16%) was invaded by Asp. flavus. Pistachio nuts that split to expose the kernels in the orchard are prone to infection by Asp. flavus and hence are likely to contain aflatoxins. The time pistachio nuts split depends largely on cultivar type. Aflatoxin contamination is also more frequent in nuts infected with the navel orange worm (Amyelois transitella) (Sommer et al., 1986).

Copra. Coconut meat is probably sterile before the nut is broken. However, the thickness of the meat means that it will dry slowly, leading to growth of bacteria and fungi. Fungal infection in copra is often very high (Pitt et al., 1993) (Table 9.2). Dominant fungi were Asp. flavus (present in 86% of 21 samples and in 20% of all copra pieces examined), Asp. niger (present in 43% of samples, and 18% of all pieces), Rhizopus oryzae (52% of samples, 25% of all

Table 9.2 Fungal infection in 21 samples of surface disinfected copra from Thailand

Range of infection Percent of particles No. of infected

Average infected

infected averaged over Fungus

particles in infected

in infected

all samples Aspergillus clavatus

3 Asp. flavus

20 Asp. niger

18 Asp. tamarii

2 Endomycopsis fibuliger

3 Eurotium amstelodami

2 Eur. chevalieri

13 Eur. repens

3 Eur. rubrum

11 Mucor spp.

5 Nigrospora oryzae

2 Penicillium citrinum

7 Rhizopus oryzae

25 Sordaria fimicola

11 Samples infected

76 Data of Pitt et al. (1993).

445 Table 9.3 Fungal infection in 45 samples of surface disinfected cashew kernels from Thailand

NUTS, OILSEEDS, AND DRIED LEGUMES

Average infected

Range of infection Particles infected

in infected averaged over Fungus

No. of infected

particles in infected

samples (%) all samples (%) Aspergillus flavus

samples (%)

samples (%)

8 2–35 5 Asp. Niger

10 2–66 5 Asp. Sydowii

18 4–70 2 Chaetomium globosum

7 2–30 3 C. fumicola

5 2–10 1 Cladosporium cladosporioides

6 2–22 2 Eurotium amstelodami

20 2–35 3 Eur. chevalieri

8 2–24 3 Eur. rubrum

20 2–90 6 Nigrospora oryzae

8 2–22 5 Penicillium citrinum

7 2–20 2 Pen. olsonii

30 8–40 2 Samples infected

40 6–90 40 Data of Pitt et al. (1993).

pieces) and the storage fungi Eurotium chevalieri and Eur. rubrum (38–43% of samples; 11–13% of all pieces).

Cashews. Dried cashews contain a range of spoilage fungi, but usually at low levels (Table 9.3). This reflects both the arboreal source of the cashew and their very thick shell. Fungal infection can occur during drying under poor conditions, but infection levels are limited. The overall infection level of Asp. flavus in 45 samples of cashew nuts from Thailand was only 5% (Pitt et al., 1993).

Others. Thirty-three species of fungi were isolated from 149 hazelnut samples, the most commonly occurring species being Rhi. stolonifer and Penicillium aurantiogriseum (Senser, 1979). Pecans (37 samples) contained a wide variety of fungi: 119 species from 44 genera. Aspergillus species accounted for 48% of the more than 1300 isolates obtained; next came Penicillium (19%), Eurotium (18%), and Rhizopus (8%). The dominant species was Asp. niger (293 isolates), followed by Asp. flavus (207), Eur. repens (132), Eur. rubrum (109), Asp. parasiticus (100), Rhi. oryzae (68), and Penicillium expansum (61) (Huang and Hanlin, 1975).

An unusual range of genera was found in pecans that had previously been invaded by weevils in the field (Wells and Payne, 1976). Nearly half of 2300 isolates from several hundred moldy nuts were Al- ternaria or Epicoccum species. Penicillium species made up 25% of the total, and Aspergillus only 1.0%.

Peanuts Peanuts. Extensive documentation of the mycoflora of dried peanuts was provided by Joffe (1969).

During a 5-year period, fungi were isolated from over 400 samples of freshly harvested and stored peanuts from Israel. By far, the most common species encountered was Asp. niger, isolated from a low of 8% of kernels in one year to a high of 71%. The other dominant members of the flora were Asp. flavus (0–8%); Penicillium funiculosum (3–16%); Pen. purpurogenum (2–8%); and Fusarium solani (0–9%).

Pitt et al. (1993) examined more than 100 samples of dried peanuts from Thailand (Table 9.4). Thirty- one fungal species were commonly encountered and a further 26 species occasionally. The dominant fungi were Asp. flavus and Asp. niger, found in 95% and 86% of all samples respectively. Asp. tamarii (31% of samples) and Asp. wentii (20% of samples) were also common. Other Aspergilli, except Asp. candidus (4% of samples) were rare. Fusarium semitectum (19% of samples) and Fus. equiseti (10% of

MICROORGANISMS IN FOODS 6

Table 9.4 Major fungal infection in 109 samples of surface disinfected peanut kernels from Thailand

Average infected

Range of infection Particles infected

in infected averaged over Fungus

No. of infected

particles in infected

samples (%) all samples (%) Aspergillus flavus

samples (%)

samples (%)

44 2–100 41 Asp. niger

38 3–100 33 Asp. tamarii

3 Asp. wentii

4 Chaetomium fumicola

2 Eurotium chevalieri

33 4–100 15 Eur. repens

1 Eur. rubrum

14 Lasiodiplodia theobromae

4 Macrophomina phaseolina

8 Penicillium aurantiogriseum

36 2–100 2 Pen. citrinum

6 Pen. funiculosum

2 Rhizopus oryzae

15 Wallemia sebi

6 Samples infected

84 6–100 84 From Pitt et al. (1993).

samples) were the only common Fusarium species found; only low percentages of nuts were infected in any sample.

Macrophomina phaseolina and Lasi. theobromae (49% and 33% of samples, respectively) were the only field fungi commonly found. Nigrospora oryzae was present in 22% of samples, but only low numbers of nuts were infected. Other noteworthy fungi were Rhizopus oryzae and Wallemia sebi (in

60% and 12% of samples, respectively). Pen. citrinum was very common (in 46% of samples) and a wide range of other Penicillium species was also detected (Pitt et al., 1993).

Among storage fungi, Eur. rubrum and Eur. chevalieri (51% and 46% of samples) were very com- monly encountered. Eur. amstelodami (9% of samples) and Eur. repens (6%) were present at much lower frequencies.

Figures for infection in more than 250 peanut samples from Indonesia were similar to those from Thailand (Pitt et al., 1998). Infection with Asp. flavus was even higher, with 40% of the 12,500 kernels examined infected by this species.

The incidence of field and storage fungi in peanut samples from farms, middlemen and retailers in Thailand was also compared (Pitt et al., 1993) (Table 9.5). Infection rates by species commonly regarded as field fungi, e.g. Fus. solani and Macrophomina phaseolina, declined during storage. Some Penicillium species usually regarded as storage fungi, notably Pen. brevicompactum, Pen. janthinellum, and Pen. pinophilum, also showed sharp declines during storage. It appears that the traditional view of these species as storage fungi is inappropriate here. In contrast, numbers of Pen. glabrum increased during storage, and those of Pen. citrinum were unaffected by sampling time.

Oilseeds. There appears to be little published information on the microflora of dried oilseeds. Legumes. Bacteria on dried legumes are of little consequence when these commodities are consumed

after boiling or other heat processing. They may be important if the legumes are incorporated into soups or other high a w products.

During the drying process, legumes are readily invaded by fungi. However, spoilage is uncommon and mycotoxin production rarely of significance. The mycoflora of various kinds of beans in Thailand

447 Table 9.5 Comparison of the fungal flora obtained from disinfected peanut kernels collected from farmers and middlemen in

NUTS, OILSEEDS, AND DRIED LEGUMES

Thailand versus those from retail outlets

Farm and middleman

Retail

Average infected

Average infected

No. of infected particles in infected Fungus

No. of infected

particles in infected

samples (%) samples (%) Aspergillus candidus

samples (%)

samples (%)

2 Asp. niger

35 Asp. Wentii

18 Chaetomium fumicola

3 Cladosporium cladosporioides

Chaetomium spp.

6 Eurotium amstelodami

14 Fusarium solani

Macrophomina phaseolina

11 Penicillium aethiopicum

10 Pen. aurantiogriseum

2 Pen. brevicompactum

Pen. citrinum

9 Pen. funiculosum

8 Pen. glabrum

6 Pen. janthinellum

Pen. pinophilum

12 Samples infected

Syncephalastrum racemosum

80 Data of Pitt et al. (1993).

Only fungi with notable differences are included. For more complete overall data see Table 9.2. Table 9.6 Fungi commonly found invading dried legumes in Thailand

Mung beans

Soybeans

Black beans

particles in

Samples

particles in

Samples particles in

infected infected samples Fungus

infected

infected samples

infected

infected samples

(%) (Av. %) Aspergillus flavus

45 4 67 8 61 6 Asp. niger

14 10 12 3 35 7 Asp. penicillioides

6 20 0 – Asp. restrictus

16 12 0 – Chaetomium globosum

14 5 33 5 26 2 Cladosporium cladosporioides

13 2 49 18 39 4 Eurotium amstelodami

16 12 13 2 Eur. chevalieri

18 6 33 12 26 4 Eur. rubrum

11 2 51 13 22 3 Fusarium moniliforme

13 7 6 3 0 – F. semitectum

55 27 29 5 52 11 Lasiodiplodia theobromae

30 6 18 7 22 12 Macrophomina phaseolina

23 9 22 9 1 – Nigrospora oryzae

11 2 39 5 13 2 Penicillium citrinum

13 5 22 13 17 2 Data from Pitt et al. (1994).

were surveyed by Pitt et al. (1994) (Table 9.6). Fungal species found were a mixture of field fungi, which presumably had invaded before harvest and during the early stages of drying, and storage fungi, which would have invaded during the late phase of drying or developed in storage. Among the most common field fungi was Fus. semitectum, which was present in 55% of mung bean samples, 52% of black bean samples and 29% of those from soybeans. The mycotoxigenic species Fus. verticillioides

MICROORGANISMS IN FOODS 6

(=Fus. moniliforme) was found in 13% of mung bean samples. It was much less common in soybeans and not found in black beans.

Lasiodiplodia theobromae, a ‘universal’ field fungus, was present in 30% of mung bean samples, and 18–22% of the other types. Macrophomina phaseolina, a common pathogen on beans in the trop- ics, was found on 22–23% of mung bean and soybean samples, but was seen only once on black beans.

Asp. flavus was common on beans, as on other commodities from Thailand (Pitt et al., 1994). It is unclear whether this species can invade beans before harvest, or is only found as a storage fungus in this commodity. Although the proportion of samples infected was high (45–67%), levels of infection in individual beans in any sample were low (3–6%).

Among storage fungi, three common species of Eurotium were present in 11–18% of mung bean samples, 16–51% of soybean samples and 13–26% in those of black beans. The higher levels in soybeans probably reflect longer storage of soybeans than other legumes in Thailand, a probability borne out by the much higher incidence of other storage fungi, Asp. penicillioides and Asp. restrictus, in soybeans also.

Coffee. Picking cherries and spreading them on drying yards frequently damages coffee cherries, allowing the ingress of microorganisms, particularly fungi. If cherries are picked from the ground, contamination is likely to be high. Drying is often a slow process, in particular because of the environment in which coffee is grown. Coffee tress will not flower above 19EC but require high temperatures to mature, so coffee is commonly grown in upland areas in the tropics. In consequence, drying is often

conducted under less than ideal conditions, with morning mists or rain common in some growing areas (Teixera et al., 2001). Fungal growth occurs, sometimes with substantial heating. After drying, most cherries were infected with Penicillium, Cladosporium, Mucor, Fusarium, and yeast species (Taniwaki et al., 1999, 2001), though little detail has been published. Most studies have concentrated on determining the source of ochratoxin A, as production of this toxin appears to be initiated during drying.

B Spoilage Nuts Tree nuts. Spoilage of tree nuts at harvest is rare. It is usually caused by excessive rain around

harvest time, which may result in splits in shells, permitting fungi to enter and cause damage to the nut meats. Spoilage is usually due to discoloration.

Tree nuts contain only low levels of soluble carbohydrates, so slight increases in moisture content lead to sharp rises in a w . As a result, stored dried nuts are very susceptible to spoilage. Increases in a w can be due to moisture movement caused by uneven storage temperatures, as may happen in poorly insulated storage facilities or shipping containers. Refrigerated storage is widely used to retard the development of rancidity in stored nuts; if effective dehumidification is not practiced, increases in moisture content may result during storage. Many spoilage fungi grow poorly at low temperature; however, return of the nuts to ambient condition for shipping may result in rapid spoilage.

Marginal increases in moisture content will permit growth of Eurotium species. Shipment of nuts in containers across the tropics is a particular hazard, as unsuitable stowage, on decks or near engines, can lead to moisture migration sufficient to cause sporadic spoilage or even total loss. Cases of rampant growth of Asp. flavus and high aflatoxin production have been observed under these conditions.

Storage of nuts in tropical countries is of concern. Inadequate facilities may lead to moisture build up from humid air, from damp floors, inadequate ventilation, or moisture migration due to uneven solar heating. Spoilage or mycotoxin production may result.

449 Peanuts. Spoilage of peanuts due to discoloration is much more common than for tree nuts, due

NUTS, OILSEEDS, AND DRIED LEGUMES

to susceptibility to pre-harvest fungal invasion. Discolored peanuts are usually sorted out before retail packaging, in developed countries by automated color sorting machines, in less developed countries by hand. Color sorting, introduced as a quality control measure, has also proved to be an effective way of controlling aflatoxin levels. Reject nuts may be used in the manufacture of peanut oil, where refining processes remove both fungi and aflatoxins.

Coconut. Coconuts are opened and cut up for drying, and this permits contamination with bacteria and fungi. More than 50 species of fungi were isolated from 25 coconut samples by Zori and Saber (1993). The most common were Asp. flavus, Asp. niger, Asp. sydowii, Pen. chrysogenum, Cladosporium cladosporioides, Alt. alternata, Rhi. stolonifer, and Eur. chevalieri.

Lipolytic rancidity in coconut is associated with growth of Micrococcus candidus, M. luteus, M. flavus, Achromobacter lipolyticum, and Bacillus subtilis during the early stages of drying (Minifie, 1989). Rancidity in dried coconut may be caused by xerophilic fungi, usually by Eurotium species, which produce off odors and flavors due to ketone formation (Kinderlerer, 1984). An unusual type of spoilage was a cheesy, butyric off flavor due to growth of Chrysosporium farinicola (Kinderlerer, 1984).

In the production of coconut milk or cream, the expressed milk typically contains 10 5 –10 6 cfu/g of bacteria. Pasteurization at 75 ◦

C for 10 min within 2 h of expression is recommended to reduce the microbial load (Hagenmaier, 1980).

Oilseeds. All oilseeds are more or less susceptible to contamination or infection by fungi during growth and drying in the field. However, little information has been published on this topic.

Cottonseed. Cottonseed, a by-product of cotton production, is a valuable source of oil and of residual meal, which is used in stock feeds. Cottonseeds develop inside a tight, effectively impervious boll, which is highly resistant to fungal attack. Nevertheless cottonseed is known to be susceptible to invasion by Asp. flavus. This may occur as the result of insect damage, for example, by the cotton bollworm, but the main invasion route has been shown to be entry through the nectaries, glands in the cotton plant near the flowers which attract insects for pollination (Klich et al., 1984).

Legumes. Spoilage of dried legumes is uncommon. Some discoloration often occurs as the result of the growth of storage fungi. However, this is not considered to be such a critical factor as in nuts and does not usually result in downgrading.

Coffee. Spoilage of coffee usually relates to off flavor or aroma development. Little is known of the causes of these problems, but they are most likely to be related to growth of fungi during drying.

C Pathogens Bacterial pathogens. Bacteria have been reported from tree nuts, but total numbers as estimated by

plate counts are usually low. Higher numbers in almonds were related to damaged shells or contamination with soil (King et al., 1970). A variety of genera were isolated, including Bacillus, Brevibacterium, Streptococcus, Escherichia coli, and Xanthomonas. Total plate count, Streptococcus spp. and E. coli dropped initially during storage of almonds, then remained constant for more than three months (King et al., 1970).

Contamination in the field or processing plant by pathogens derived from animals can be important. Growth can only occur when favorable conditions of a w , pH, and temperature permit, but it has been found that very low numbers of bacteria such as Salmonella can be a hazard, as oil protects them after

MICROORGANISMS IN FOODS 6

ingestion. One outbreak of salmonellosis in peanuts has been reported, in Australia in 1996 (Ng et al., 1996; Oliver, 1996; Scheil et al., 1998) with more than 50 cases of infection and one death (Rouch, 1996; Burnett et al., 2000). Salmonella Mbandaka and S. Senftenberg were detected in suspect jars of peanut butter at levels as low as 3 cells per gram. This outbreak was probably due to bird or rodent droppings in improperly cleaned equipment in a peanut shelling plant.

Canned peanuts processed by an unlicensed canner resulted in an outbreak of botulism in Taiwan in 1986. Nine people were affected, of whom two died. Only a single batch was apparently contami- nated, and the cause was not determined. Lack of distribution records hampered recall, but mass media announcements are credited with a reduction in the severity of the outbreak (Chou et al., 1988). In the United Kingdom, contamination of hazelnuts with Clostridium botulinum led to botulism from yoghurt to which inadequately processed hazelnut pur´ee had been added (O’Mahoney et al., 1990).

Aflatoxins. The principal microbial hazard associated with nuts and oilseeds lies with the potential for production of mycotoxins, notably aflatoxins. Aflatoxins are produced by Asp. flavus and the closely related species Asp. parasiticus during growth in foods or feeds. Crops with high susceptibility to growth of these species and aflatoxin production have a high oil content, though the physiological reasons for this remains unclear. Peanuts, maize, and cottonseed are the three most economically important crops

affected by aflatoxins. Aflatoxins were discovered as the result of the deaths of 100 000 young turkeys in the United Kingdom in 1960 (Sargent et al., 1961). The toxicity was traced to peanut meal originating in Brazil. Subsequent work showed that the cause was the common mould Asp. flavus, and the closely related species Asp. parasiticus. The use of the new technique of thin layer chromatography soon established that four

toxins were implicated, named aflatoxins B 1 ,B 2 ,G 1 , and G 2, the names being based on the compounds’ blue or green fluorescence under ultraviolet light, and their positions on TLC plates (Broadbent et al., 1963).

It was later shown that Asp. flavus produces only B aflatoxins, and occurs throughout the tropics and subtropics, while Asp. parasiticus produces both B and G toxins, and is much less widespread (Klich and Pitt, 1988; Pitt and Hocking, 1997). Other species capable of producing aflatoxins have been discovered more recently, Asp. nomius (Kurtzman et al., 1987), Asp. ochraceoroseus (Klich et al., 2000), and Asp. pseudotamarii (Ito et al., 2001). However, these three species are very uncommon, have no known food safety significance, and will not be discussed further here. Aspergillus flavus, Asp. parasiticus, and Asp. nomius are closely related and have similar physiology (ICMSF, 1996b; Pitt and Miscamble, 1995).

Soon after their discovery, the acute toxicity of aflatoxins to all domestic animal species was es- tablished. Their potential carcinogenicity to animals and, by implication, man, became evident a few years later (Stoloff, 1977). A possible role in human cancer was supported by epidemiological stud- ies (e.g. Peers and Linsell, 1973; Peers et al., 1976). These apparently convincing data were soon confounded by the realization that the hepatitis B virus, endemic in many areas high in liver can- cer, was also a liver carcinogen, or at the least a strong potentiator of liver cancer. Indeed, the role of aflatoxins in human liver cancer was dismissed by some authors (Stoloff, 1989; Campbell et al., 1990).

Most recent studies have again supported the position that aflatoxins have a role in human liver cancer.

A careful epidemiological study established that liver cancer rates in different areas of Swaziland correlated well with aflatoxin intake, but was independent of hepatitis B, which varied little with geographic region (Peers et al., 1987). A good correlation was also reported between the incidence of liver cancer and the extent and severity of aflatoxin contamination of foodstuffs in Guangxi Province,

China (Yeh et al., 1989). In an authoritative report, IARC (1993) considered aflatoxin B 1 to be a Class

1 human carcinogen.

451 Evidence has been presented that both aflatoxins and hepatitis B virus are involved in the very high

NUTS, OILSEEDS, AND DRIED LEGUMES

incidence of primary liver cancer in some areas of the world, notably parts of Africa, Southeast Asia and China (Wild et al., 1993).

However, the human toxicology of aflatoxins still remains a source of uncertainty. The Joint FAO/WHO Expert Committee on Food Additives (JECFA) summarized this position: “Risks from specific exposures to aflatoxins [in humans] are difficult to estimate and predict. Ques- tions remain regarding the independence of aflatoxin as a human carcinogen, the extent to which hepatitis

B, hepatitis C, and other factors modify the effect of aflatoxin, how findings from countries with high liver cancer rates and high prevalence of hepatitis B may be compared to those from countries with low rates, and how to describe the dose–response curve over the wide range of aflatoxin exposure found worldwide” (JECFA, 1997).

Observations concerning the interaction of hepatitis B and aflatoxins in humans suggest that two separate aflatoxin potencies exist, one in populations in which chronic hepatitis B infections are common and the second where such infections are rare. In consequence, JECFA divided potency estimates for analyses based on toxicological and epidemiological data into two basic groups, applicable to individuals with and without hepatitis B infection (JECFA, 1997). Potency values chosen for hepatitis B positive individuals were 30 times higher than for hepatitis B negative individuals, implying a 30-fold synergy between hepatitis B and aflatoxins.

Very sensitive techniques have now been developed to monitor aflatoxin intake by humans, including DNA–aflatoxin adducts in tissues, aflatoxin adducts in serum and urine (Harrison et al., 1993), and

excretion of aflatoxin M 1 in breast milk (Wild et al., 1987).

Detectable aflatoxin levels have been recovered from human serum or milk by these techniques, in developed countries including Australia (El-Nazami et al., 1995), Denmark (Autrup et al., 1991), and the United Kingdom (Harrison et al., 1993), as well as less developed regions, including Kenya (Autrup et al., 1987), China (Ross et al., 1992), Gambia (Allen et al., 1992), Taiwan (Hatch et al., 1993), and Thailand (El-Nazami et al., 1995). In the Taiwan study, a much greater correlation was found between aflatoxin intake and primary liver cancer than with exposure to hepatitis B virus (Hatch et al., 1993).

In Southeast Asia, the major sources of aflatoxins in the human diet are dried peanuts and maize (Pitt and Hocking, 1996). Using their data, and equations relating aflatoxin intake to liver cancer incidence (Kuiper-Goodman, 1991), it has been estimated that aflatoxin causes 12 liver cancer deaths per 100 000 population per annum in Indonesia; the relative contributions of maize and peanuts were estimated to

be 7 and 5, respectively. That figure translates to more than 20 000 deaths per annum in Indonesia from aflatoxins (Lubulwa and Davis, 1994). Aflatoxins are regulated in foods traded throughout the world. Levels permitted by country legislation range from as low as 1 µg/kg total aflatoxins to 50 µg/kg (van Egmond, 1992), though in some developing countries, limits in products for domestic consumption are not enforced. It seems probable that the Codex Alimentarius Commission will set a limit of 15 µg/kg total aflatoxins in peanuts and other commodities traded internationally in the near future. Under that international control, it will still

be permissible for individual countries to set lower limits for commodities for local consumption, but not to refuse importation of commodities meeting that international standard. Meeting any lower internal limit will be the responsibility of processors within the country, not international producers or traders.

Levels of aflatoxins that can occur naturally in peanuts and other susceptible commodities are very high. Peanut plants affected by drought can produce kernels without obvious signs of damage but which may contain up to 1 000 µg/kg of aflatoxins (Hill et al., 1983). Kernels affected in this way may be difficult to detect visually or by color sorters, but may cause serious contamination of lots. Much higher levels have been observed in discolored nuts, 20 000 µg/kg or more (Urano et al., 1992).

MICROORGANISMS IN FOODS 6

Aflatoxins are very heat resistant. Heating cottonseed meal of 6.6% moisture content in a jacketed kettle at 100 ◦

C for 30 min reduced aflatoxins by 30% (Stoloff and Trucksess, 1981); and 40–50% reduction was achieved in a commercial roasting process (Waltking, 1971; ICMSF, 1996b). However, boiling in a pressure cooker at 116 ◦

C for 2 h reduced aflatoxin levels by 50% (Mann et al., 1967); boiling at 100 ◦

C with 5% salt, a commercial process in Brazil, reduced aflatoxins by 80–100% (Farah et al., 1983).

Ochratoxin A. The principal public heath hazard associated with coffee is the formation of ochratoxin

A. This toxin is discussed more fully in Chapter 8—Cereals. Nuts

Peanuts. Early work assumed that the presence of Asp. flavus in peanuts and subsequent afla- toxin production was primarily a function of inadequate drying or improper storage (Austwick and Ayerst, 1963). While aflatoxin production undoubtedly does occur due to poor storage, it soon be- came clear that the main time of entry of the fungus was pre-harvest (McDonald and Harkness, 1967; Pettit et al., 1971; Cole et al., 1982). Indeed evidence indicates that, in developed countries such as the United States and Australia, invasion of peanuts by Asp. flavus after harvest is rare (Pitt, 1989). Although insect damage or shell cracking may be responsible for nut infection (Graham, 1982),

such damage is not a necessary prerequisite for Asp. flavus to penetrate the peanut shell. Pre-harvest invasion of peanut seeds most commonly occurs directly through the shell from the surrounding soil.

Asp. flavus may also invade developing peanuts through flowers or the pegs on which the nuts develop (Griffin and Garren, 1976), or systemically (Pitt et al., 1991). Invasion of peanuts, and hence aflatoxin production, is promoted by pre-harvest drought and/or high temperatures, rather than post-harvest rain (Sanders et al., 1981, 1984, 1985; Cole et al., 1982, 1989). The likely explanation is two-fold. First, in the absence of drought, a vigorously growing peanut plant has a variety of defence mechanisms against invasion, including phytoalexin production (Dorner et al., 1989, 1991). Such defences become ineffective when the plant becomes physiologi- cally weak during drought stress. Secondly, Asp. flavus is a xerophile, capable of growth when drought stress reduces competition in soil from its natural enemies—bacteria, amoebae, and soil fungi (Pitt, 1989).

Asp. flavus and Asp. parasiticus are capable of growth down to about 0.8 a w (Pitt and Miscamble, 1995). Toxin production is most abundant at high a w (greater than 0.95), and is limited below 0.85 a w . The optimum temperature for growth of both species is between 30 ◦

C and 37 ◦

C, with a minimum near

C and a maximum at about 42 ◦

C (ICMSF, 1996b).

Peanuts are readily invaded by both Asp. flavus and Asp. parasiticus—in contrast to maize, where nearly all aflatoxin is produced by Asp. flavus. In the United States and Australia, 50% or more of aflatoxin in peanuts is derived from Asp. parasiticus, so that both B and G aflatoxins are normally present (Read, 1989). However, Asp. parasiticus appears to be a rare species in some other parts of the world (Pitt et al., 1993; Pitt and Hocking, 2004). Of more than 400 food commodity samples from Thailand, Indonesia and the Philippines, which contained aflatoxin, 95% contained only B aflatoxins (Pitt et al., unpublished).

A survey of Californian almonds from the 1972 season showed that 14% of 74 samples were contaminated with aflatoxins, but levels were low, mostly below 20 µg/kg. Commercial sorting was effective in removing almonds containing aflatoxins. Sliced almonds were more uniformly contaminated (Schade et al., 1975). A more comprehensive survey of almonds and walnuts later found that less than

Almonds.

453 one nut in 25 000 was contaminated (Fuller et al., 1977). Of 256 almond samples checked under US Food

NUTS, OILSEEDS, AND DRIED LEGUMES

and Drug Administration compliance programs from 1980–1984, 2.3% contained detectable aflatoxins; only 0.8% exceeded 20 µg/kg (Pohland and Wood, 1987).

Pistachios. Available evidence indicates that pistachio nuts are essentially sterile before harvest, due to thick shells and surrounding tissues. However, in some varieties of pistachio, shells dehisce at maturity, providing an opportunity for microorganisms to enter. Pistachios are also treated by mechan- ical abrasion and washing to remove the outer nut layers around the shell, providing an opportunity for microbial entry if shell dehiscence has occurred (Doster and Michailides, 1994; Heperkan et al., 1994).

Pistachio nuts have usually been reported to be free of aflatoxin (e.g. Burdaspal and Gorostidi, 1989; Abdel-Gawad and Zohri, 1993; Taguchi et al., 1995). However, of 835 imported pistachio sam- ples checked under US Food and Drug Administration compliance programs from 1980–1984, 2.0% exceeded 20 µg/kg (Pohland and Wood, 1987). In other countries, some reported levels have been very high: up to 400–800 µg/kg (Shah and Hamid, 1989) or even 1300 µg/kg (Tabata et al., 1993).

Pecan nuts. Of 446 pecan nut samples checked under US Food and Drug Administration com- pliance programs from 1980–1984, only 1.1% contained detectable aflatoxins and only one exceeded

20 µg/kg (Pohland and Wood, 1987). The cause of aflatoxin contamination in pecans has not been determined, as both damaged and undamaged kernels may contain aflatoxin (Pohland and Wood, 1987).

Brazil nuts. Brazil nuts were recognized as a cause for concern before 1980. Analyses of 135 samples in Germany showed 58% aflatoxin free, 22% up to 5 µg/kg and 21 higher, up to 8,000 µg/kg (Woller and Majerus, 1979). In the United States, in 1980, 4.4% of 158 lots examined under Food and Drug Administration compliance programs were found to contain unacceptable levels (>20 µg/kg) of aflatoxins. However, improvements to quality control in producing countries resulted in no unacceptable lots from more than 300 sampled in 1983–1984 (Pohland and Wood, 1987).

Cashews. Spoilage or mycotoxin production by fungi in cashew nuts is very uncommon. Coconuts. Due to less than ideal drying conditions for coconuts in most producing areas, bacterial

contamination and aflatoxin production are major problems in copra and desiccated coconut production. Salmonellae were frequently been isolated from desiccated coconut (Schaffner et al., 1967; Gilbert, 1982). This was traced to the unhygienic state of collecting centers at farms, where keeping free ranging fowls was the source of the contamination. The Salmonella problem has been largely overcome as a result of improved controls.

Bongkrek poisoning, from the Indonesian fermented coconut press cake product “tempeh bongkrek”, was first reported in 1895 and 1901, and caused a total of 850 deaths from 1951 to 1975, and 69 in 1977, among unsuspecting Indonesian consumers (Arbianto, 1979). This disease is caused by con- tamination of the fermenting press cake with Pseudomonas cocovenenans, now known as Burkholde- ria cocovenenans (Zhao et al., 1990), and is apparently confined to the Seraryu Valley of Central Java (Cox et al., 1997). Burkholderia cocovenenans produces two powerful toxins, bongkrek acid and toxoflavin (Arbianto, 1979; ICMSF, 1996a; Cox et al., 1997). The onset of illness apparently occurs 4–6 h after ingestion, with death following 1–20 h after the onset of symptoms (Cox et al., 1997). Because of the many deaths, the Indonesian Government declared the manufacture of tempeh bongkrek illegal in November 1988, but this is unlikely to eradicate the problem (Buckle and Kartadarma, 1990).

MICROORGANISMS IN FOODS 6

Like other nuts and oilseeds, coconut is susceptible to invasion by Asp. flavus during drying. Aflatoxin in copra is a well-recognized hazard, but quantitative data are limited (e.g. Saxena and Mehrotra, 1990; Zohri and Saber, 1993).

Oilseeds. The major microbiological problem in oilseed crops is the growth of Asp. flavus and consequent aflatoxin production. High levels of aflatoxins have been found in a variety of oilseeds. In

73 samples of various oilseeds from South Africa, 43% were positive, with aflatoxin levels ranging up to 2 000 µg/kg (Dutton and Westlake, 1985); in linseed oil from India, 44% of 105 samples contained aflatoxins, with levels ranging from 120–810 µg/kg (Sahay et al., 1990); and in sesame seed oil from Pakistan, 55% of 24 samples were positive, ranging up to 440 µg/kg (Dawar and Ghaffar, 1991). Of

45 samples of maize destined for oil manufacture, 87% contained aflatoxins ranging up to 2300 µg/kg (average 250 µg/kg), and 100% of 50 similar peanut samples also contained aflatoxin, ranging up to

22 000 µg/kg, average 1600 µg/kg (Urano et al., 1992). Dried legumes. As noted above, growth of Asp. flavus or other mycotoxigenic fungi in dried legumes

(other than peanuts) is rarely sufficient to produce significant mycotoxin levels. Aflatoxins have been reported from a variety of beans including haricot beans and broad beans (Abdalla, 1988; Abdel-Rahim et al., 1989; Mahmoud and Abdalla, 1994), butter beans imported into Japan (Tabata et al., 1993) and soybeans (Pinto et al., 1991; El-Kady and Youssef, 1993; Jacobsen et al., 1995). In general, levels reported were low and of little consequence in trade.

Fusarium semitectum, which appears to have a strong association with beans, is not considered to

be an important producer of known mycotoxins (Miller, 1994). The potentially mycotoxigenic species Fus. verticillioides (=Fus. moniliforme) was found in 13% of mung bean samples in Southeast Asia (Pitt et al., 1994), but it is unclear whether toxin production is a potential problem. Levels of Asp. flavus were low in all kinds of Southeast Asian beans, indicating that significant aflatoxin production was unlikely (Pitt et al., 1994, 1998).

Soybeans, with their high oil content, might be expected to be a suitable substrate for aflatoxin production. However, available information indicates that aflatoxin formation in soybeans and soy products is not a commercial problem. Like other beans, soybeans possess factors that are highly inhibitory of the extensive fungal growth needed for toxin production.

Coffee. The principal problem associated with coffee is the production of the mycotoxin ochratoxin

A. Although the possibility of significant levels of this toxin being present in coffee beans has been known for some time, only recently has it been established that Asp. ochraceus is a major source of ochra- toxin A in Brazilian coffee (Pitt et al., 2001) and likely to be the major source world wide. Other known ochratoxin A producers, Asp. niger and Asp. carbonarius, have also been isolated from coffee, but do not appear to be as important (Frank, 2001; Pitt et al., 2001). Available evidence indicates that the sources of these fungi are environmental, and that entry to cherries is gained during picking and drying (Taniwaki et al., 1999; Pitt et al., 2001). Ochratoxin A is produced during drying (Taniwaki et al., 1999; Bucheli et al., 2001; Teixera et al., 2001). Coffee picked and dried under good agricultural practice appears to contain ochratoxin A only rarely (Taniwaki et al., 1999; Pitt et al., 2001). Nearly all ochratoxin A is produced in the husk rather than in the coffee bean, and immature fruit were less likely to become contaminated than mature or overripe cherries. Removal of defectives (damaged beans) and careful removal of husks re- duces ochratoxin A levels in dried green coffee beans (Bucheli et al., 2001). High levels of ochratoxin A in green coffee beans are related to poor manufacturing conditions, i.e. where drying is slow, waste ma- terial is admitted to the processing stream, or drying is incomplete and storage inadequate (Teixera et al., 2001).

NUTS, OILSEEDS, AND DRIED LEGUMES