Binary suspensions of bacteria isolated from the gastric juice of achlorhydric patients were used ... to form carcinogenic N-nitroso compounds (Mirvish, 1975).
Journal of General Microbiology (1987), 133, 1845-1849. Printed in Great Britain
1845
Nitrite Accumulation during Anaerobic Nitrate Reduction by Binary Suspensions of Bacteria Isolated from the Achlorhydric Stomach By S . J . F O R S Y T H E * t A N D J . A . COLE Department of Biochemistry, University of Birmingham, PO Box 363, Birmingham B1.5 2TT, UK (Received 26 September 1986; revised 4 February 1987)
Binary suspensionsof bacteria isolated from the gastric juice of achlorhydric patients were used to determine conditions which favour nitrite accumulation during nitrate reduction. Suspensions of Veillonellaparvula and Haemophilus parainzuenzae accumulated nitrite during nitrate reduction in the absence of nitrite-reducing Neisseria subflava or Streptococcus sanguis. The maximum concentration of nitrite that transiently accumulated decreased predictably as the ratio of nitrite-removing bacteria to nitrite-accumulating bacteria increased. This ratio, but more importantly the bacterial density, determined the duration of nitrite accumulation. These results are correlated with the previously reported tendency of nitrite to accumulate in the gastric juice of hypogammaglobulinaemic and pernicious anaemic patients, and with the extremely high incidence of gastric cancer in the two groups.
INTRODUCTION
Nitrite is an intermediate product in the bacterial reduction of nitrate to ammonia or dinitrogen. Nitrite will therefore tend to accumulate whenever the rate of nitrate reduction exceeds that of nitrite reduction. In certain microniches nitrite can react with secondary amines to form carcinogenic N-nitroso compounds (Mirvish, 1975). Clostridia and some species of streptococci can use nitrite to nitrosate secondary amines at a neutral pH although the reaction has not been proven to be enzymic (Hawksworth & Hill, 1971a, b). The achlorhydric stomach provides an environment which might favour bacterial formation of N-nitroso compounds because there is an established bacterial flora predominated by streptococci, an input of salivary and dietary nitrate and the pH is nearly neutral. Nitrite has been shown to accumulate in the gastric juice of achlorhydric patients, so this might be a contributing factor to their raised incidence of gastric cancer compared with the general population (Dolby et al. 1984; Kinlen et al. 1985). Forsythe et al. (1987) identified the predominant nitrate-reducing and nitrite-reducing bacteria from the achlorhydric stomach and determined the nitrate reductase and nitrite reductase activities of washed cell suspensions of these organisms. Haemophilus parainzuenzae and Veiltonellaparvula were suggested to be the predominant organisms responsible for nitrite accumulation because their nitrate reductase activities’ exceeded their nitrite reductase activities. Conversely, Neisseria subzava and Streptococcus sanguis were the dominant components of the gastric juice flora that reduced nitrite in preference to nitrate. N. subpava rapidly reduced nitrite to gaseous products with glucose or lactate as the electron donor. In order to understand more fully bacterial nitrate metabolism in the achlorhydric stomach, the present study was designed to investigate how nitrite accumulation was dependent upon the ratio of these two types of organism and the total cell density. ~~
Present address: North East London Polytechnic, Romford Road, London El5 4LZ, UK.
0001-37370 1987 SGM
1846
S . J . FORSYTHE A N D J . A . COLE METHODS
Bacterialstrains. Strains of V .puruulu (strain F740), H . puruinfluenzue (strain F829) and N.subfluvu (strain F813) were isolated from the stomachsof achlorhydric patients as reported by Dolby et ul. (1984) and, with the exception of S . sunguis, were the same isolates as those described by Forsythe et ul. (1987). S. sunguis was isolated from the stomach of a patient with a duodenal ulcer (Muscroft et ul., 1981). The rate of nitrite reduction by this S . sunguis strain was in the middle of the range of reduction rates previously determined for S. sunguis strains (Forsythe et ul., 1987). N . subfuvu was maintained by subculturing it every 3 d in Brain Heart Infusion Broth (Oxoid) supplemented with 5 rnhf-sodium nitrite. After 24 h at 37 "C, liquid cultures were stored at 4 "C for a further 48 h. Although N. subj¶uvu will tolerate more than 0-1 M-nitrite during anaerobic growth, the growth-limiting concentration used, 5 mM, was sufficient to allow adequate yields of bacteria and to induce nitrite reductase activity but was only slightly higher than the highest concentration detected in gastric juice of achlorhydric patients (Dolby et ul., 1984). Anaerobic growth conditions. Bacteria were grown as described by Forsythe et ul. (1987) except that the concentrationof sodium nitrite was raised to 5 m~ for culturing N.subfluuu and S . sunguis. Medium for V .puruuh and S. sunguis was prepared on the day it was required. Determinationof the rates of nitrate and nitrite reduction by washed cell suspensions. Anaerobic pure cultures were harvested in late-exponential phase and resuspended to a known optical density. Bacteria were washed and resuspended in nitrogen-saturated 50 m-sodium phosphate buffer (PH 7-5)except V. puruulu, which was washed and resuspended in 1% (w/v) Bactopeptone-50 m-sodium phosphate buffer (pH 7.5). Nitrate and nitrite reductase activitieswere determined as reported by Forsythe et ul. (1987). Glucose (50 m ~was ) the electron donor in binary suspensions of H.puruintwnzue and N . sub$uuu or S . sunguis. Lactate (racemized mixture, 50 msodium salt) was the electron donor for binary suspensionsof V .puruufu and N . subfluou. Both glucose and lactate were provided as electron donors for binary suspensionsof V .puroulu and S. sunguis. Rates of nitrite accumulation and removal by binary suspensions were determined by linear regression analysis of the mid-points of the respectiveslopes of graphs of nitrite concentrationagainst incubation period. Units of activity (U)are nmol nitrite formed from nitrate or reduced min-l. Determinationof nitrite concentration, Nitrite concentrations were determined colorimetricallyby the method of Radcliffe & Nicholas (1968). RESULTS A N D DISCUSSION
Rates of nitrate and nitrite reduction by suspensions of individual bacteria In preliminary experiments, rates of nitrate and nitrite reduction by the four organisms used in this study were determined. Values comparable to those reported by Forsythe et al. (1987) were obtained. No nitrate reductase activity was detected with suspensions of N. subflava or S. sanguis, but H . parainfluenzaehad an activity of 210 U (mg dry wt)-l in the presence of glucose; the corresponding rate for V . parvula with lactate as electron donor was 400 U (mg dry wt)-l. Rates of nitrite reduction by H . parainfuenzae and N . subfava with glucose as the electron donor were 60 and 120 U (mg dry wt)-l, respectively. In the presence of lactate V . parvula reduced nitrite at a rate of 9 U (mg dry wt)-l. Streptococcalnitrite reductase activities vary greatly between strains and so a strain with an intermediate activity of 20 U (mg dry wt)-l in the presence of glucose was selected for our experiments. Furthermore, suspensionsof streptococci rapidly lost their ability to reduce nitrite, so they were assayed within 3 h of harvesting. Bacto-peptone, which stabilizes nitrite reduction by veillonellae, did not stabilize streptococcal nitrite reductase activity. Nitrite accumulation during nitrate reduction by binary bacterial suspensions Initial experiments with binary bacterial suspensions were designed to determine conditions in which the nitrite-removing bacteria N. subflava or S. sanguis were able to prevent nitrite accumulation during nitrate reduction by V .parvula or H . parainfluenzae. In these experiments, the nitrite accumulationpotential was constant at lo9 bacteria ml-I but the density of the nitriteremoving bacteria was varied. Nitrite accumulated in binary suspensions of V . parvula and N . subfava when the ratio of veillonellae to neisseriae was greater than 1:9 (Fig. 1a). The maximum transient nitrite concentration decreased from 1.0 mM to 0.9 mM and 0.7 mM as the ratio of veillonellae to neisseriae decreased from 9 :1 to 3 :1 and 1 :1 respectively. At a ratio of 1 : 3, the duration of the
1847
Nitrite accumulation by binary suspensions
0.4
B '
**
A
20
C
lo
40
60
20
A A A
'4
40
60
80 100 120
C
u
0.8
Po
tAoo
0.4
A
L O
A
10
20
e ***
00ooo 0 0
20
Incubation period (min)
40
60
Fig. 1 . Nitrite accumulation curves for binary suspensions of nitrite-accumulating and nitriteremoving bacteria. In (a),(b)and (c)the density of V .p a d a or H . paruinJuenzae was held constant at lo9 ml-l and the ratioof the two organismswas varied by adding differentquantitiesof N. subJuvu or S . sanguis. These ratios were as follows. (a) V.p a d u to N. subJava: 0 , 1 : 9 ; 0 , l :3; A, 1 :1 ; 0 , 3 :1 ; 0, 9 :1 . (b) V.pumula to S. sanguis: A, 1 :1 ; e,1 :lo; 0 , 1 5 0 . (c) H.paruinJuenzae to S. sanguis: A, 1 :1; 0 , 1 :5 ; 0 , 1 :25. In (d)the ratio of V.parvulu to N . subJuva was varied but the total cell density was constant at lo9 bacteria ml-l; the ratios were 9 :1 (a),3:l (O), 1 :1 0,1 :3 (0)and 1 :9 (0).
transient accumulation was 8 min and the maximum concentration of nitrite that accumulated was only 0.2 mM. No nitrite was detectable after 18 min when the ratio of veillonellae to neisseriae was 1 :1 or after 55 min for the ratio 9 :1. The rates of nitrite reduction by the individual components of the binary suspensions were estimated by solving simultaneous equations for the rates of nitrite removal. The rate of nitrite reduction by the two species was 7 U (mg dry wt)-l for V .parvula and 111 U (mg dry wt)-* for N . subjava. These rates were comparable with those determined using pure cultures [9 and 115 U (mg dry wt)-l, respectively]. N . subjava and N . gonorrhoeae are unable to grow anaerobically in the absence of nitrite (Knapp & Clark, 1984; S. J. Forsythe unpublished results). Growth of N . subjava in the achlorhydric stomach under anaerobic conditions might, therefore, be sustained by nitrite produced by veillonellae. Binary suspensions of V. parvula and S. sanguis accumulated nitrite even when the ratio of streptococci to veillonellae was 50 :1 (Fig. 1b). This confirmed that nitrate is reduced far more rapidly by V. paroula than the slow nitrite reduction by S. sanguis. The maximum concentration of nitrite that accumulated with a 1 :1 ratio of veillonellae to streptococci was 1.1 mM, equivalent to the initial concentration of nitrate, but this decreased to 0.8 m~ when the ratio was 1:50. The rate of nitrite reduction by S. sanguis and by V .parvula in the binary suspensions, determined by solving simultaneous equations for the rates of nitrite removal with suspensions of streptococci and veillonellae in the ratios 1 :1 and 1 : 10, was 4 U (mg dry wt)-l for streptococci and 16 U (mg
1848
S. J . FORSYTHE AND J . A . COLE
8
0.6
.,
20 40 60 Incubation period (min)
80
Fig. 2. Nitrite accumulation curves for 1 :1 binary suspensionsof H . parainjluenzae and N. subJava at various total cell densities: 4 x lo9 bacteria ml-l; A, 2 x lo9 bacteria ml-l; .,1 x lo9 bacteria ml-' ; 0,5 x lo8 bacteria ml-'. The rates of nitrite accumulation and removal were 640 and 580 U ml-l at a densityof 4 x lo9 bacteria ml-l, 370 and 330 U ml-l with 2 x lo9 bacteria ml-l, 200 and 140 U ml-l with 1 x lo9 bacteria ml-l, and 120 and 100 U ml-l with 5 x lo8 bacteria ml-l.
dry wt)-l for veillonellae. These values are again similar to those determined with pure culture suspensions (Forsythe et al., 1987). The slower rate of nitrite reduction by streptococci than by neisseriae was reflected in the times taken for nitrite to disappear from binary suspensions (Fig. la, 6). Nitrite did not accumulate in binary suspensionsof H. parainfluenzae and S. sanguis when the streptococcioutnumbered the haemophili by 25 :1 but did transiently accumulate to a maximum concentration of 0.6 mM at a ratio of 5 :1. These results confirm the prediction from results of experiments with pure cultures that nitrite would not accumulate with streptococci to haemophili ratios greater than 8. No nitrite was detectable after 20 min in binary suspensionsof haemophili and streptococci in the ratio 1 :1 (Fig. 1 c) compared with 120 min for a similar ratio of veillonellae to streptococci (Fig. lb). This difference was due to the more rapid complete reduction of nitrite by haemophili than by veillonellae. In the previous experiments the total cell density varied due to the variation in the population of nitrite-removing bacteria. Similar results were obtained, however, when the ratio of the nitrite-accumulating to nitrite-removing bacteria was varied at a constant total density. For example, nitrite again accumulated to almost the maximum possible (1.1 mM) when veillonellae outnumbered neisseriae by 9 :1 at a total density of lo9 ml-l, and also transiently accumulated to 0.05 mM when the veillonellae to neisseriae ratio was as low as 1 :3 (Fig. 1 d ) . The predictable nature of the results from all of these experiments suggests that there is no antagonism between the organisms, and that the apparent K, values for nitrate and nitrite are sufficiently low for linear rates of reduction to be obtained. Eflects of total cell density on the duration of nitrite accumulation The preceding experiments established that the maximum concentration of nitrite that accumulated was dependent upon the ratio of the potential rate of nitrite formation to the rate of nitrite removal. If so,the maximum concentration of nitrite accumulated by a binary suspension should be independent of the total cell density. This was confirmed by varying the total cell density in the range 5 x lo8 bacteria ml-l to 4 x lo9 bacteria ml-l at a constant 1 : 1 ratio of haemophili to neisseriae (Fig. 2). The maximum nitrite concentration that accumulated was 0.7 to 0.8 mM in each case but the period for which nitrite was detectable increased with decreasing total cell density (Fig. 2). The rates of nitrite accumulation and removal also decreased as expected as the total cell density decreased. The decrease in rates of nitrite accumulation and removal were not directly proportional to the cell densities, suggesting that there was some inhibition of nitrite accumulation and reduction at the higher cell densities. If nitrite is a precursor of the carcinogenic N-nitroso compounds which are formed in the gastrointestinal tract by a combination of chemical and enzymic reactions, then the longer it
Nitrite accumulation by binary suspensions
1849
persists, the greater will be the tendency for the carcinogens to be formed. We have clearly demonstrated that nitrite accumulation and persistence are primarily determined by two factors. These are the ratios in the achlorhydric stomach of nitrite-accumulating organisms, such as veillonellae and haemophili, to nitrite-removing organisms, such as neisseriae and streptococci, and the total population density. Nitrite accumulates and persists when this ratio is high, but the total cell density is low. It is perhaps significant that the average bacterial density in gastric juice from hypogammaglobulinaemic patients is lo6 ml-l compared with lo7 ml-l in pernicious anaemic patients (Dolby et al., 1984; Borriello et al., 1986). Although nitrite is found in gastric juice from both types of patient, the highest concentrations detected in hypogammaglobulinaemic patients, 1 mM, is tenfold greater than the 0.1 m~ maximum nitrite concentration detected in pernicious anaemic patients (Dolby et al., 1984; Forsythe et al., 1987). This correlates with the observation that hypogammaglobulinaemic patients are ten times more prone to gastric cancer than those with pernicious anaemia and 50 times more prone than the general population (Kinlen et al., 1985). Collaboration with Drs J. Dolby and P. Borriello, Clinical Research Centre, Northwick Park Hospital, London is gratefully acknowledged. This research was supported by a grant from the Cancer Research Campaign. REFERENCES
BORRIELLO, S.P., REED,P. J., DOLBY, J. M., BARCLAY, F. E. & WEBSTER,A. D.B. (1985). Microbial and metabolic profile of the achlorhydric stomach : comparison of pernicious anaemic and hypogammaglobulinaemia. Journal of Clinical Pathology 38, 946-953. DOLBY,J. M., WEBSTER,A. D.B., BORRIELLO, S. P., BARCLAY,F. E., BARTHOLOMEW, B. A. & HILL, M. M. (1984). Bacterial colonization and nitrite concentration in the achlorhydric stomachs of patients with primary hypogammaglobulinaemia or classical pernicious anaemia. Scandinavian Journal of Gastroenterology 19, 105-1 10. FORSYTHE, S. J., DOLBY,J. M., WEBSTER, A. D.B. & COLE,J. A. (1987). Nitrate and nitrite reducing bacteria from the achlorhydric stomach. Journal of Medical Microbiology (in the Press). HAWKSWORTH, G. & HILL, M. J. (1971a). The formation of nitrosamines by human intestinal bacteria. Biochemical Journal 122, 28-29P. HAWKSWORTH, G. & HILL,M. J. (1971 b). Bacteria and
the N-nitrosation of secondary amines. British Journal of Cancer 25, 520-526. KINLEN,L. J., WEESTER, A. D.B., BIRD,A. G., HAILE, R., PETRO,J., S ~ ~ T H IJ.LF. L ,&THOMPSON, R. A. (1985). Prospective study of cancer in patients with hypogammaglobulinaemia. Lancet i, 263-266. KNAPP,J. S.&CLARK,V.L. (1984). Anaerobic growth of Neisseria gonorrhoeae coupled to nitrite reduction. Infection and Immunity 46, 176-181. MIRVISH,S. S. (1975). Formation of N-nitroso compounds : chemistry, kinetics, in vim occurrence. Toxicology and Applkd Pharmacology 31, 325-35 1. MUSCROFT, T. J., DEANS,S.A., YOUNGS,D.,BURDON, D.W. & KEIGHLEY, M. R. B. (1981). The microflora of the postoperative stomach. British Journal of Surgery 68, 560-564. RADCLIPPE,B. C. & NICHOLAS, D.J. D.(1968). Some properties of a nitrite reductase from Pseudomonas denitrificans. Bibchimica et bwphysica acta 153, 545554.
