Bacteriological Observations on the Leaves of Sarracenia purpurea

by John Lindquist

Thesis and two general review items in the Carnivorous Plant Newsletter:

The following are a few of the highlights from my thesis research and some subsequent work on pitcher leaf samples and can be added to the above list as a third general review item. Far from being comprehensive, this is just a start, and – as time permits – more will be added to make this a more representative review. This was basically an exercise in "classical bacteriology" done in the early 1970s when I could find some time during my employment as a bacteriology lab instructor. One should expect genome-based methods to be at the core of any similar project being attempted today, although the plate-dilution frequency technique discussed below could become a standard procedure for such a study.

As library and laboratory research progressed, various factors known to affect the environment of the pitcher leaf neatly sorted themselves into the cycle shown on the right. One might think that digestive enzymes produced by Sarracenia purpurea (if any) would be diluted significantly by accumulated rain water and possibly utilized as a nutrient by the indigenous bacteria. Perhaps a plant-produced surfactant could be found, but I digress.

Regarding the question mark associated with N2, bacterial nitrogen-fixation in the pitcher water was initially considered unlikely due to the inhibitory presence of organic nitrogen-containing compounds (including the secretions from the commensal larvae of the mosquito Wyeomyia smithii) and also ammonia produced by the bacterial population. Klebsiella pneumoniae was readily isolated, and cultures of the various strains were shown by the acetylene-reduction test to perform N2-fixation (a process associated with this common plant and soil-dwelling organism), but acetylene-reduction tests of the water samples themselves showed no N2-fixing activity. (So far, anyway!)

For plate counts of the usual mesophilic chemoheterotrophic bacteria that one might expect to find in natural samples, making dilutions of the samples and surface-inoculating plates of Plate Count Agar (Standard Methods Agar) is probably the best way to go. These plates would be good for picking representative colonies. What I consider to be a "classic" paper on the subject of quantitating and isolating bacteria from plants is "The Bacterial Flora of Beech Leaves" by V. Jensen – in Ecology of Leaf Surface Microorganisms (1971), edited by T. F. Preece and C. H. Dickinson (Academic Press, London). The author looked for at least eight groups of bacteria among the colonies on his dilution plates; the groups were based on general properties such as gram reaction, shape and colonial pigmentation. These groups could be quantified specifically. If I had the project to do over again, I would use this approach for the "general bacteria" and subsequently make as many specific identifications as possible. I think that the "standard" approach I used – making isolations from macrocolonies on the plates in the "plate-dilution frequency technique" discussed below – fell short of being as comprehensive and quantitative as one would expect from a study of this sort.

One can expect an individualized ecosystem in each leaf, and one can also expect each leaf to contain a mixed population of bacteria that could vary considerably in kinds and numbers, depending on nutrients, temperature and other changeable cultural factors. Counts of approximately 104 to 109 CFUs per ml were obtained for samples collected in summer and early fall.

Sodium caseinate2.0 g
Yeast extract2.0 g
Glucose1.0 g
K2HPO40.2 g
MgSO40.2 g
FeSO40.01 g
Agar15.0 g
Distilled water1 liter

To enumerate proteolytic bacteria – the numbers of which would be expected to rise and fall during the insect digestion process – I used a casein (milk protein) agar medium whose formula is shown at left. After incubation, a positive result is a clearing around the growth after the application of 1M HCl (around the colonies) which precipitates unhydrolyzed casein. (Any isolations are best made before application of the reagent.) Non-proteolytic bacteria also grow well on this medium, and a "total aerobic plate count" could be inferred.

Now the usual technique of doing plate counts would make it difficult to quantitate the proteolytic colonies, as the enzyme is extracellular and non-proteolytic colonies can be confused with proteolytic colonies on a crowded plate. To get around this problem, I utilized a technique to inoculate plates which is similar in concept to the MPN (most probable number) method, and it can be found in "Plate-Dilution Frequency Technique for Assay of Microbial Ecology" by R. F. Harris and L. E. Sommers, Appl. Microbiol. 16:330-334 (1968). Various ten-fold dilutions are made of the pitcher sample, and eight spot-inoculations (0.01 ml each) are put on a single plate for each dilution of the sample. With this technique, one can get an estimate of "total" numbers of those that can simply grow on the medium as well as numbers of proteolytic bacteria which were often found to be very much in the minority – usually about 1 to 10% of the total number. This method can be used with other substrates degradable by extracellular enzymes such as chitin and starch.

Note the photo on the right of a series of plates starting (at the upper left) with a plate that received eight 0.01 ml inoculations from a 10–1 dilution of pitcher water. The plate in the lower right received inoculations from a 10–8 dilution. One can see how the frequency of positive growth responses decreases with increasing dilution. Likewise the number of positive proteolytic reactions can be noted, and one can assume a general positive response for all inocula in the top five plates from the general clearing of the medium. (The concentric rings seen in some plates were on the old glass petri dishes used here.) From the number of inoculated areas which subsequently produced growth and the lower number that indicated a positive reaction for proteolysis, one comes up with the following counts per ml of pitcher water (after applying the table in the above-cited paper with the use of the second plate in this series as the "dilution level 1" plate): "total" – 2.45 X 108/ml; proteolytic – 7.96 X 107/ml. As with the MPN method, these are very approximate numbers and could be rounded off considerably. The technique is good and depends greatly on well-dried plates. Isolations can be made by subsequent streaking on separate plates.

Basal Medium:
Distilled H2O1 liter

Dominant numbers were sometimes seen for non-proteolytic, yellow colonies later identified as Flavobacterium and non-proteolytic, pink colonies showing similarities to Acinetobacter but unable to be identified at the time. Proteolytic bacteria that I often found in high numbers included Pseudomonas, Serratia (non-pigmented) and Chromobacterium. The latter two were significant among my chitinolytic isolates as well.

The basal formula for the chitin medium I used is shown at the left and is taken from A Guide to the Identification of The Genera of Bacteria (1967) by V. B. D. Skerman (Williams & Wilkins, Baltimore). Plates are poured in two layers: Bottom layer is 15 ml of basal medium. Top layer is 5 ml of basal medium mixed with 0.5 ml of a purified chitin suspension in distilled water. As the chitin makes the medium semi-opaque, the degredation of this substrate results in clear areas around the growth, similar to what is seen above but without the requirement of a precipitating agent.

For lactic acid bacteria, which include some proteolytic members (most notably the proteolytic strains of Enterococcus faecalis), I used APT Agar with 0.02% sodium azide and then incubated the plates aerobically at 30°C. Virtually anything coming up on those plates would be lactic acid bacteria. Colonies of different types are all small and are not too distinctive although there is some variation in opacity and colors (usually white or yellowish).

Among various other groups of organisms, I looked for coliforms according to the usual water analysis procedure of MPN enrichments followed by platings for isolations. I found Escherichia coli on occasion and used that fact to suggest that flies and other insects contributed that organism (found exclusively associated with fecal matter) and probably other organisms as well from outside sources.

It was thought that utilization of oxygen by respiring microorganisms would create anaerobic conditions in the deeper part of the pitcher water. Indeed, the anaerobic genus Clostridium was found in substantial numbers in several samples – approx. 103 to 106 per ml.

The real surprise was the general finding of purple non-sulfur photosynthetic bacteria in the pitcher water samples – sometimes in extremely high numbers such as in the sample shown on the right in which the concentration of these organisms was found to be 2 X 109 per ml. (Some remains of dead frogs can be seen stuck to the leaf.) Rhodopseudomonas and (subsequent to the thesis research) Rhodomicrobium were identified among the isolates. Actually, these organisms are primarily aquatic and their presence in any aqueous, anaerobic environment exposed to sunlight (their basic habitat) should not be surprising. In more recent investigations, I found these organisms in a lot of unexpected places including snow, rain, and water from bromeliads.

As for the microbial population in bromeliad water, I found total bacterial counts averaging 2.9 X 106 per ml from six greenhouse plants; the number of proteolytic bacteria from the same plants averaged 6.8 X 105 per ml. Naturally the thought comes to mind that maybe these plants are passively carnivorous, benefitting from amino acids and ammonia released from the decomposition of flies that fall in. I posed this question in the 1981 letter to the Carnivorous Plant Newsletter (cited above) and have recently become happily aware that this subject has been looked into as can be found here.

Determination of pH was really a problem, as the pH paper I was using in the 1970s was for highly-buffered solutions and just didn't work with the low-buffered pitcher water. When I used a pH meter, I found that slight shaking of the sample – which released carbon dioxide – caused the pH to change considerably. New kinds of pH paper which are designed for low-buffered solutions function reliably.

Return to the
Pitcher Plant Project
Index Page

Page last modified on 1/14/03 at 12:45 PM, CST.
John Lindquist, Department of Bacteriology,
University of Wisconsin – Madison