1.
INVESTIGATIONS OF THE EFFECT OF SOME ANTIBACTERIAL AGENTS ON A BACTERIOPHAGE.
A THESIS submitted by WILLIAM ROBERT LAING BROWN for the degree of DOCTOR OF PHILOSOPHY in the UNIVERSITY OF LONDON.
School of Pharmacy, University of London. April 1962. 2.
ABSTRACT.
The effects on coliphage T6r of 11 commonly used chemical antibacterial agents have been examined.
The effect of the agents on the multiplication of the phage was shown by their effect on the mass lysis of fluid cultures of the host by the phage. Only aminacrine hydrochloride and sodium lauryl sulphate showed significant inhibition of the growth of the phage at concentrations below those inhibiting the host.
For the examination of the effect of the agents on free phage particles, a method of assessment giving precise inactivation time estimates has been applied and two methods of interpreting the results have been compared. Large differences have been shown in the effectiveness of different antibacterial agents in inactivating the free phage. The effect on the inactivation time of changes in concentration of the antibacterial agent has been examined for five of the agents and the regression between log inactivation time and log concentration has been shown to be linear. The dilution coefficients for these agents fell into two distinct groups, those for ohioramine T and formaldehyde (about 2 and
3 respectively) being very much smaller than those for crystal violet, cetrimide and phenol (about 11,13 and 15 respectively). It has been postulated that this difference in dilution coefficients indicates different mechanisms of inactivation between the two groups and the possible nature of the mechanisms has been discussed.
Methods of cultivation of the phage and its stability under different conditions of storage have been investigated. A method of plaque counting by surface drop plating has been developed and its reproducibility examined. 3. CONTENTS. Title 1
Abstract 2 Contents 3 Acknowledgements 6
PART I. INTRODUCTION Definitions 8 Inhibition of the interaction of phage and host 9 Inactivation of free bacteriophage 13 Assessing the inactivation of free phage 16
Counting methods 16 Extinction methods 17 Phage extinction methods 18 Bacterial extinction methods 20 The viricidal activity of specific antibacterial agents on phage 23 Chlorine compounds 23 Formaldehyde 25 Medicinal dyes 27 Triphenylmethane dyes 27 Acridine derivatives 28 Mercury Compounds 29 Phenol and its derivatives 31
Surface active agents 33 Quaternary ammonium compounds 33 Anionic surface active agents 34
Object of present work 35 4.
PART II. EXPERIMENTAL TECHNIQUES AND RESULTS. A. Apparatus 37 1. Dropping pipettes 37 2. Volumetric apparatus 39 3. Automatic tilt measures 39 Miscellaneous containers 40 B. Materials 42 1. Water 42 2. Media 42 3. Antibacterial agents 43 C. The Bacterial Host 46 1. Characters 46 2. Maintenance of cultures 47 3. Stability of characters 47 4.. Viability of cultures 48 D. The Bacteriophage 51 1. Cultural characteristics 51 2. Counting of phage particles 52 3. Preparation of phage stocks 71 4. Storage of phage stocks 79 E. Inhibition of the action of the phage on its host 81 1. The effect of inocula size on the rate of action of the phage on its host. 81 2. The effect of antibacterial agents on the action of the phage on its host 84
3. The effect of size of phage inoculum on the rate of lysis in the presence of antibacterial agents. 90 5. F. Inactivation of free phage by antibacterial agarlia 94 1. Method of evaluation 94 2. Reproducibility of the method 98 3. Estimation of the Mean Single Survivor Time 98 h- Relation of Mean Inactivation Time to Mean Single Survivor Time 110
5. Two factors affecting the reproducibility of the 113 Mean Inactivation Time 6. The systematic examination of the antibacterial agents 116
PART III. DISCUSSION . The Bacteriophage 128 Phage plaque counting method 128 Cultivation of phage 129 Storage of phage 130 Inhibition of the action of the phage on its host 130
Inactivation of free phage 135 The method of evaluation 135 The precision and reproducibility of the method 139
The Mean Single Survivor Time 140 The relation of Mean Inactivation Time to the Mean Single Survivor Time. 141 The systematic examination of the antibacterial agents 1)i2
Phage reactions and other viruses 149 Suggestions for further work. 151
Bibliography. 151_ 6.
ACKNOWLEDGEMENTS.
I wish to record my gratitude to Dr. A.M. Cook for first suggesting the subject of this thesis and for his helpful and enthusiastic direction of the work presented here. I am also grateful for the technical assistance of Mr. A. Edwards and his staff in preparing the apparatus and materials used. 7.
PART I. INTRODUCTION. J. DEFINITIONS.
The term antibacterial agent may be used to describe any agent, physical or chemical, which inhibits the growth of or kills bacteria. Heat, sonic vibration, white light, ultraviolet and ionising radiations may all be classed as physical antibacterial agents while the term chemical antibacterial agent may be used to describe those chemicals having a
bacteriostatic or bactericidal action and referred to variously as disinfectants, antiseptics and germicides.
The investigations which are the subject of this thesis were concerned with examining the effect of chemical antibacterial agents on a baoteriophage and the term "antibacterial agent" will be used throughout the thesis as meaning this type of agent.
The agents considered have been further limited to include only those which belong to the classes of compounds commonly used as disinfectants or antiseptics, namely chlorine compounds, formaldehyde, medicinal dyes, mercury compounds, phenol and its derivatives and surface active agents. The antibiotics and other chemotherapeutic agents have been largely excluded. The study of the action of antibacterial agents on bacteriophages has two main aspects. The first is the action of the agents on extracellOrm phages and the second is the interaction of phages and their bacterial hosts in the presence of the agents. The present work has been primarily concerned with the former aspect but any such study must include a consideration of the latter. This arises because the examination of antibacterial agents for possible viricidal activity 90 requires a clear distinction to be made between viristatic and viricidal effects. Tests for virioidal action. must therefore be so designed as to ensure that conditions are provided in which any phage particle surviving exposure to the agent can subsequently multiply,
INHIBITION OF THE INTERACTION OF PHAGE AND HOST. Inhibition of the growth of phages by the presence of an antibacterial agent may result from inhibition of the growth of both the bacterial host and the phage. Alternatively, phage production may be arrested by a selective inhibition of phage growth with no effect on the growth of the host. Several surveys have been undertaken in which large numbers of substances have been examined for possible selective phage inhibitory action in the hope that compounds showing this action may have some application in the development of chemotherapeutic agents for viral infections in man. Notable amongst these surveys are those carried out by Asheshov and his co—workers (Hall et al, 1951; Asheshov et al , 1950, Bourke et a] (1252), Boyd and Bradley (1951), Chantrill et al (1952) Dickinson (1948), Graham and Nelson (1954), Mills (1955) and Wooley et a] (1952). From all these surveys the same general picture emerges; that, of thelundreds of compounds tested, the majority inhibit both bacteria and phage and only a very few show any selective phage inhibitory action.
Of these the majority have proved to have no application as chemotherapeutic agents against animal viruses.
In investigating the mechanism of selective inhibitors of phage growth 10.
a wide variety of techniques, some relatively oomplex, are required to show the effect of the inhibitor on all or some stages of miltiplication, A useful, indication of whether or not a compound possesses phage-speoifio inhibitory activity can, however, be obtained by the relatively simple procedure of examining its ability to inhibit visible or mass lysis of a bacterial culture infected with phage. In preliminary screening programmes this effect is conveniently shown on a qualitative basis using phage cultures on solid media the agent under test being localised in the medium by use of the cup-plate and filter-paper disc teohnique
(Jones and Schatz, 1946; Ball et a]., 1951; Asheshov et al, 1954). This type of procedure must then be followed by a quantitative determination of the concentration of phage inhibitor whioh will prevent mass lysis in fluid culture of the phage without affecting the growth of the host, the progress of lysis being followed by estimation of the turbidity of the culture. Finally, if the agent shows promise, the detailed examination already referred to may be undertaken using techniques such as the single-step growth curve (F1iis and DelbrUck, 1939; DelbrUckand Luria, 1949) or electronmicrographio and radiochemical examinations.
A detailed examination of the mode of aotion of the few selective phage inhibitors known is outside the scope of the present work and the subject has been extensively reviewed by Gots (1959). Two examples of selective phage inhibition are, however, of particklar interest in the investigations to be reported here.
The first example is the action of the acridine derivatives.
Modern interest in the action of these compounds on phage was stimulated by 11. the demonstration by Fitzgerald and Babbit (1946) that a number of acridines, in sub-bacteriostatic concentrations, inhibited the lysis of Escherichia coli infected with phage. From the large number of reports on the subject subsequently published (cots, 1959),it became clear that - (a)not all acridines had a selective action in inhibiting all phages,
(b) aoridines have little effect on free, mature phage00 (c)the inhibitory action of acridines can be reversed by the addition of ribonucleio acid (RNA) or desoxyribonuoleic acid (DNA), (d)the acridines act by blocking the final steps in the maturation of the phage particles.
De Mars et al (1953)confirmed the last point by demonstrating the presence of empty phage heads ('aoughnuts") in lysates obtained in the presence of proflavine. Later, De Mars (1955) showed that synthesis of protein and DNA occurs in the presence of proflavine and the bacteria lyse as usual
although infective particles are not liberated. Astrachan and Volkin (1957) suggested that the DNA produced under these conditions may, however,
be different from normal phage DNA, and Kay (1959) believes that proflavine acts by combining with newly formedDNA, so interfering with the incorporation of polyamines into the DNA, and preventing its orientation to fcrm the core of the phage head. The mechanism of the action of the acridines generally on the multiplication of phage is still not fully known. The second example of selective inhibition of phage growth of particular interest here is the action of crystal violet. Graham and
Nelson (1954) reported that the concentration of crystal violet which i2 inhibited the multiplication of L. strains of lactic streptococcus phage was 3 to L,4-.1 of the concentration required to inhibit the growth of their hosts. Viristatic concentrations of crystal violet had no effect on extracellular phage and the inhibitory action could be overcome by the addition of DNA. The authors suggested that crystal violet acts by combining with phage DNA or fraction thereof at some critical stage in the incorporation of the DNA into the phage. This effect has not since been confirmed for any other phage —host system. 3.3.
INACTIVATION OF FREE BACTERIOPHAGE Much of the work reported on the effect of antibacterial agents on free phage particles has, understandably, been inspired by the hope that some indication may be obtained on the possible action of these compounds on animal and plant viruses. Such comparisons must be made with caution. There are similarities in chemical, physical and biological properties between phages and other viruses but there are differences which may be, in some cases, significant. The nucleic acid present in many animal viruses is DNA, like that of phage, but some other viruses contain only
RNA. Further, the morphologically distinct head and tail structures shown by many phages, the mode of attachment of phage to its host and the complex injection meohanism by which the phage DNA enters the host cell has not been found in other viruses.
A number of investigations have been made on the effect of antibaoteriel agents on phage compared to theirefect on other viruses. Without exception the phages have proved more resltant to inactivation than other viruses. Burnet and Lush (1940) tested the action of 3 surface active agents (sodium lauryl sulphate, sodium desoxycholate and saponin) on 7 different phages and a large number of animal viruses.
Under the conditions of the test the infectivity of all the phages were unaffected while the other viruses were all inactivated by varying concentrations of the surface active agents. Klein et al (1945) tested 20 synthetic detergents on 3 phages, onlinfluenza virus and vaccinia virus and found the phages to be generally more resiant than the other 2 viruses. A similar difference in sensitivity to mercuric chloride 34.
between an influenza virus And a staphylococcus phage was shown by Klein et al (1948). In a review of the susceptibility of viruses to
ethyl ether, Andrewes and Horstmann (1949) tested 25 viruses including 3
phage and found the phage to be among the most resistant of the viruses
tested. In addition to the general difference in sensitivity between phage and other viruses, different phages show marked differences in sensitivity to inactivation by antibacterial agents. For example, in a long eries of investigations into the action of various antibiotics on a collection of
60 phages, Asheshov and his colleagues (Hall et al, 1951; Asheshov et al, 1952; Hall et Asheshov, 1953; Asheshov et al, 1954; Strelitz et al, 1955; Strelitz et al, 1956) found that some of the antibiotics inactivated some of the phages in the free state, There was, however, a great variation in the sensitivity of the phages, not only between phages with different hosts but between different strains of phage with the same host. No correlation existed between the sensitivities of
the phages and their size or the relative sensitivities of their hosts. Similar variations have been reported, by Jones (1945) for 2 coliphage
and a staphylococcus phage treated with several antibiotics; by Klein et al (1945) for coliphage T2, a shigella phage and a staphylococcus phage treated with 20 synthetic detergents (T2 being completely resistant
to all the detergents); by Schinzel and Lingau (1955) for coliphages T1, T2, T3, T4 and T5 treated with ethyl alcohol (the variation in sensitivities being greater than that shown between 8 species of vegetative
bacteria tested at the same time); by Heicken and Spicher (1959) for all 7 15. of the T-group coliphases treated with formaldehyde (the T-even phages being geaerally more resistant than the T-odd); and by Tamimi (1959) for 3 shigella phages treated with a formalised buffered glycerol saline.
It seems highly likely that the differences in sensitivity between viruses is due to differences in their detailed chemical structure.
Our present state of knowledge of both the chemical structure of phages and the mode of action of chemicals in inactivating phage is relatively limited, and previous surveys in which many phages have been tested against many potential inactivators have yielded little information on the mode of action of the compounds which showed antiphage action and even less information on the detailed structure or function of phages.
A more promising line of approach would appear to be to conduct a detailed examination of the action of a number of potential inactivators of diverse chemical types on a single phage. Such an investigation might yield information on the mechanism of action of chemical inactivators of the phage. Once this has been fully elucidated, the inactivators can then be tested against other phages with the object of explaining the difference in sensitivity in terms of biochemical structure. The work which forms the subject of this thesis, is, it is hoped, the beginning of such a programme. ,ISSESSING THE INACTIVATION OF FREE PHAGE. CountlEaMethods, In the majority of previously reported investigations into the inactivation of free phage, some form of plaque counting method has been used to assess the degree of inactivation of the phage. In essenoe, suoh a method consists of exposing phage particles to the agents and oarrying out plaque counts on the phages surviving after various intervals of time. The inactivating efficiency of the agent is then expressed as the percentage of the original phage inactivated by a given dose of the agent, as the dose required to produce a given percentage inactivation or as the velocity constant of the reaction under given conditions. The inactivation of phage by most lethal agents appears to be exponential with time but some departure from this relationship has been reported for the inactivation of phage by antibacterial agents. Krueger and Baldwin (1934) found that the rate of inactivation of staphylococcus phage K by mercuric chloride followed the kinetics of a first order reaction only to 99 per cent inactivation, the rate then slowing considerably. A similar effect was reported by Mills (1954) for the inactivation of pseudomenas phage Pb by an unspecified disinfectant.
The kinetics of the inactivation of the T6-group of coliphages and 2 staphylococcus phages by formaldehyde were found by Heicken and Spicher (1956, 1959) to show some deviation from that of a first order reaction. With the T.-odd coliphages and the staphylococcus phages the rate of inactivation was exponential with time except at concentrations of formaldehyde of the order of 0.01 per cent or less. With the T-even 17,,
coliphages they found an initial rapid inactivation followed by a slow
phase which they believed was due to a small proportion of phages reacting slowly with the formaldehyde, The authors claimed to have shown that
this small proportion were not mutants of the original phage.
Such alterations in the rate of inactivation place limitations on the use of plaque counting methods in comparing the inactivation of phage by different antibacterial agents, The comparisons are invariably based on something less than complete inactivation; commonly the reaction is followed in detail to no more than 99 or 99.9 per cent inactivation. If the rate of inactivation is not constant then the time to give say 99 per cent inactivation gives little indication of the time to inactivate
all the phage present and, with initial titers commonly of 107 or more, gives no indication of the reaction of large numbers of the original population. Extinction Methods.
Little attention appears to have been paid to the precise determinatio
of the time required to produce complete inactivation (that is, an extinction time estimate) as a means of comparing the effect of different
agents on phage. Such a determination will yield no information on variations in the rate of inactivation but will give an estimate of the action of the agent on the most resistant particles in the phage population being investigated.
Extinction methods have been widely used in investigating the action
of antibacterial agents on bacteria. They consist essentially of exposiri the organisms to the antibacterial agent, then, at various time intervals, 1£1,
removing samples of the reaction mixture into nutrient medium to test for
the presence of surviving organisms. The extinction time is the minimum time required to produce sterility in the reaction mixture. The applications and limitations of these methods have been reviewed by Wills (1955) and Sykes (1958). The main criticisms which can be made of most of the bacterial extinction methods described in the literature are lack of precision and lack of reproducibility of the estimates obtained. The first shortcoming results from the use of wide intervals of time between the taking of samples from the reaction mixture. The second results from failure to obtain a representative sample of the very small numbers of resistant organisms surviving towards the end of the disinfecting period.
The same criticisms can be made of the few examples given in the literature of extinction methods used in examining the effect of lethal agents on phage. asp Extinction Methods.
In experiments which were admittedly not designed to elucidate the dynamics or mechanics of the inactivation, Hunter and Whitehead (1940) tested the action of several chemical disinfectants on 8 representative phage for the lactic streptococci. The phages, suspended in cheese whey, were mixed with solutions of the disinfectants and held at l5-16°C. After various time intervals the reaction mixtures were sampled with a loop into milk containing the host organisms. The presence of bacterial growth after incubation indicated absence of viable phage. The time intervals used were wide, the "Time for complete destruction"'being quoted as, 19, for example, 5-30 minutes and 1-24. hours. No information was given on replication in the tests. Klein et al (1945) investigated the action of detergents on phages by preparing serial dilutions of the detergents and inoculating the dilutions with buffered saline suspensions of the phages. After 10 minutes at room temperature each reaction mixture was diluted by serial, tenfold dilutions first in broth then in host culture and volumes of the latter dilutions plated and incubated. Absence of plaque formatior indicated absence of surviving phage. The results were quoted as the approximate dilution of detergent giving complete inactivation under the conditions of the test but "complete inactivation" was, in fact,reduction of the original phage titer by 99.99 per cent or more. The tests were carried out in duplicate and "some variation" in end7pqints was obtained between individual estimates. The viricidal action of a number of disinfectants on T2 uoliphage was investigated by Deutsch and Rohr (1955). Their techuique consisted of mixing equal volumes of broth suspension of the phage and disinfectant solution and, after periods of 2 to 60 minutes, diluting the mixture 1 in 15, the disinfectant being chemically neutralised when necessary. Small volumes of these dilutions were then mixed with a host culture, plated and incubated. Complete inactivation was here considered to have occurred when no plaques were produced. The level of replication used is not clear. The disinfectants tested were classified into 3 groups - those giving complete inactivation in 20 minutes, those doing so in 60 minutes, and those having no effect on the phage.
Investigations of this type cannot, and in fact are not claimed to, 2O
give more than a qualitative indication of the presence in the antibacterial
agent of viricidal aotivity towards phage.
Bacterial Extinction Methods
It was suggested by Cade (1937) that the lack of reproducibility in
estimates of extinction times for bacteria was largely due to errors in
sampling the reaction mixture. He further suggested that this sampling
error could be reduced by more extensive replication and greater accuracy
in the measurement of the sample volumes. His recommendations, though
sound, were largely ignored until the introduction of the method of
Berry and Bean (1954). This method answered many of the criticisms of
earlier extinction methods and the essential features of it were as follows:--
(1) The reaction mixture of antibacterial agent and bacteria was sampled
immediately after inoculation and the samples maintained at the
required temperature of the disinfection process. This procedure
was adopted in an attempt to minimise the errors due to clumping of
the test organisms by the bactericide, errors which would become
progressively greater with increased time of contact before sampling
was carried out. No experimental evidence was offered to show that the
sampling procedure could be completed before a significant amount of
clumping occurred but unless clumping is instantaneous the procedure
will certainly reduce its effect on the distribution of organisms
between samples. Whatever the validity of the original justification
of the procedure, its adoption does permit the exact control of the
time of contact of organism and bactericide and facilitates the
simultaneous performance of several replicate determinations. To 21.
minimise the variation in sample volumes, the volumes used were relatively large, about 0.1 ml., and measured using a dropping
pipette. (2)The reaction was stopped at short time intervals by quenching the samples with suitable culture media. The time intervals used were between 1/5 and 1/10 of the expected extinction time and produced a more precise estimate of the extinction time than that obtained from
the wider time intervals of earlier workers. The procedure resulted, however, in.a wide variation between the individual estimates from replicate determinations, the variation increasing as the time intervals were decreased.
(3)Each estimation was performed with extensive replication and the result expressed as the mean of the extinction times obtained in individual replicate estimations. (4)The samples of reaction mixture were incubated at 37°C immediately after quenching. Temperature fluctuations at this point in the technique were shown to have a large bearing on the reproducibility between estimates due, presumably, to the incomplete recovery at sub-optimal temperatures of organisms damaged by exposure to the bactericide. Using results obtained by Berry and Bean,Mather (1949) devised an
analysis by which an estimate is obtained of the time at which there is,
on the average, one surviving organism per sample volume. The analysis
made use of the regression existing between contact times and log (—log
proportion of negative samples), theoretical grounds for this regression 22. being deduced, from the random chances of obtaining a negative sample from reaction mixtures containing different numbers of surviving organisms. The reliability of the extinction time data was assessed by testing the significance of departure from the regression line mid the analysis permitted computation of the standard error of estimations of mean single survivor times. A detailed examination of factors influencing the repreducibiliti of extinction time estimates in the evaluation of bactericidal agents was carried out by Cook et al (1956), Cook and Wills (1954., 1956a, 1956b, 1959) and Wills (1955) using the method of Berry and Bean with the analysis of Mather. These investigations represent the first attempt to examine extinction time estimates by a technique which took adequate account of sampling variations. The object of a major part of the work reported in this thesis was to examine the application of an extinction method, based on the principles of the Berry and Bean method, to the examination of the viricidal action of antibacterial agents on a phage. 2 3 .
THE VIRICIDAL ACTIVITY OF SPECIFIC ANTIBACTERIAL AGENTS ON PHAGES. Comparison of the results obtained by different workers for the reaction of phage with antibacterial agents is generally of limited value due to the previously discussed variations in the resistance of different
strains of phage and to the variety of experimental techniques and
conditions used. A brief review of previous reports will, however,
now be given since at least some impressions of the relative activities of different agents can be gained and the range of the recorded activities is of interest.
Chlorine compounds. Compounds containing available chlorine appear to be quite effective in inactivating free phage although the required concentrations of
available chlorine is generally higher than those required to inactivate bacteria. The inactivation 4hypochlorite of 8 lactic streptococci phage contained in cheese whey was founA by Hunter and Whitehead (1940) to be markedly reduced by an increase in the protein content of the whey. When the protein content was adjusted to the same level (0.5 per cent) in all samples then all the phages were inactivated by 0.05 per cent w/v available chlorine in less than one minute. Inkley (1948) found that 10 per cent sodium hypochlorite (presumably equivalent to about 2 to 5 per cent w/v available chlorine) was required to completely inactivate a pseudomonas phage, suspended in broth, within 10 minutes. Six per cent hypochlorite inactivated the phage in one hour but 1 per cent failed to do so in 48 hours. Using chloramine T, the same author found that 211..
10 per cent (equivalent to 2.5 per cent available chlorine) was required to inactivate the phage in 10 minutes whereas Deutsch and Ro1r (1955) found that 0.5 per cent chlornmine T would inactivate coliphage T2 within 20 minutes. Both workers carried out the reactions in broth.
No attempts have been made to investigate the mode of action on phage of compounds containing available chlorine. 25.
Formaldehyde. The action of formaldehyde on phage has been more widely studied than that of any other antibacterial agent.
The pseudomonas phage of Inkley (1948) was found to be completely inactivated in broth by 4. per cent formaldehyde in 10 minutes, by one per cent in 1 hour and by 0.4 per cent in 0 hours. Labaw (1949) found that the titer of a suspension of T1 coliphage containing 1.5 per cent formaldehyde was scarcely affected after more than one hour, while the
titres of T2 and T4 were reduced to 10-5 of the original by 0.2 per cent formaldehyde after 15 minutes contact. Hershey and Chase (1952) found a reduction of 99.9 per cent or more in the titer of T2 exposed to 0.035 per cent formaldehyde for one hour at 37°C while Deutschand Rohr (1955) reported that T2 (with an initial titer of 106) was completely inactivated by 0.4 per cent in 20 minutes. While the exact mode of action of formaldehyde on phage has by no
means been fully elucidated, it seems likely that it attacks the protein
coat of phage. It was clearly demonstrated by Hershey and Chase (1952)
that T2 coliphage, inactivated by formaldehyde, is still adsorbed on to
its host and kills it but the phage DNA is not released. A similar reaction was reputed by Bourgaux (1957) for coliphage N treated with formaldehyde. The experiments of Kubinski (1960) on'the effect of
substances extracted from E.coli B in increasing the sensitivity of T2 coliphage to formaldehyde indicated that the formaldehyde was not acting on DNA since treatment with the extract did not make the phage
more sensitive to methylene blue which is known to react with nucleic acid. 26.
It has further been clearly established (Staehelin, 1958; Berns and
Thomas, 1961; Grossman et al,, 1961) that formaldehyde will not react with amino groups involved in strong hydrogen bondings as is the case in
DNA. That formaldehyde forms an unstable combination with phage is indicated by the ease with which formaldehyde inactivated T-group phases are reactivated by storage in the presende of histidine (Heicken and Spicher, 1959). Sueh an unstable combination implies that the formaldehyde is combined with some part of the protein coat of the phage.
On the other hand, Sauerbier (1960) suggested that formaldehyde inactivates Tl phage not by protein damage but by reacting with the phage
DNA. He therefore concluded that some amino groups in the DNA of T1 must be free and not all the possible hydrogen bonds are established.
The findings of Sauerbier have so far been unconfirmed although the demonstration by Hutsaar3(1957a, 1957b, 1959) that multiplicity reactivation occurs in formaldehyde inactivated coliphage-N implies that inactivation here occurred by damage to the DNA. 2.7
Medicinal Dxls.
Interest in the medicinal dyes in relation to phage has been directed mainly to their inhibitory effect on phage multiplioation and to their photosensitising effect on free phage. This latter effect, being an aspect of inactivation by irradiation, is outside the scope of the present work. Little information is available on the effect of dyes on freo phage without irradiation. Tri02almethane Dyes. This class of compounds includes malachite green, brilliant green and crystal Tiolet,
Inkley (194.8) claims that his pseudomonas phage was completely inactivated in 6 hours but not in one hour by both crystal violet and brilliant green at a concentration of one per cent in broth.(The strength of the solutions indicates that the crystal violet used was the material originally described as crystal violet but which was less pure, though more soluble, than the material used today). No information is available on the mode of action of the triphenylmethane dyes on free phage but crystal violet is known to have a strong affinity for DNA with which it combines in preference to amino acids or histones. (Stearn, 1930; Mirsky and Ris, 1951; Graham and Nelson, 1954). Acriaine derivatives.
The numerous reports on the effect of the acridines on phage multiplication contain many comments on the absence of any inactivating effect on free phage of viristatic concentrations of the acridines. Experimental details of the viricidal tests are rarely stated. Inkley (1948) found that his pseudomonas phage was completely inactivated by one per cent proflavine in 10 minutes, but that 0..1 per cen failed to completely inactivate the phage in 6 hours. Hotchin (1951) showed that some acridine compounds had a slight inactivating effect on free phage. Among the more effective compounds were 5-amino-acridine and Rivanol (2:5 diamino-7-ethoxyacridine lactate) although even their activity was slight. The mechanism of action of the acridines generally on phage is unknown but proflavine has been shown to combine with free DNA
(Peacocke and Skerrett, 1956; Luzzati et al, 1961). 29. h:erct+ Compounds.
Inorganic compounds of mercury appear to be relatively efficient in inactivating free phage but only in relatively high concentrations.
Inkley's (1948) pseudomonas phage was completely inactivated in 10 minutes when exposed to one per cent mercuric chloride but survived exposure to 0.1 per cent for over 6 hours and Deutsch and Rohr (1955) found that T2 phage was inactivated by one per cent mercuric chloride in less than 20 minutes. On the other hand Krueger and Baldwin (1934) reported that several days exposure to 2.8 per cent mercuric chloride were required to completely inactivate staphylococcus phage K. Klein et al (1948) found that 0.01 per cent mercuric chloride reduced the titer of a staphylococcus albus phage by only 99 per cent in 30 minutes. Less information is available on the activity of organic compounds of mercury but Latarjet and Morenne (1954) found T2 phage to be quite resistant to inactivation with phenyl mercuric borate. Treating the phaga with 0.1 per cent of the mercurial for 30 minutes at 37°C reduced the original titer by 99.9 per cent. Deutsch and Rahr (1955) reported that a 4 per cent solution of an organic mercury compound of vaguely specified composition failed to inactivate T2 in 60 minutes.
That phage inactivated by mercury salts can be reactivated by treatment with hydrogen sulphide (Krueger and Baldwim,1954) or sodium thioglycollate (Klein et a30 1948) indicates that the mercurials act on the phage protein, as they do on bacterial protein (Fildes, 1940; McCulloch, 1945), by combining with compounds of the P-SH type, the combination being a loose one. 30.
Latarjet and Morenne (1954) claimed that treatment of free phage
T2 with phenyl mercuric borate did not affect the adsorption of the phege to its host although infectivity was reduced. .31.
Phenol and its derivatives.
Although phenol and., more particOarly, its derivatives are widely used as antibacterial agents little precise information is to be found on their action on phage. Such information as is available indicates that phenol itself is relatively ineffeotiTe in inactivating phage. A lactic streptococcus phage was not completely inactivated by 2.5 per cent phenol in the presence of 0.5 per cent milk protein after 14 days (Hunter and
Whitehead, 1940), a pseudomonas phage was only slowly inactivated by one per cent phenol in broth (Inkley 1948), T2 coliphage was unaffected by 2 per cent phenol in broth after 60 minutes (Deutsch and Rohr, 1955) and purified T2 in phosphate buffer was incompletely inactivated by 3 per cent phenol in 30 minutes at 4°C(Flose.t/NespeAi r4ez) Phenol appears to have no effeot on free DNA. Berns and Thomas (1961) extracted DNA from T2 and T4• coliphages using very high oonoentrations of phenol. That the free DNA obtained is unaltered by exposure to the phenol is shown by the fact that it did not react with formaldehyde until subjected to a controlled degradation which broke the hydrogen bonds but not the ester links.
The results of Inkley (1948) indicate that phenol derivatives inactivate phage only in high concentrations, pseudomonas phage being inactivated by 10 per cent Lysol in 10 minutes, by 6 per cent in one hour but not by one per cent in 24 hours. A proprietary preparation containing chloroxylenol, which is roughly equivalent to Chloroxylenol Solution B.P., also inactivated the phage in 10 minutes at a concentration of 10 per cent but failed to do so in 24 hours at one per cent. Deutsch and 32.
Rohr (1955) found. that T2 ooliphage was unaffected by exposure to a 2 per cent solution of cresol for one hour. 33.
Surface Active Quaternary Ammonium Compounds.
The most interesting feature of the few published reports on the action of quaternary ammonium compounds on free phage is the differences in effectiveness of different quaternaries on the same phage under similar conditions of test. This is in addition to the usual variation in sensitivity found between different phage.
Klein et al (1945) showed that concentrations of 7 quaternaries required to inactivate a shigella phage and a staphylococcus phage to the same extent in 10 minutes varied from 1 in 250 to 1 in 4000 or more. The activity of the compounds generally decreased with decreasing alkyl Chain length and no activity was foundina.quaternary with an 8 carbon chain. None of the compounds were effective against T2 coliphage. The high resistance of T2 is confirmed by a report by Deutsch and Rohr (1955) that T2 was unaffected by 60 minutes contact with a 10 per cent solution of two quaternary compounds. In developing a technique using quaternaries for the isolation of phage from sewage, Halter et al (1946) found that one coliphage was unaffected by 10 minutes contact with 4. different quaternaries at a concentration of 1 in 5,000 while a second coliphage was
90 to 99.9 per cent inactivated by 3 of the compounds and unaffected by the fourth.
No suggestions have been made as to the mechanism of action of the quaternaries on phage. 34.
Anionic Surface Activegents.
Anionic surface active agents are very much less effective than the quaternaries in inactivating free phage. Klein et al (1945) found that of 9 anionic detergents tested against 3 phages with 10 minutes contact, only sodium lauryl sulphate and one other proprietary compound, both. at concentrations of 1 in 500, inactivated the shigella phage and the staphylococcus phage used. T2 coliphage was unaffected by any of the compounds tested. Burnet and Lush (1940) found that 7 different phage strains all resisted inactivation by 1 in 500 sodium lauryl sulphate for
2 hours at 36°C.
Under appropriate conditions, sodium lauryl sulphate has been shown by Mayers and Spizizen (1954) to cause the separation of DNA from the coliphages T2r, T6r and T7. The procedure used included treatment of the phages with 1 per cent sodium lauryl sulphate in the presence of sodium acetate at half saturation and at a temperature of 60°C for 15 minutes.
It is a matter for conjecture if the disruption of the phage particle by this treatment has any bearing on the mode of action of sodium lauryl sulphate on phage under the conditions used by Klein and his co—workers. 35.
92.2(t of the work.
The object of the present work was to study the effect on a bacteriophage of a number of antibacterial agents representative of the various classes of bactericides in common use. The main interest in the study was the effect of the agents on free phage particles using an extinction time method to obtain a precise estimate of the activity of the agents. It was hoped that the activity of the various types of bactericide would give some indication of the mechanisms involved in the inactivation of phage particles with the probable extension of the work* to other viruses.
While this was the primary object of the work, it was found necessary to develop methods of cultivation and assay of the phage and the results of these studies form part of the subject of this thesis. 36.
PART II. EXPERIMENTAL TECHNIQUES AND RESULTS. 37. A. APPARATUS
1. Dropping Pipettes. These were prepared from 7-8 mm, glasstubing, one end being drawn out, reinforced and ground to fit a "Record" fitting hypodermic needle, as described by Cook and Yousef (1953). The tips of No.20 gauge "Record" needles were ground to remove their bevels and polished to give a smooth square end. In using the pipettes a needle and a 1 ml. rubber teat are fitted to the glass barrel, the pipette is clamped vertically and charged with the liquid which is to be measured. Drops are delivered from the pipette at a constant rate of one per second (in time with a metronome),the first few drops being rejected to ensure uniformity in drop size before the required number of drops is collected. Cook and Yousef (1953) investigated the accuracy of these pipettes in measuring small volumes and showed that, using 5-drop samples, the variance between samples from separate needles was statistically significant compared with the variance between samples from any one needle but that the accuracy of the pipettes compared favourably with that given by all-glass dropping pipettes. These findings were confirmed for sample sizes ranging from 1 to 10 drops by Cook (1954), Wills (1955) and Steel (1959). The latter also showed that the variance in single-drop weights obtained on successive days was of no practical importance. The object of the calibration procedure used here was to obtain an estimate of the mean drop weight (and hence the mean drop volume) of water delivered from a batch of needles under conditions of practice. To minimise the inaccuracies incurred in weighing small 38. volumes of water the calibration was based on the weights of 60-drop samples of water. A preliminary calibration of 12 needles was carried out by determining the mean weight of water delivered in 10 drops from each needle. Using the results dbtained, 8 needles were selected, the mean drop weights of which were most nearly equal. The weight of 60 drops of water delivered from each of these needles was then determined on 3 separate days and an analysis of varianoe carried out on the weights obtained. The result of the analysis agreed with the findings of the workers previously referred to in that the between needles variance and the between drop variance were not significantly 2 = 3.38) while both were significantly greater than the different (F 7 residual varianoe. The total variance was, however, small and the mean weight of 60 drops of water with 95% confidence limits was found to be 1.0389 I 0.00268 g. The reliability of the technique used in delivering smaller numbers of drops with the pipettes had previously been tested on a different batch of needles by weighing 10 samples of 6 drops from each of 4 needles. The mean in this case was found to be 109 mg. and the 95% confidence limits of the mean + 0.19 mg. The variation in the drop weights obtained was considered to be acceptable and the volume of one drop of water from any of the 8 needles used in this work was taken as 1/58 (0.0172) ml. Throughout this thesis, the term "drop" shall mean a drop delivered from one of the standardised needles. The routine cleaning and sterilisation of the pipettes was as follows: the needles were removed from the glass barrels and sterilised by boiling for not less than one minute in sterile water while the 39.
barrels were immersed in a 5% solution of phenol for not less than 12 hours. When necessary, the lumens of the needles were washed with distilled water from a hypodermic syringe and the needles boiled in several changes of sterile water. When not in daily use, the needles were autoclaved and stored in a screw-capped bottle. The barrels of the pipettes, on removal from the phenol solution, were rinsed in tap water, immersed in chromic acid cleaning fluid fox• 12 hours, rinsed repeatedly in tap water in an automatic pipette washer and then rinsed in 3 changes of distilled water. They were finally dried in hot air, packed in glass tubes and sterilised by dry heat,150°C for one hour.
2. Volumetric Apparatus. All pipettes and graduated flasks used were of normal laboratory Class B accuracy. The 1, 5, 10 and 25 ml. graduated pipettes used were calibrated for delivery from a graduation mark down to the jet. The 2, 5, 10 and 25 ml. one-mark bulb pipettes used complied with B.S. 1583:1950 and the 50, 100, 250, 500 and 1,000 ml. one-mark graduated
flasks with B.S. 1792:1952. After use all pipettes were treated in the manner described for
dropping, pipettes, being packed in glass, cardboard or aluminium tubes or wrapped in paper before sterilisation. Volumetric flasks were cleaned by being rinsed in tap water, washed in warm tap water containing Teepol (Shell Chemical Company Limited), rinsed in tap water, rinsed twice in distilled water and dried in a hot-air oven. Finally, the flasks were capped with aluminium foil and
sterilised in a hot-air oven at 150°C for one hour,
3. Automatic Tilt Measures.
For the measurement of approximate volumes of media, use was made of automatic tilt meaeures ("Er-mil" Brand, Eipps measures, manufactured by H.J. Elliott, Ltd. Glamorgan). The measures were attached to standard B.24 ground-glass joints so that they fitted the 500 ml. B.24 conical flasks in which the media were normally contained. They were most commonly used to measure 20 ml. volumes and occasionally 5 ml. and 10 ml. volumes of media. The volume delivered by a 20 ml. measure, as determined by weighing 10 separate volumes of water delivered by ea9h of 3 measures was 19.98 : 0.04 ml. (P = 0.95). The measures wore cleaned and sterilised in the same way as the volumetric flasks.
4. Miscellaneous containers. The containers used for the preparation of cultures of he bacterial host were bottles closed with aluminium screw-caps fitted with rubber washers, 4 fl. oz, flat-sided bottles normally being used for fluid cultures and 1 fl. oz. round bottles for slope cultures. The fl. oz. bottles were also used as containers for peptone and sterile water iii the preparation of dilutions of phage suspensions and bacterial host cultures. Tubes of approximately 20 ml. capacity and fitted with B.24 ground glass stoppers were used as medication tubes in the investigations on viricidal activity. These tubes were also used in the preparation of dilutions of phage and bacterial suspensions when the volume of diluent involved was 10 ml. or less. The test-tubes used for incubation of quenched samples in the investigation on viricidal activity were 6 inch x 1 inch hard glass tubes closed with loose fitting metal caps.
The pet .-dishes used complied with B.S. 611:1952 and were of pressed glass. 41.
After use, infected containers were sterilised by autoclaving at
115°C for not less than 15 minutes before being subjected to the cleaning procedure described for volumetric flasks. The glass-stoppered tubes and petri -dishes were packed in metal boxes and the test-tubes in wire crates before being sterilised in a hot-air oven at 150°C for one hour. The screw-capped bottles were not normally sterilised until after media had been filled into them when sterilisation was effected by autoclaving at 115°C for an appropriate time. 4.2. B. MATERIALS. 1. Water. The water usod throughout this work was distilled from two
"Manesty" stills, type 0.0 BE. When required sterile it was packed in screw-capped bottles and sterilised by autoclaving at 11500 for 30 minutes. 2. Culture Media. The liquid medium used, referred to subsequently as "broth", was a peptone water containing 1% "Oxoid" peptone and 0.5% sodium chloride. The medium was prepared in 50 litre batches with double the final concentrations of peptone and sodium chloride dissolved in distilled water and the reaction adjusted to pH 7.4. This double-strength medium was distributed in 14. litre quantities into 1 gallon sorewcapped bottles and sterilised by autoclaving for 45 minutes at 1150C. When required, the double-strength medium was diluted with an equal volume of distilled water, distributed in 100 ml. quantities into It fl. oz. screw-capped bottles or in 4.00 ml. quantities into 500 ml. plugged conical flasks and sterilised by autoclaving for 20 minutes at 115°C. The reaction of the medium after the seoond autoclaving was checked on several occasions and found to vary between pH 6.9 and 7.0. For some experiments the medium in its double-strength form was distributed into conical flasks and sterilised as above. With this treatment the reaction was again found to be between pH 6.9 and 7.0. Since this broth was used as the diluent in the preparation of phage suspensions and small volumes of these suspensions were measured using dropping pipettes, the mean drop volume of broth delivered from the dropping pipettes was determined. The method used was that L3.
described by Withell (1938) and the mean drop volume was found to be 1/63 (0.0159) ml. The solid medium used was peptone agar prepared by the addition of 2:i per cent of shredded agar to the peptone water. Solution was
effected by steaming, the medium clarified by passage through washed paper pulp, distributed into 20 fl. oz. screw-capped bottles and sterilised by autoclaving for 15 minutes at 115°C. When required for use the medium was remelted in an autoclave by heating to 115°C momentarily. 3. Antibacterial Agents.z. All the antibacterial agents used were obtained from British Drug Houses Ltd. Those compounds described below as of "B.P. quality" were supplied as complying with the standards of purity and strength
specified in the British Pharmacopoeia 1958. From the time of receipt from the manufacturer the agents were stored at ambient room temperature (20-25°C) in tightly closed screw-capped or glass-stoppered containers. Unless otherwise indicated below, experimental solutions of the agents were prepared by dilution of stock solutions. The stock solutions were prepared by dissolving the agent in sterile water and were stored in glass-stoppered bottles, at ambient room temperature and protected from light, for up to six months. The solutions showed no deterioration during storage.
(a) ANINACRINE HYDROCHLORIDE .(5-aminoacridine hydrochloride). The sample used was of B.P. quality and the stock solution
prepared from it contained 0.207 per cent w/v of the sample (8.32 x 10-3 M). 41f.
(b) CETRIMIDE (a mixture of dodecyl, tetradecyl and hexadecyl
trimethylammonium bromides). The sample used was of B.P. quality and the stook solution prepared from it contained 0.0518 per cent w/v of the sample (equivalent to 1.42 x 10-3 M assuming sample to be pure oetyltrimethylammonium bromide).
(o) CHLORAMINE T. (toluene -p-sulphonsodiochloramide). The sample used was supplied as a B.D.H. Laboratory Reagent and solutions of it were prepared on the day on which they were intended to be used. The molarity of solutions was galoulated using a molecular weight
of 281.7. (d) CHLOROCRESOL (2 -chloro-5-hydroxytoluene). The sample used was of B.P. quality and the stock solution prepared from it contained 0.204 per cent w/v of the sample (1.1+3 x 10-2 M).
(e) CHLOROXYLENOL SOLUTION. The solution used was Chloroxylenol Solution of the B.P. 1958 and was manufactured in the laboratory. It contained 5 per cent w/v ohloroxylenol (4-chloro-3:5xylenol)
(equivalent to 3.19 x 10-1 M chloroxylenol), 10 per cent w/v terpineol and 20 per cent alcohol (95 per cent) in an aqueous solution of potassium ricinoleate. Dilutions of this solution were
prepared on the day on which they were intended to be used.
(f) CRESOL AND SOAP SOLUTION. The solution used was Cresol and Soap Solution of the B.P. 1958 and was manufactured in the laboratory. It contained 50 w/v of Cresol (equivalent to 4.62 M calculated as pure cresol, C6H4(OH)CH3) in an aqueous
solution of a Soap obtained by the action of potassium hydroxide
on linseed oil, Dilutions of this solution were prepared on the day on which they were intended to be used. 45.
(g) CRYSTAL VIOLET (hexamethylpararosaniline hydrochloride). The sample used was of B.P. quality and the stock solution prepared -3 from it contained 0.101 per cent w/v of the sample (2.47 x 10 M).
(h)FORMALDEHYDE. This was used as Formaldehyde Solution of IANALARt quality which, when assayed by the method of the B.P. 1958, was found to contain 36.83 per cent CH2O (12.27 M ). This solution was stored exposed to the light and dilutions prepared from it on the
day on which they were intended to be used.
(i)PHENOL. The sample used was of 'ANALARt quality and the stook solution prepared from it contained 5.00 per cent w/v of the
sample (5.31 x 10— M).
(j)PHENYLMERCURIC NITRATE. The sample used was of B.P. quality and the stock solution prepared from it contained 0.0018 per ceLt of the
sample (2.84 x 10 5 M).
(k)SODIUM LAURYL SULPHATE (a mixture of sodium normal primary alkyl sulphates, consisting chiefly of sodium lauryl sulphate). The sample used was of B.P. quality and solutions of it were prepared
on the day on which they were intended to be used. The molarity of solutions was calculated as if the sampie was pure sodium lauryl
sulphate with a molecular weight of 288.38.
C. THE BACTERIAL HOST.
The bacterium used as the host for the phage being investigated was a phage sensitive strain of Escherichia coli. 1. Characters of the bacterial host. The organism was a motile, gram-negative, non-capsulated, short rod, about 1 ja in diameter and 2 -3)along. It grew well in peptone water at 37°C producing a uniform turbidity in cultures up to about 18 hours old; older cultures showed a slight and readily dispersible sediment. On peptone agar convex to umbonate, finely granular, greyish-white colonies with a slightly crenated edge were formed.
The following carbohydrates were fermented within 18 hours with the production of acid and gas: Arabinose, Dextrose, Fructose,
Galactose, Lactose, Mannitol, Mannose, 3 is ose, SorbitoI, Trehalose and Xylose. Fermentation of Dulcitol, Glycogen, Maltose and Salicin was apparent in four days and Adonitol, Aesculin, Cellobiose, Dextrin, Erythritol, Glycerol, Inulin, Inositol, Raffinose, Sucrose and Starch were not fermented.
Acid and gas were produced in MacConkey broth at both 37°C and
44°C; acid and a clot were formed in litmus milk; indole was produced; nitrate was reduced to nitrite; a positive reaction to the methyl red test and a negative reaction to the Vosges -Proskauer test were shown and citrate could not be utilised as sole carbon source.
The characters of the organism are those of a typical
Escherichia coli. Because of its sensitivity to the phage used the strain will be referred to as E. coli B although exact identity with 47, the strain for which this designation was first introduced by Delbrfick and Luria (1942) was not attempted,
2. Maintenance of cultures of the bacterial host. A culture of the organism was maintained on peptone agar slopes in screw-capped bottles. After inoculation the slopes were incubated at 37°C for 16-18 hours with the caps slackened then stored, with caps tightly closed, at room temperature (20 - 25°C) in the dark. Duplicate subcultures were made every 4 weeks; one of these cultures being used to continue the stock culture, the other being subcultured into broth at weekly intervals. Each of these broth cultures became the first in a series of daily subcultures in broth, the series being maintained for 5 to 6 days after which a fresh subculture from the slope was used to initiate a new series. The broth cultures consisted of 100 ml. volumes of broth contained in 4 fl, oz. screw-capped bottles, inoculated from a slope using a loop-or from a previous broth culture using a 0.1 ml. inoculum and incubated with the caps slackened. They were normally used after between 18 and 30 hours incubation.
3. Stability of the characters of the bacterial host. A 24 hour broth culture of E. soli B was normally dense but homogeneous. Periodically, however, such a culture showed more granular growth with a tendency to produce a readily dispersible pellicle. If this culture was plated out on peptone agar, two types of colonies were readily distinguishable; one had the typical appearance of E. soli B colonies, the other was smaller (normally about half the diameter of the Lypical colonies), more convex and had a more finely granular and
rather whiter surface. Pure cultures prepared from the two types of colony were identical in microscopical and. biochemical characters and both were sensitive to the phage. In broth culture, organisms from the
small colonies grew more slowly than did typical E. coli B cultures and, in the presence of phage, the cultures lysed more slowly. The small colony organisms appear to be simply a slow growing variant of the normal host cells. The characters of this variant were maintained during artily subcultures for up to four weeks. When phage counts were
performed using both the normal E. coli B and its variant as bacterial host, no difference could be detected in the counts obtained and the occurrence of this variant in cultures of the bacterial host was considered to be of no practical importance. The cultural, microscopical and biochemical characters of the
bacterial host were examined periodically during the course of the present work and, apart from the occurrence of the slow growing variant discussed above, the characters remained constant.
Ii., Viabilitmsf daily cultures of the bacterial host. In order to obtain an estimate of the mean number of viable organisms in daily broth cultures of the host and of the day-to-day variation in the viability of these cultures, viable counts were performed on the cultures at intervals during the time the present work was being carried out. The counting method used was the "surface viable" method
described by Miles and Misra (1938). The cultures were serially diluted using 6-drop samples delivered from a standardized dropping 49. pipette and suitable volumes of sterile water contained in screw- capped bottles. The level of dilution used was chosen so that at least one of the dilutions finally plated would contain between 5 and
20 viable organisms per drop. The plates used for the counts were of peptone agar, dried by incubating overnight at 37°C with their lids in place and then noverdried" by incubating, inverted and with the lids removed, at 37°C for 30 to 40 minutes in an incubator with forced air circulation. At least 3 separate drops of the dilutions selected for plating were placed on the surface of each of 5 plates and the plates left undisturbed until adsorption of the drops was complete.
Colonies were counted after incubation at 37°C for 12 to 18 hours. 3 The counts for the dilutions yielding between 5 and 20 colonief per 2 drop were subjected to X tests for homogeneity and any result 2 giving a value of X. corresponding to a probability of less than 0.05 was rejected. The remaining results were multiplied by the appropriate dilution factors to give an estimate of the number of viable organisms in the original culture.
The mean viable count obtained from 9 separate counts on daily broth cultures performed at intervals over a period of 3 years, the age 8 of the cultures counted varying from 18 to 28 hours, was 4.18 x 10 organisms per ml. and the 95 per cent confidence limits of the mean 8 1.59 x 10 .
It is of interest to compare this result with the mean viable count obtained from a set of 2+ broth cultures in plugged flasks, inoculated and incubated simultaneously. The cultures were counted after from 22 to 24 hours incubation, 3 separate counts being 50, performed on each culture. The mean viable count of these cultures 8 was 3.88 x 10 organisms per ml. and the 95 per cent confidence limits 8 of the mean 1.60 x 10 . The close approximation of this mean to that obtained from the counts performed over 3 years on cultures maintained in screw-capped bottles indicates:- (a)that the apparent variation in the viability of the daily broth cultures over 3 years was of the same order as the variation found in replicate cultures prepared on the
same day, and
(b)that the organism grows equally well in bottles and in flasks. It was found that in aerated broth cultures of E. coli B counts of 3.61 (1.- 1.20 x 109 could be obtained with 19 to 26 hours incubation but for most routine procedures the static cultures were found satisfactory. 51,
D. THE BACTERIOPHAGE.
1. Cultural Characteristics.
The bacteriophage used throughout this work was the colipnage classified as T6 by Demerec and Fano (1945). The strain used was supplied by Dr. Meynell of the School of Hygiene and Tropical Medicine,
London.
Broth cultures of E. coli B incubated until a perceptible turbidity appeared and then inoculated with the phage were rapidly cleared, the cultures remaining clear for a time before becoming turbid again as phage-resistant mutants of the host multiplied. The rate at which the reaction proceeded increased with increasing concentrations of phage in the inoculum and no evidence of "lysis inhibition" (Doermann,191i.8) was seen. When suitably dilute suspensions of phage were dropped on to the
syrface of peptone agar plates previously surface-seeded with a broth
culture of E.coli B end the plates incubated at 37°C for 6 to 8 hours, plaques were produced which closely resembled those described by Demerec
and Fano for T6, the plaques being small (about 0.5 mm in diameter) and
having no halo. However, on incubating the plates for a further 12
hours, the majority of plaques increased in size (to about 1 to 1.5 mm
diameter) and developed definite, clear halos, between 1 and 2 per cent
of the plaques remaining small and undifferentiated. A number of
cultures prepared from these small plaques produced large plaques on platin
and it seems likely that the small plaques were produced by unstable
variants of the normal phage strain.
The cultural characteristics of the phage strain correspond to the art (rapidly-lysing) form of T6 (Hershey 1946), 52.
The constancy of plaque morphology, host specificity and rate of lysis of broth cultures of the host were accepted as satisfactory evidence of absence of significant variation in the phase during the course of the work which is the subject of this thesis.
2, Counting of phase particles. (a) Definition of "phage particle" ang_plasue count. The concentration (or titer) of phage particles present in the suspensions of phage prepared and investigated in this work was determined by plaque counts. Such counts determine the number of particles, present in a given volume of suspension, which will produce plaques on or in a solid medium seeded with the host bacterium after suitable incubation. A plaque count is therefore a measure of the number of infective particles; that is particles of phage which can initiate lysis of the host under the conditions of assay being used. There are indications that all the phage particles present in a preparation may not be infective in this sense. Hershey et al (1943) found that crude lysates of coliphage Ti and T2 contained several times more specific phase antigen than can be accounted for as infective units by plaque counts, while Luria et al (1951), comparing the plaque counts obtained from suspensions of several of the T group of coliphage with the number of specific phage particles present in the suspension as determined by electron micrographic counts, found ratios of infective units to electron microscopically visible particles ranging from 1.4 to 0.4.5. It has, however, been clearly established that plaque counts, are strictly 530 proportional to the aliquot of phage plated and, within the limits of experimental error, are reproducible. It has also been shown that, at least for some phage, close agreement can be obtained between plaque counts and counts of the number of particles capable of multiplying in liquid cultures as estimated by dilution end-point methods (Hershey et al,' 1943; Eleczkowski and Kleczkowski, 1951). These facts have led to the general acceptance of the view that, under given conditions, plaque counts of a phage preparation at least provide a reproducible measure of the number of infective particles present and that this number is characteristic of the phage. An estimate of the titer of a phage preparation based on the result of a plaque count should, strictly, be expressed in terms of "plaque producing particles". For the purpose of the present work the terms "phage particle" and "infectious particle" will be used as synomymous with "plaque producing particle". (b) Plating phage suspensions for plaque counts.
A number of methods of performing plaque counts have been described but these differ essentially only in the method by which the phage suspensions, suitably diluted, are finally plated. The two most widely used methods of plating are the "spread plate" method of DtHerelle (1917, 1920) in which a mixture of the phage and the bacterial host are spread on the surface of agar medium plates and the "agar layer" method of Gratia (1936) and Hershey et al (1943) where the phage and the host are mixed in small volumes of "soft agar" (0.7 per cent) which are then poured on to the surface of normal agar plates.
A third method of plating, briefly described by Williams-Smith (1951, 1953) 54.
is, in effect, a modification of the method of Miles and Misra (1938) for
counting bacteria. The procedure described consists of spreading a small
volume (0.08 ml) of host culture over the surfaee of agar plates previously "overdried" by incubating with their lids raised, and then dropping small volumes of the phage suspension on to the surface of the plates. The drops are absorbed and, provided the phage suspension is suitably dilute, discrete plaques are produced on the surface of the medium during incubation. This Miles and Misra type of plating technique appeared attractive in that it offers a high level of replication of samples plated with economy in materials and apparatus. Little attempt, however, has previously been made to assess the reliability of phage counts using this method of plating and a detailed examination of the method has therefore been undertaken in the present work. (i) Method of inoculating plates with host.
After some preliminary experimentation the following procedure was adopted.
Peptone agar plates containing about 15 ml. of medium were dried by
incubating overnight at 37°0 with their lids in place and then "overdried" by incubating, inverted and with the lids removed, at 37°C for one hour in an incubator with forced air circulation. Each plate was then flooded with 3 ml. of a 21i. hour broth culture of E. Goli B, allowed to stand for about one minute then drained for about 5 minutes by inverting and standing them at an angle in a tray. The plates were finally redried
with their lids raised, at 37°C for 15 to 20 minutes in an incubator without forced air circulation. 55.
Plates prepared in this way absorbed a drop of phage suspension in broth in less than 10 minutes and up to 12 drops could be conveniently and separately located on one plate. Incubation of the plates at
37°C for 6 to 7 hours resulted in the production of a uniform lawn of bacterial growth and clearly defined plaques. The plaques from a single drop were usually distributed fairly regularly over a circular area of 1 to 1.5 cm. diameter and up to 20 to 25 plaques per drop area could be counted with ease. It was found that "overdrying" the plates without the preliminary drying overnight produced wrinkling of the surface of the agar which resulted in an irregular bacterial lawn and indistinct plaques. Plates flooded with the host culture before overdrying also produced an irregular bacterial lawn. (.ii) Effects of variations in plates. A series of investigations were performed in order to determine the effect on plaque production of variations in the method of preparation of the plates and their subsequent treatment. Depth of Medium. Five batches of plates were prepared, eaoh batch consisting of 4 plates containing 5 ml., 10 ml., 15 ml., 20 ml., or 30 ml. of peptone agar per plate. All the plates were dried and inoculated with E. soli B by the method previously described and 10 drops of a dilute suspension of phage in broth placed separately on eaoh plate. The plates were incubated at 37°C for 6 hours and the plaques produced were examined and counted. The 5 ml. Plates were excessively dry and the plaques which had been 56. produced were indistinct and almost confluent. All the other plates showed clear, discrete plaques. The mean number of plaques per drop produced on the 10, 15, 20 and 30 ml. plates were 12.75, 13.12, 11.37 and 12.00 respectively. An Analysis of Variance, carried out on the mean number of plaques per drop calculated from each individual plate, showed that the variance due to different depths of medium in the plates was not significantly greater than the variance between replicate plates (42 = 1.88). It was considered that, provided the volume of agar in each plate is not less than 10 ml„ variations in the depth of medium have no effect on the plaque count. Dree of drying. Twenty approximately 15 ml. peptone agar plates which had been incubated at 37°C overnight were divided into 4 batches of 5 plates each. One batch received no further drying treatment and the other three batches were noverdried" as before for 30 minutes, 60 minutes and 90 minutes respectively. Each plate was then treated as in the previous experiment. On the plates which had received no "overdrying" the area of the drops of phage suspension were large and not sufficiently discrete for accurate counting. The remaining plates all showed discrete, easily counted plaques. The mean number of plaques per drop produced on plates ftoverdried" for 30 minutes, 60 minutes and 90 minutes were 14.40, 14.62 and 14.28 respectively. An Analysis of Variance on the mean plaques per drop calculated from each plate showed that there was no variance between the three different treatments. The plaque counts obtained are not, therefore, affected by variations in "overdrying" time M'7. within ±30 minutes of 60 minutes. Density and age of the host culture A 24 hour broth culture of E. coli B and 5 dilutions of the culture prepared by mixing samples of the culture with broth in the proportions 4:5, 3:5, 2:5, 1:5 and 1:10 respectively were each used to inoculate 4 plates by the standard procedure. Each plate was then treated as in the previous experiments. The viable count of the original culture was 8 1.9 x 10 . The mean number of plaques per drop for the different host inocula are given in Table 1 . Table 1 The effect of the densit of the host culture on la ue counts.
Dilution of host Undiluted 4:5 3:5 2;5 1:5 1:10 culture 1 Viable cells per ml. of host inoculum 19 15 11 7.6 3.8 1.9 (no7)
Mean number of 14.25 13.40 14.03 11.35 11.03 9.40 plaques per drop i Plaque count as percentage of 100 94 99 80 77 66 highest count.
The plaque count is therefore dependent on the concentration of host cells initially present in the bacterial lawn although (in this case) the effect became appreciable only when the concentration of cells fell
below 60 per cent of the original. The explanation probably is that
as the bacterial concentration falls the chances of multiple infection increases and each phage particle is less likely to produce a discrete
plaque. 58.
To test the significance of the effect of day-to-day variation in the density of the host cultures used in seeding plates for plaque counts, plates in batches of 4 were inoculated by the standard technique using broth cultures of the host ranging in age from 4 to 48 hours. The plates were inoculated with phage and incubated as before. The host cultures were all prepared using a 0.1 per cent inoculum from 24 hour broth cultures of E. coli B except for one 4 hour culture as indicated in the table where a 1 per cent inoculum was used. Viable counts were performed on each host culture used and the results of these and the plaque counts are shown in Table 2 Table 2 The effect of a e and densit of the host culture on la ue counts.
Host culture Age(Hrs) 4 L:K.i 6 10 12 18 24 24 24 48 Viable 1 ccunt i (X1081per. 0.311'2 1.3 1.8 2.7 6.4 5.0 7.9 6.6 7.2i ml) ------Plaque count Mean plaques per 8.92 12.37 11.95 12.92 14.00 14.15 13.27 13.50 14.271227 drop
As per cent of highest 62 87 84. 90 98 99 93 95 :00 86 count.
2.?1 culture produced from 1 per cent inoculum of a 24 hour broth culture of E. coli B
Apart from the first "4 hour" culture, where the viable count is substantially lower than in the remainder of the cultures, no obvious relationship exists between the plaque count and the viable count or 59- age of the host culture. Omitting the first "4 hour" result an Analysis of Variance was carried out on the mean number of plaques per drop calculated from each individual plate. The variance due to 'Llhe use of different cultures was not significantly greater than the variance between replicate plates (F:7 = 1.38). It was concluded that, while the use 8 of a host culture containing less than 1 x 10 viable cells per ml. can result in an appreciable decrease in the plaque count obtained, the variation in the viability normally found in 18 to 21 hours daily broth cultures of the host has no significant effect on the plaque count (see page 18 ). Time Flooded. Five batches of 5 plates were inoculated using a 24 hour broth culture of E. coli B and the standard technique except that the time for which the plates were allowed to remain flooded was varied for eaoh batch. The times used were 10 seconds, 30 seconds, 1 minute, 5 minutes and 10 minutes. After inoculation and secondary drying, drops of phage suspension were placed on each plate as before and the plates incubated.
The 10 minute plates were not satisfactory since they had absorbed enough of the culture to produce the same effects as are seen on a plate which has not been "overdried". The remaining plates produced satisfactory localisation of the drops and discrete plaques. The mean number of plaques per drop on the 10 second, 30 second, 1 minute and 5 minute plates were 11.24, 11.32, 11.08 and 13.25 respectively. .An
Analysis of Variance on the mean plaques per drop calculated from
each individual plate showed that the variance due to the different treatments was not significantly greater than the variance between replicate plates (Ff0 = 2.98). 6o.
A flooding time of about 1 minute therefore seems to be suitable and slight variations in this time are not likely to produce variations in the plaque count. Storage after inoculation with host. Three batches of 5 plates each were inoculated with E. coli B by the standard technique. One batch was inoculated with drops of a phage suspension as before after standing at room temperature for 20 minutes and the other two batches after standing for 3i hours and 61 hours respectively. After incubation the mean plaques per drop on the 2C minute, 3'i hour and 6i hour plates were 19.55, 21.90and 22.20respectively. An Analysis of Variance of the mean plaques per drop from individual plates showed the variance between batches to be not significantly greater than the variance within batches (F13 = 2.36). A similar experiment was performed on plates stored in the cold after inoculation with the host. Three batches of 5 plates each were inoculated with E. coli B using the standard technique at intervals over a period of
8 days and the plates stored at 5°C until used. On the eighth day each batch of plates, now 8 days, 5 days and 1 day old respectively, were transferred to an incubator and held at37°C for 30 minutes in order to dry off the surface film of moisture collected during refrigeration. The cold stored plates, together with a batch freshly inoculated with host, were then inoculated with drops of a phage suspension as before and incubated. The mean number of plaques per drop for the four batches of plates was 15.00, 15.65, 16.50 and 15.00 respectively and an Analysis of
Variance again showed that the between batches variance was not significantly greater than the variance within batches (46 = 1.3).
61,
It was concluded that a batch of plates, after inoculation with the hostp could be stored on the bench for up to 6 hours or in the refrigerator for up to a week with no appreciable effect on the ability of phage
particles to produce plaques on the plates. (iii) Reliability of Plating Technique. The reliability of bacterial counts obtained by plating samples of a bacterial suspension can be assessed by testing the agreement of the counts with the Poisson distribution (Fisher 1948). The agreement is tested by calculating the index of dispersion
2 S(x X)2 mt
for the counts where x is the number of colonies produced by any sample and 2 n is the number of samples plated. If the value of X calculated is less than the tabulated value for a level of probability (P) of 5 per cent with (n - 1) degrees of freedom then the agreement of the counts with the theoretical distribution is acceptable and the experimental conditions can be considered to be satisfactory. The mean count of the samples from
the bacterial suspension can then be accepted as a reliable estimate of the viable count of the suspensions. The accuracy of this estimate can be calculated since, provided the product nx is larger than about
30 (Bailey,1959), the standard error of the mean count is + yr. 2 It should be noted that thelk test is not satisfactory when x is less than about 5 and, since the test is insensitive for low values of n, a minimum of 5 replicate samples is usually considered necessary. 2 TheX test has been used in the work reported here as the basis of the assessment of the reliability of the plating technique used in the 62. performance of plaque counts. Ijnj_Lauz'coneLitionsininaiviallatesancire1iabilitofnitof sampling technique. The uniformity of the conditions for plaque formation existing within individual plates together with the suitability of the technique of sampling of a phage suspension to be plated was tested by inoculating each of a batch of 10 standard plates with 12 single drops of a broth suspension of phage containing about 1000 particles per ml. Two further batches of 10 plates were inoculated in the same way using phage suspensions containing about 600 and 300 particles per ml. After incubation, the number of plaques produced from each drop of the suspensions was counted and the )(2 calculated for the counts on each individual plate. The results are shown in Table 3 Table I. The Uniformit of conditions in individual plates and the reliability of samoliEg technique. phage 10 Replicate plates, 10 single drops of phage suspension Suspensions on each.
; 18.7 17.9 17.7 16.3 18.1 17.2 18.4 16.8 16.8 18.7 )(2 13.0 12.6 14.1 8.3 10.9 10.8 7.5 5.1' 2.8 11.5 2 i 10.0 11.8 11.8 11.2 10.0 11.9 11.3 10.8 11,0 11.2 7.2 13.4 11.8 14.2 10.0 13.3 13.8 )(2 9.3 16.7 8.9 3 R 5.1 5.8 5.0 5.7 4.3 5.6 5.8 5.0 4.8 5.1 12.0 12.5 6.4 9.7 11.8 5.9 10.1 9.6 10.3 8.5 V Tabulated )((for P = 0.05 and d,f. = 11) = 19.7 x2(for P = 0.10 and d.f. = 11) = 17.3
The values of V obtained show a close agreement of the plaque count per drop with the Poisson distribution for each individual plate. This b3. agreement indicates that the method of preparation of the plates provides satisfactory conditions for the production of plaques, that the conditions are uniform within each plate and that the sampling technique is satisfactory. Using all 120 values of plaque count per drop available for each of the three suspensions the mean plaque count per drop,;(, standard error of the mean and 95% confidence limits of the mean were calculated for each suspension. The calculations are summarised in Table 14. . T2,322.e xiatioh232.escunts to the Poisson distribution with the confidence limits of the mean count for 3 phage suspensions.
Phage 2 Standard 95% 95% limits suspension x R sx2 (sxA error of limits as per ) mean of mean, cent of +/x/n t=1.98 mean. 4. 1. 21144 17.87 40226 38306 107.4 - 0.38 !0.76 - 4
2. 1339 11.16 16135 14941 107.0 0.30 2'0.60 + 5 + + + 3. 625 5.21 3811 3255 106.7 0.21 -0.41 - 8
Each of the )(2 obtained here has 119 degrees of freedom which is beyond the scope of the )(2 tables but substitution of the values in the expression.44e-hd.f - 1 gave values for "d" of 0.46, 0.42 and 0.42 respectively. These are appreciably lower than 1.645, the value of td' for P = 0.05 in the Normal Distribution Table, which indicates that conditions for plaque production are satisfactory and uniform over all the plates in each batch. The size of the mean plaque count per drop has no obvious effect on 2 the value of ): obtained and it is concluded that the method of plating is reliable for plaque counts on suspensions in which the mean number 64..
of phage per drop is 5 or more (315 per ml. or more).
Reliabilit of routine counts,.
The results of all routine plaque counts on phage suspensions were 2 2 subjected to a )( test and only when the X corresponded to a level of
probability of not less than 0.05 was the count accepted.
A count was not considered reliable unless the mean number of plaques
per drop of suspension plated was between 5 and 25; the lower limit being 2 set by the limitations of the X test and the upper by the number of
plaques per drop area which can be counted with accuracy. Not less
than 20 single-drop samples were taken from each suspension plated, )(2 this level of replication making the test reasonably sensitive and
giving a percentage standard error of the mean count of from 10 per
cent (for a mean of 5) to 4.5 per cent (for a mean of 25).
In order to minimise the effect on a count of any one plate in a
batch being grossly different from the rest, the drops plated for each
count were distributed between 5 plates. Commonly, L. drops from each of
3 separate phage suspensions were plated on each of 5 plates to give the required 20 samples from each suspension. 2 While the use of the )( test on a single count is a useful test
of the reliability of that count, a more stringent test of the reliability of the counting technique generally is obtained by testing X2 the "Goodness of Fit" of the distribution of the obtained from
a large number of counts with the distribution expected from random
samples of a Poisson series (Fisher et a1,1922). This test was applied 2 to the )( obtained from 100 routine plaque counts in each of which 20 2 samples had been plated. The number of )( expected to fall into each 65. range of values was taken from Elderton's Tables of (2 (Elderton 1902) and the comparison is shown in Table 5.
Table . "Goodness of Fit" of the distribution of X 2 in la ue...TOOIOLV•201C counts with the distribution e ected from random samples of a Poisson series.
2 Observed Expected 0 - E (0 - E) /E (0) (E)
less than 12 14 11.44 +2.56 0.573 12 - 15 20 16.30 +3.30 0.84.0
15 - 18 16 20.02 -4.02 0.807 18 - 21 20 18.56 +1.44 0.119 21 - 24. 8 14..06 -6.06 2.612 24 - 27 9 9.15 -0.15 0.002 27 - 30 6 5.29 +0.71 0,095 Over 30 7 5.18 +1.82 0.639 Totals 100 100.00 0.00 5.68C 2 When X. = 5.68 with 7 degrees of freedom, P = 0.5-0.7 2 The fit of the distribution of the observed values of X to the distribution expected for samples from a Poisson series is good. This
indicates that in routine plaque counts the technique used for preparing the plates, plating the phage suspension and counting the resultant plaques is sound Lnd that the plaque counts obtained are reliable,
LL Preare -enofdilatj22L2ns 8 t9Ls In the performance of a plaque count on a phage suspension it is
normally found necessary to prepare a series of dilutions of the
suspension and to plate out an aliquot portion of the final dilution in
order to obtain a number of discrete plaques suitable for counting. 66.
(i) The diluent
It is common practice in performing bacterial counts to use solutions of eleo+.rolytes such as normal saline or Ringer's solution as diluents
for bacterial suspensions. Phage, however, are generally inactivated in s'ich solutions and all the T group ooliphages have been shown by Adams (1948) to be rapidly inactivated at the air/liquid interface when suspended in a saline buffer solution and shaken vigorously. He further
showed that the inactivation could be prevented by the addition to the medium of a fewildml. of protein. To minimise the possibility of
inactivation of phage during the preparation of dilutions for plaque counts these dilutions were invariably prepared using peptone broth as the
diluent. The remarkable stability of the phage particles in this medium is shown by the tests discussed later under "Storage of Phage Stooks". (ii)Method of preparation of dilutions Dilutions ware prepared by a process of serial dilution taking S drop samples of original phage suspension or dilution thereof and adding them to 100 ml., 30 ml. or 10 ml. volumes of broth. Since the volume of one drop of broth is 1/63 ml. the dilution ratios thus produced are 1:1051, 1:316 and 1:106. Suitable combinations of these dilution steps were used so that at least one of the dilutions finally plated would contain an average of between 5 and 25 phage particles in the 1/63 ml. volumes dropped on to the plates. Each dilution was shaken vigorously for 20 seconds before being sampled, and in order to minimise the effect of any adsorption of phage to glass (Puck et al91951), the dropping pipettes were charged and emptied three times before being used to deliver a sample. The 100 ml. and 30 ml. volumes of broth were measured with an 67, accuracy of ± 0.5 ml. into screw-capped bottles using an automatic filler and were then sterilised and stored until required.
The 10 ml. volumes were measured aseptically using a pipette with an accuracy of ± 0.1 ml. into glass-stoppered tubes when required.
On occasion, dilutions were prepared by adding 1 ml. samples, measured using a 1 ml. graduated pipette, to 100 ml. or 9 ml. volumes of broth. This technique was found generally less convenient than the dropping pipette techniques because of the larger volumes of the initial phage suspension and the greater number of dilution steps required.
(iii) Reproducibility of dilution technique
In order to estimate the reproducibility of the dilution technique five separate plaque counts were made on each of a number of phage suspensions of different titers. Various dilution levels were therefore required and the dilutions were made using both dropping pipettes and 1 mi. pipettes with appropriate volumes of diluent. In each count L. separate standard drops of the final dilution were placed on each of
5 plates.
After incubation the mean number of plaques per drop of the final dilution was calculated for each individual count. From these means the mean plaque counts for the original suspension was then calculated for each group of 5 replicate counts together with the 95 per cent confidence limits of this mean count, the limits being expressed as percentages of the appropriate mean for ease of comparison of the variation shown in the dilution tehniques. The results are shown in Table 6.. The reproducibility, as expressed by the per cent confidence limits, is considered to be satisfactory for all the dilution techniques tested.
fiable 6. Reproducibili of plaque counts obtained by various dilution techniques.
Preparation of Dilutions Dilution Mean number of Mean plaque 95 per cent plaques per count, as limits of Sample Diluent Number Factor drop (1/63 ml) phage per mean size volume of of final ml. of plaque dilutions dilution:- original count as with each ION suspension. percen:-age diluent
NS of the volume for each Grand Mean of five mean count.
USPE separate S counts. , 8 1. 6 drop 100 ml 2 1.17x10 11.45 10.85 10.6 11.07 8.11x1010 +8 10 ml 1 11.40 11.05 - 1. 1 ml 100 ml 1.04x10 10.85 10.50 11.55 11.15 7.30x1010 +5 . 11.30 11.55 1. 1 ml 9 ml 8 8 in 1.00 x10 10.05 12.35 10.85 11.50 7.24x10-w +11_. 11.70 12.55 ...... 1 10 2. 6 drop 100 ml 2 1.17x10 5.80 5.50 5.55 5.85 4.3 x10 +8 10 ml 1 5.95 6.45 8 10 2. 6 drop 100 ml 1 3.52x10 19.25 18.70 .17.55 19.10 4.24x10 +4 30 ml 1 19.00 21.00 10 ml 1
3. 6 drops 100 ml 2 1.10x10 11.95 10.40 11.25 11.81 8.18x10 +12 13.55 11.90 .. Table 6. Reproducibility of plaque counts obtained by various dilution techniques. (contd.)
Preparation of Dilutions Dilution Mean number of Mean plaque 95 per centi plaques per count, as limits of Sample Diluent Number of Factor drop (1/63 ml) phage per mean size volume dilutions of final , ml. of plaque with each dilution:- original count as
ON diluent suspension percentage 1 I volume of the meal NS for each !Grand:Mean E of five count separate SUSP counts.
drops 30 ml 2 1.00x105 12.50 9.40 11.85 11.47 7.23x107 +13 12.20 11.40 6 5. 6 drops 10 ml 2 1.12x104 9.40 10.65 11.40 10.71 7.56x10 +10 10.60 11.50 4 5. 1 ml 100 ml 2 1.02x10 12.90 13.40 11,90 12.68 8.15x10 +6 1 12.25 12.95 4 6 55. 1 ml. 9 ml 4 1.00x10 12.60 12.80 11,75 12.67 7.98x10 +7 13.80 12.40 odeesaamar-,.ar-1,...car.a' 2 4 6. 1. ml 9 ml 2 1.00x10 13.10 12.40 11.95 12.58 7.92x10 +6 13.35 12.10 ____ 70.
No relationship is obvious between the method of preparing the dilutions, the number of dilutions involved or the titer of the suspension finally plated and the reproducibility of the count obtained.
The close approximation of the estimates of the plaque counts on those suspensions diluted by more than one procedure (suspensions 1, 2 and 5) furnishes further evidence of the reproducibility of the dilution techniques involl d.
The variations between replicate counts shown here are considered to represent the limits of accuracy of the counting melbalgenerally and to incorporate the variations due to (a) measuring sample volumes
(b) measuring diluent volumes (c) mixing the dilutions (d) sampling the dilutions and (e) counting the plaques produced.
(d) Time of incubation of plates.
In 10 routine plaque counts the mean number of plaques per drop of suspension plated was determined after incubating the plates at 37°C for
6 to 7 hours and again after further incubation either at 37°C for 14. to 24 hours or at room temperature for 72 hours. The mean number of plaques per drop in the 10 counts ranged from 9.2 to 22.6 and in no case did the mean obtained in the first count differ from that obtained in the second count by more than 2 per cent of the first count. The second counts. were sometimes larger and sometimes smaller than the first counts, an increase in count being explained by the late emergence of some plaques and a decrease by the fusion of closely adjacent plaques on prolonged incubation. Another possible source of variation between the two counts is the personal error in counting the plaques but as the largest number of plaques on any one drop area was 32 this source of 71. variation was considered to be of no practical significance.
The results indicate that the great majority of the plaques produced in a plaque count are visible after about 6 hours incubation at 0 37 C and that more prolonged incubation (up to a total of about 30 hours at 57°C) will result in counts which are not significantly different from those obtained after 6 hours.
3. Preparation of phage stocks.
Phage may be cultivated in cultures of the host either on a solid or in a liquid medium. Each of these methods were investigated with the object of determining which of them could best be adapted to the production of high titer stocks of T6r coliphage with reasonable convenience using the facilities available.
(a) Cultivation in a livid medium.
The method used consisted of inoculating broth with the bacterial host, adding phage in amounts which would ensure an Initial excess of the host cells, incubating the culture until the maximum phage titer had been attained and then removing bacterial cells and debris by differential centrifugation and filtration leaving a suspension of phage in broth.
A series of experiments were performed to examine the effect of the conditions of inoculation and cultivation on the yield of phage by this method.
(i) ha1122EELL202211121221p host cells in inoculum
100 ml. volumes of peptone broth contained in 250 ml. plugged conical flasks were inoculated with 1 ml, of a 21. hour broth culture of 6 E. coli B. The inoculated broth was then assumed to oontain i- x 10 72.
viable cells per ml. Dilutions in broth of a phage suspension were
prepared and 1 mi. of one of these dilutions added to each of the volumes of broth already inoculated with bacteria. The concentration of phage particles in the dilutions was such that the ratio of phage particles to bacterial cells initially present in the cultures was
1:1, 1:10, 1:1x102, 1:1x103, 1:1x104 or 1:1x105. The cultures were incubated at 37°C and 5 ml. samples withdrawn from them at intervals. Each sample was centrifuged at 3,000 r.p.m. for 30 minutes to remove the majority of the bacteria present and a plaque count carried out on the supernatant. The counts obtained, therefore, represented fairly accurately, the number of free phage particles present and could be expected to include relatively few plaques produced by infected bacteria not yet lysed. The results are shown in Table 7.
Table 7. The effect of varying the ratio of phage particles to host cells on the yield of phage obtained.
Initial Ratio Expt. Initial Concentration of phage (x 109 of concentration particles/M1.) after incubation Phage:Bacteria of phage as for:- particles/ml. 4. hrs. 8 hrs. 24. hrs. 48 hm 6 1:1 1 4 x 10 2.4 2.6 3.5 1:10 1 4. x 105 3.9 3.9 5.1 1:1 x 102 1 4 x 104 6.8 8.2 8.6 1:1 x 102 2 4 x lo 8.1 9.6 9.5 1:1 x 10 29.24 x 103 11.0 10.9 1:1 x 103 3 4 x lo 7.o 8.1 6.8 2 1:1 x 104- 3 4 x 10 7.7 8,7 7.7 1:1 x lo5 3 4 x 10 8.6 11.0 8.3 73.
It was concluded from these results that, with a constant size of inoculum of bacterial cells, the initial ratio of phage particles to bacterial cells had little effect on the final titer when the ratio 2 used lay between 1:12: and 1:1x10 and that the final titer was reached with 4. hours incubat- (ii) Size of inoct, m of host cells Two cultureF of phage were prepared by adding 1 ml. of a 24 hour broth culture of E. coli B to one 100 ml. volume of peptone broth and 10 ml. of the culture to another. To each was added enough phage to give an initial ratio of phage to bacteria .of 1:1x103. The cultures were sampled and plaque counts performed as before. The experiment was repeated on another occasion using 300 ml. volumes of media in 500 ml. flasks and increasing the volumes of inocula proportionately. On this occasion the cultures were counted only after 6 hours incubation. The results of these experiments are shown in Table 8. Table 8. The effect of size of inoculum of host cells on yield of phage obtained.
Size of Initial Concentration of phage (x109 particles,/ inoculum of concentration ttI m2.) after incubation for:- bacterial cells, of phage as (viable cells per particles per 6 hrs. hrs. 24 hrs. 4.8 hrs. ml. of culture). ml. 9 in 100 ml.medium 6 4- x 10 4- x 103 9.2 10.0 11.0 10.0 4 x 107 4 x 104 4.8 8.0 10.0 9. 6 in 300 ml, medium 6 4. x 10 4. x 103 5.9 4 x 107 4 x 104 8.5 74. A tenfold increase in the size of the bacterial inoculum therefore resulted in no appreciable increase in phage titer.
(iii)Composition of the medium
100 ml. volumes of three different media were inoculated with
1 ml. of a 24 hour broth culture of E. coli B and enough phage to give an initial phage to bacteria ratio of 1:1x103 The three media used were normal peptone broth, a digest broth and Nutrient Broth, Burroughs Wellcome. The cultures were incubated and plaque counts performed on samples withdrawn from them as before. The phage titers obtained are shown in Table 9. Table 9. The effect of composition of the medium on the phage yield obtained.
Medium Concentration of phage (x109particles per ml.) after incubation for:-
4 hrs. 8 hrs. 24. hrs.
Peptone Broth 5.9 6.9 5.6 Digest Broth 6.7- 6.4 6.5 , I Nutrient Broth, 1 Burroughs Wellcome 5.1 4.6 5.1
No appreciable difference in phage titer was found with the three media used.
(iv)Effect of Aeration of Cultures All of the investigations on cultivation of phage in liquid medium so far described were carried out without aeration of the medium. Since E. coli B could be grown to a concentration of 3 x 109 viable cells per
75. 8 ml. in aerated broth compared to the 4 x 10 cells per ml. attained in static cultures, it was thought that aeration of phage cultures might result in higher yields of phage. 300 ml, volumes of broth in 500 ml. flasks fitted with wash-bottle heads were inoculated with 3 ml. of a 24. hour culture of E. cell B and enough phage to give initial ratios of phage to bacteria ranging from 1:10 to 1:1x105. As an anti-foam agent, 0.3 ml. of a sterile 1 per cent solution of capryl alcohol in arachis oil was added to each flask, Air from a compressed air line was sterilised by passage through a cotton wool filter, saturated with water vapour by bubbling through water and
then fed into the depths of each culture. The air flow was adjusted so
as to keep the cultures vigorously agitated throughout the incubation period and the flasks maintained at 37°C by immersion in a water bath. lie cultures were sampled at intervals, and plaque counts performed as
before. The titers obtained are shown in Table 10. Table 10. The effect of aeration of the culture on the Yield of phase obtained.
Initial ratio of Initial Concentration of phage (x109 particles phage:bacteria concentration per ml.) after incubation for:— of phage as particles per 3hrs. 5 hrs. 8 hrs. 24. hrs. ml.
1:10 4 x 105 1.7 2.1 2.1 5.0 2 1:1x10 4 x 104 2.3 11:1x103 4 x 103 3.2 6.3 7.1 7.6
1:1x105 4 x 10 4. 4 8.4 6.4 11.0
Aeration of the phage cultures therefore produced phage tit's of the same order as those obtained in the static cultures. 76.
(v) Routine production in liquid culture On the basis of these investigations, the routine production of phage in liquid culture was carried out by inoculating 300 ml. volumes
of broth contained in 500 ml. plugged flasks with 1 per cent v/V of an 18-24 hour broth culture of E. coli B, adding 1 per cent v/v of a 5 broth suspension containing 4 x 10 phage particles per ml. to give an initial ratio of phage to bacteria of 1:1000, incubating at 37°C for 6 to 8 hours without aeration and removing bacteria and debris by
centrifugation and filtration as described later. Using this procedure,
stocks of phage containing 4-9 x109 infectious particles per ml. were regularly obtained, these yields comparing favourably with those obtained by Putnam et al (1949) in the cultivation of coliphage T6 by
a similar procedure. (b) Cultivation on a solid medium The method used for cultivation of the phage on a solid medium was that described by Hershey et al (1943). This consists, essentially, of surface seeding agar plates with a mixture of E. coli B and phage in a ratio which will give barely confluent lysis when the plaques have fully
developed. After incubation the phage are harvested by flooding the plates with broth, allowing them to stand for 10 to 20 minutes, decanting
the broth and removing bacteria nna debris by centrifugation and filtration. The method applied was as follows: 20 ml. peptone agar plates were
dried by incubation at 37°C overnight. A 24 hour broth culture of
E. coli B was inoculated with enough phage to product* 4. x 105 particles per ml. of and 3 ml. of this culture added to each plate. After standing for about 1 minute, the excess culture was removed from 77. the plates by inverting and draining them at an angle for 10 minutes the volume of culture retained by each plate being about 0.5 ml.
Incubation at37°C for 4 to 5 hours produced almost confluent but incomplete lysis of bacterial growth on the surface of the plates.
Each plate was then flooded with 3 ml. of broth, allowed to stand for 20 to 30 minutes and the broth drained off. Bacteria and debris were removed from the phage suspension so obtained by centrifugation and filtration as described later.
By this procedure it was found that between 2.2 and 2.6 ml. of
phage suspension per plate could be collected; the phage titer varying 10 from 6 to 9 x 10 infectious particles per ml.
During the harvesting of the phage considerable amounts of soluble
material are extracted from the agar. The amount of soluble solids
present in phage suspensions obtained as above was determined by
evaporating 5 ml. volumes of the suspensions to oonstant weight in a vacuum dessicator over silica gel and found to be 2,38 per cent weight in volume compared with a soluble solid content of 1.54 per cent
obtained for a sample of broth treated in the same 'way. The sodium
chloride content, determined by the method of assay of the British
Pharmacopoeia 1958, for phage suspension and broth were 0.75 and 0.53
per cent respectively. Harvesting of the phage was, therefore, carried
out latterly by adding 3 ml. of sterile water to each plate in place of
broth. The phage suspensions so obtained contained 1.54 per cent soluble
solids and 0.50 per cent sodium chloride — and were therefore equivalent
to suspensions of phage in normal broth. The volume per plate and phage
titer in the final suspension obtained with "water harvest" were identical to those obtained with "broth harvest". 78.
In the original application of this method of cultivating phage on a solid medium, Hershey et al (1943) obtained titcrs of 12 about 1 x 10 phage particles per ml. using coliphage T2 (Demerec
and Fano,1945) and believed that the high yields obtained were due partly to the negligible loss by readsorption of the phage, ooliphage
T2 being characterised by a slow rate of adsorption to its host. Swanstrom and Adams (1951) investigated some of the variables affecting
the phage yield when a modification of this method, the agar layer method, was used for the production of stocks of all seven of the T—coliphages. Using certain strains of each of the seven T—ooliphages, 11 12 they obtained yields of from 1 x 10 to 1 x 10 phage particles per ml. but pointed out that different strains of the same phase may give different yields. It is possible that some increase in phage yield could be obtained with the strain of T6 coliphage and the method of cultivation used here by further adjustment of the conditions of 10 cultivation. The average titer of 7 x 10 phage per ml. obtained, was, however, considered satisfactory for the present work. () Separation of phage from insoluble matter. Bacteria and insoluble debris were removed from the crude lysates
obtained from both methods of cultivation by centrifugation followed by filtration. The lysates were normally processed in batches not exceeding 300 ml. in volume. Each batch was distributed into 75 ml. centrifuge tubes and centrifuged horizontally at 7,000 r.p.m. in swing buckets for 20 minutes. The upper half of the supernatant in each tube was then removed, the remaining fractions decanted, mixed and transferred to fresh tubes and the process repeated until all the 79.
remaining lysate was centrifuged in one tube. The lowest 25 ml.
fraction in this last tube was normqlly discarded. After
centrifugation the lysates were passed through a sterile 5/3 sintered
glass filter and a 2 ml. sample of the filtrate incubated at 37°C for
24 hours to test for freedom from bacteria.
On several occasions plaque counts were performed on the lysatea
before and after filtration. It was found that, provided most of the
bacteria were removed from the lysate before filtration, the loss of
phage on filtration never exceeded 10 per cent of the titer before
filtration. If centrifugation was inadequate, filtration became
extremely slow and up to 90 per cent of the phage could be lost on the
layer of bacteria deposited on the filter bed.
4_ Storage of phage stocks.
Stocks of phage were stored as suspensions of the phage in broth
contained in glass-stoppered bottles and kept at 5°C. Stored in this way
no decrease in the phage titer of the suspensions was detected over
periods of up to two years.
The stability of the phage at room temperature is illustrated by an
experiment in which suspensions of phage in broth and in 1 per cent
dilutions of broth in water, each suspension initially containing 8 1 x 10 phage per ml., were stored at ambient room temperature
(20 - 25°C) for 32 weeks with no detectable decrease in titer. After
87 weeks the titers had decreased by not more than 50 per cent of their initial value.
Storage of broth suspensions of phage at 37°C showed the phage 30„ to be remarkably stable even at this temperature. A suspension 10 containing 1 x 10 phage per ml. showed no decrease in titer after being maintained at 37°C for 7 weeks. More dilute suspensions were rather less stable and the titars of suspensions initially containing 7 2 2 1 10 , 7 x 10 and 4 x 10 phage per ml. were reduced by approximately 50 per cent in 10, 9 and 9 weeks respectively. It was concluded from the results of these storage tests that any inactivation of the phage found in the subsequent tests with antibacterial agents will certainly not be due to any inherent instability of the phage. 81.
E. INHIBITION OF THE ACTION OF THE PHAGE ON ITS HOST.
When broth is inoculated with E. coli B and a number of T6 phage
particles approximately equal to or less than the number of host cells
in the inoculum, the lytic action of the phage on the host is shown
by a characteristic sequence of changes in the opacity of the culture on incubation. The culture first becomes turbid, then decreases in
opacity until completely clear, remains clear for a time and finally
becomes turbid again. The first turbidity is produced by the growth
of initially uninfected host cells but as phage particles are produced
by the lysis of infected cells in sufficient numbers to infect the remaining host cells, lysis of these cells occurs and the culture clears.
It then remains clear until the phage resistant mutants of the host have multiplied sufficiently to produce the second turbidity. If sufficient phage are introduced into the culture initially to ensure that all the sensitive host cells are infected, then the first turbidity is not
produced and the culture remains clear until the growth of the mutants manifests itself,
The effect on the ability of the phage to infect its host of concentrations of antibacterial agents at which the host will grow was studied by observing the effect of sub—bacteriostatic concentrations of the antibacterials on the typical lytic sequence of changes in phage cultures.
1. The effect of inocula size on the rate of action of the phage on its host.
The effect of the size of inocula of phage and of E. coli B on the C2. rate of growth of the host cells and on the time of incubation required to produce visible lysis in the culture in the absence of any antibacterial agent was determined by the following procedure. Ten ml, volumes of broth were inoculated with 6-drop samples (0.095 ml.) of broth suspensions of phage cf various titers so as to produce a range of phage inocula of from 1x108 to.0.5 infectious particles per ml. of culture (shown in Table 11.). To these suspensions were added one-drop (0.016 ml), 6-drop or 1 ml. samples of a 24. hour broth culture of E. soli B, or 3-drop samples of a 4.8 hour culture giving 6 6 host inocula of approximately 7 x105,.4x10 , ha107 and 3x10 viable cells per ml. of culture. Each combination of phage and host inoculum was prepared in quintuplicate and three control tubes without phage were inoculated with each of the host inocula. The number of phage particles in the cultures with the two weakest phage inocula was estimated by plating 10 separate 6-drop samples of each of the two suspensions used for these inocula. The mean nu►bers of plaques produced per 6-drop sample were 5 and 10 respectively•but the numbers ranged from 3 to 8 and 6 to 15 respectively. The phage inocula in the other cultures were estimated by normal plaque counts on the suspensions used for the inoculations. All the cultures were incubated at 3700 and the opacity of the phage cultures compared visually with that of the control at hourly intervals for the first 8 hours then again after 22 and 30 hours incubation. The times of incubation required for the appearance of turbidity in the host control cultures and for lysis of the host to become apparent in the phage cultures are shown in Table 11. 83.
Table 11. The effect of inoc size on the rate of lysis of E. coli
Phage inoculum Time of incubation (hours) to show lysis with host (infectious inoculum of: — particles per (approximate number of viable cells per ml. of ml. of culture) culture)
4 7 x 105 ' 4.A10 4 x 103 x 10 no complete no complete no complete no comfHP lysis lysis lysis lysis lysis lysis lysis lysis 1 0.5 5 7 4 8 8 22 4. 7 1.0 4 7 4 7 8 22 4 6 10.0 4 5 3 4 8 22 3 4 3 1 x 10 4 5 3 4 7 22 1 x 105 3 4. 2 3 5 22 7 i 1 x 10 2 3 2 4 5 8 1 x 10 1 2 0 1 1 2
Time of incubation (hours) for appearance of 1 — 2 0 — 1 0 ..:. 1 0 — 1 turbidity in control host cultures
(t inoculation from 4.8 hr. culture of host) The first time quoted in each case is the maximum time at which the opacity of all 5 replicate phage cultures matched that of the controls, the second is the minimum time at which the opacity of all the replicates was less than that of the controls. The presence of phage in the cultures with a phage inoculum of 8 6 1 x 10 particles per ml. and host inocula of 7 x 105 and 24. x 10 cells per ml. was no longer apparent after 22 hours incubation since their opacities then equalled that of the host control cultures. The opalities of all the other phage cultures remained less than that of the controls for at least 22 to 30 hours.
As expected, the rate of growth of E. coli B is affected by the size of the bacterial inoculum and the rate of lysis of the culture is affected by the size of both the bacterial and phn.ge inocula. The difference in rate of lysis produced by small variations in inoculr size is, however, relatively small and the rate of lysis was not appreciably affected by the use of a 48 hour host culture as inoculum in place of a 24 hour culture. Using a host inoculum of 6 crops of a 24 hour culture of E. coli B, a volume which is conveniently and accurately measurable, the presence of phage initially present in 8 numbers ranging from 1 x103 to 1 x 10 per ml. of culture can be detected by exnmining the cultures after 4 to 8 hours incubation and numbers of from 0.5 to 10 per ml. of culture by examination after 8 to
30 hours incubation. Small variations in the viability or age of the host inoculum will have no appreciable effect on the rate of lysis of the culture at least within the limits of precision of the measurements used here.
2. The effect of antibacterial agents on the action of the phage on its host.
With each of the eleven antibacterial agents, tests were carried out
(a) to determine the concentration of the antibacterial which inhibits the growth of the bacterial host in peptone broth (the bacteriostatic 85,
concentration) and (b) to investigate the ability of the phage to infect,
and cause the lysis of, the bacterial host in peptone broth containing concentrations of the antibacterial at which the host cells can grow. (a) Determination of bacteriostatic concentrations. The procedure followed was to prepare, for each antibacterial, a series of tubes of peptone broth containing a range of concentrations of the antibacterial by adding from 1 to ml. of different strength
solutions of the antibacterial to 5 ml. volumes of double-strength
peptone broth and adjusting the volume in each tube to 9.8 ml. with water. Three tubes of media were prepared for each antibacterial at each concentration tested and each tube was inoculated with 6 drops of a 24 hour culture of E. coli B and 6 drops of broth. Each series of cultures included, in addition, 2 control cultures of E. coli B which received the same inocula as the test cultures but contained no added antibacterial agent. The cultures were incubated at 37°C, examined at hourly intervals for 4 hours then daily for a week and weekly for 3 weeks. The concentrations of each antibacterial at which growth of the host occurred after various times of incubation are recorded in Table 12.
The minimum value quoted in the table for each antibacterial/time of incubation is the maximum concentration at which growth was visible in all 3 replicate cultures, the maximum value quoted is the minimum concentration at which no growth occurred in any of the 3 replicates.
The control cultures of E. coli B showed a definite turbidity in 1 hour
and any test culture in which the opacity increased over the first 24 hours at a rate approximately equal to that in the control cultures, the 86. comparison being made by eye,was considered to be showing a normal rate of growth of the host. Table 12. Growth of E.coli B in the presence of antibacterial agents.
Antibacteriail Maximum Concentration of antibacterial agent permitting:- Agent Approx. Visible growth after normal rate of 4 hours 24 hours 2 days growth. AMINACRINE HYDROCHLORIDE < 3 6.4-7.2 7.2-8.0 7.2-8.0
(1) CETRIMTDE (x 10- b M) 2.7-5.5 5.5-11 5.5-27 5.5-27 (xLORAM3 INE T 10- M) <1.8 <1.8 2.5-2.8 2.5-2.8 CHLOROCRESOL (x 1 -4 M) <:. 5.6 8.4,9.8 9.8-11 9.8-14 (2) CHLORXYLENOL 6.4-13 26-32 32-6)- ISc010011\8 32-64
(3) CRESOL (AND SOAP SOLUTION) 18-27 18-27 > 4.6 (x 10-4 m) CRYSTAL VIOLET (x 10-6 M) < 1 5-10 20-50 20-50 FORMALDEHYDE (x 10- 4 M) 2.3-3.0 3.0-6.0 6.0-12 12-18 PHENOL (x 10-3 M) < 5 21-27 21-27 21-27 PHENYLMERCURIC NITRATE (x 10-4 M) < 1 < 1 1.3-1.9 1.3-2.5 toSODIUM LAURYL SULPHATE ( 1.7 1.7-17 17->340 17)340 (x 10-4m) footnote overleaf 87.
Footnote to table on previous page.
(1)Molarity calculated assuming sample to be pure cetyltrimethylammonium bromide.
(2)Molarity calculated as the equivalent of pure chloroxylenol.
(3)Molarity calculated as equivalent of pure cresol.
(4)Molarity calculated assuming sample to be pure sodium lauryl sulphate. No further growth became visible in any of the cultures upon incubation for longer than 2 days. By comparing the concentrations at which bacterial growth eventually became visible with the concentrations at which growth proceeded at a
normal rate or at least became visible in I+ hours, it can be seen that all the antibacterials caused a reduced rate of bacterial growth at concentration approaching the final bacteriostatic concentration. This effect was particularly pronounced with Cresol and Soap Solution, crystal violet,
formaldehyde and sodium lauryl sulphate. With Cresol and Soap Solution no real bacteriostatic concentration could be obtained since visible growth was eventually produced in all the cultures containing the highest concentration tested (=4.6 mM or 0.05% w/v cresol). With sodium lauryl sulphate also, growth occurred in the highest concentration tested
(E34. mM or 1%) but in this case not all the replicates showed growth and 1 or 2 negative cultures were produced at each concentration tested down to 1.7 mM. A similar, but less widespread scatter of negatives occurred with cetrimide where the difference between the concentration showing growth and that showing inhibition in all replicates was fourfold. With
nil other antibacterials a relatively precise bacteriostatic concentration was obtained, the difference between the concentrations showing growth and those showing inhibition being no more than two-fold. (b) Aotivit of ha e in sub-bacteriostatic concentrations of antibacterials. Simultaneously with the baeteriostatio tests, identical tubes of peptone broth containing the antibacterials were prepared and inoculated with 6 drops (0.095 ml.) of a broth suspension of phage containing 6 1 x 10 particles per ml. and 6 drops of a 24 hour broth culture of E. coli B. The inocula therefore consisted of approximately 1 x 104 6 phage particles and 4. x 10 viable host cells per ml. of test culture. Five such cultures were prepared for each concentration of each
antibacterial tested. As in the bacteriostatic tests the cultures were incubated at 37°C, examined at hourly intervals for L. hours then daily
for a week and weekly for 3 weeks. The sensitivity to lysis of each host culture used was confirmed by 2 control phage cultures with the same inocula as the test cultures but containing no antibacterial agent. At each examination, the presence or absence of lysis in the phage cultures was decided by visually comparing each culture with the host cultures containing the same concentration of the same antibacterial agent. When the opacity of the phage culture was appreciably less than that of the host culture (and remained so for about 24 hours) lysis was considered to have occurred. Lysis was found to occur in the presence of all the antibacterials tested, with the exception of sodium lauryl sulphate, at all concentrations at which the host cells could be grown. In the case of
sodium lauryl sulphate, lysis occurred in phage cultures containing
0.34 mM, but not in those containing 0.68 mM, whereas growth of the host 89.
occurred in all cultures containing 1.7 mM.
The design of the experiment did not permit the precise measurement
of the rate of lysis in the phage cultures, but, with the exception of
aminacrine hydrochloride, lysis, where it occurred, was obvious within
not more than 214. hours of the appearance of growth of the host in the
corresponding host cultures. pit the concentrations of cetrimide,
Chloroxylenol Solution and formaldehyde showing a normal rate of growth
of host, phage cultures showed a normal rate of lysis. In the phage
cultures containing aminacrine hydrochloride at concentrations of
5.6 x 10-5 M or greater, lysis was not obvious until 48 hours after the appearance of growth in the corresponding host culture.
'While none of the antibacterial agents, excepting sodium lauryl sulphate, appear to completely inhibit the growth of the phage at concentrations which are sub -bacteriostatic for the host, they all, in concentrations approaching the final bacteriostatic concentration, at least cause a reduced rate of growth of the host cells and a corresponding increase in the time of incubation required to demonstrate the presence of the phage. Therefore, in testing for the presence of infectious phage in the presence of an antibacterial agent it is desirable that the concentration of the antibacterial should be reduced to a level which is not only sub -bacteriostatic for the host, but which will permit the growth of the host cells at an approximately normal rate.
Precipitation was found to occur in peptone broth containing concentrations of cetrimide of 3 x 10-4 M or greater, in dilutions of 3 Chloroxylenol Solution equivalent to 1.5 x 10 gchloraxylenol or greater and in dilutions of Cresol and Soap Solution equivalent to 90.
2.5 x 10-3 M cresol or greater. No precipitation occurred with any of the other antibacterials in the concentrations tested.
3. The effect of size of phage inoculum on rate of lysis in presence of antibacterial agents.
A series of experiments were performed to determine if the presence of both very small and large initial numbers of phage particles could be readily detected in cultures containing sub-bacteriostatic concentrations of the antibacterial agents. In these experiments an examination was made of the rate of lysis produced in phage cultures with 8 inocula of 1.5 x 10 or 0.15 phage particles per ml. of culture and containing concentrations of the antibacterials which (a) were equal to or less than the maximum concentrations at which growth of the host had previously been shown to become visible in 1 to 4. hours, and (b) were such as could be produced conveniently by dilution of the maximum concentrations which it was intended to test for phage inactivation. The procedure used was to prepare for each antibacterial agent, nine 20 ml. volumes of peptone broth containing the concentration of the antibacterial shown in Table 13. Three of the tubes in each series were inoculated with 6 drops (0.095 ml.) of a broth suspension of phage 10 containing 3 x 10 particles per ml., 3 with 6 drops of a suspension containing 30 particles per ml. and 3 with 6 drops of broth. Each tube was then inoculated with 6 drops of a 24. hour broth culture of E. coli B and incubated at37°C. Control host cultures and phage cultures were also prepared as before. The cultures were examined after .14 to 2 hours incubation and thereafter at hourly intervals for 81 to 9 hours, then 91.
Tabaejlate of lysis in phase cultures Containing sub-bacteri ostatic concentrations of antibacterial agents.
TIBACTERIAL Concentration Time of incubation (hours)to show:- GENT, of anti- bacterial Visible growth Lysis in phage agent. of host, in cultures with absence of inocula of (phage (Molar) phage. per ml. of culture)
1.5 x 108 0.15
comp- no 'comp lysis late lysis'lete lysis iNsis IIMINACRINE 1+.0x10-5 2 48 HYDROCHLORIDE 4.0x10-6 b 2 7.5 8.5 CETRIMIDE 1.4x10- 6 0 - 1.5 0 1.5 2.5 20
CHLORAMINE T 8.9x104+ 1.5- 2.5 1.5 2.5 6 20
CHLOROCRESOL 7.0x10-5 0 - 2 0 2 7.5 21
CHLOROXYLENOL (Solution) 3.2x164 0 - 1.5 1.5 2.5 3.5 RESOL (and -I+ oap Solution) 4.6x10 0 - 1.5 0 1.5 2.5 20 CRYSTAL VIOLET 1.2x10-6 0 - 2 0 2 6.5 26 FORMALDEHYDE 3.3x10-4 3.5- 4.5 3.5 4.5 3.5 20 PHENOL 2.7x10-3 0 - 2 0 2 6.5 21 PHENYL- IMRCURIC 1.6x10-7 0 - 1.5 0 2 8 21 NITRATE SODIUM LAURYL 1.7x10-4 0 - 1.5 2.5 3.5 4.8 SULPHATE 1.7x10-5 0 - 1.5. 0 1.5 9 20
Control cultures with no 0- 1 antibacterial agent 92,
Footnote to table on previous pale
The first time quoted in each case is the maximum time at which the opacity of all 3 replicates matched that of the host controls, the second is the minimum time at which the opacity of all the replicates was less than that of the controls. again after 19 to 21 hours, 26 to 28 hours and 48 hours. The times of incubation required for lysis to become apparent are shown in Table 13. At the higher concentration—of aminacrine hydrochloride tested growth was visible in the host cultures within 2 hours but the rate of growth of the host was noticeably less than that of a broth control. Failure to detect the presence of the phage in the lightly inoculated cultures may have been due to delay in lysis caused by slow vowth of the host. At the higher concentration of sodium lauryl sulphate the rate of growth of the host was approximately norml and it seems likely that the action of the phage on its host was being specifically inhibited. At the lower concentrations of these two substances tested and at the tabulated concentrations of all the other antibacterials the rate of growth of the host, as shown by comparison of the opacities of the host test cultures with that of host broth culture, was approximately normal. The rate of lysis in the phage cultures inoculated with 0.15 phage particles per ml. of culture was slightly decreased in the presence of all the antibacterials at the concentration tested except for chloroxylenol where the rate of lysis was normal. No decrease in the rate of lysis could be 8 detected in the cultures inoculated with 1.5 x 10 phage particles per ml. except, again, for chloroxylenol where, in this case, a slight Q3. decrease in the rate of lysis was seen.
There is, however, within the limits of precision of the measurements used in these experiments, no evidence that the concentrations of antibacterials used caused any variation in the rate of 1ys.•s in the phage cultures greater than could be accounted for by differences in the rate of growth of the host cells with the possible exception of sodium lauryl sulphate. 8 The opacity of the cultures inoculated with 1.5 x 10 phage per ml„ remained less than that of the corresponding host cultures for at least
82 — 9 hours and that of the cultures inoculated with 0.15 phage per ml. for at least 26 hours. Using the concentrations of antibacterial agents used here, the presence of phage initially present in numbers ranging 8 from 1.5 x 10 to 0.15 per ml. of culture can therefore be detected by examining the cultures after 4 to 8 hours incubation and again after 20 to 24 hours. 94.
Y. INACTIVATION OF FREE PHAGE BY ANTIBACTERIAL AGENTS. 1. Method of evaluation.
The technique used to evaluate the viricidal activity of an antibacterial agent on the phage consisted of inoculating an aqueous solution of the antibacterial with a standardised number of phage particles, immediately taking a number of samples of the mixture and maintaining the samples at the temperature of the test (25°C). At appropriate intervals of time the samples were diluted with broth to reduce the concentration of the antibacterial to a level at which it was known the phage could normally infect its host and the dilutions inoculated with a standardised inoculum of host cells. On incubation the presence of surviving infective phage particles was shown by the occurrence of visible lysis of the culture. (a)Phage inoculum.
The phage inoculum used in each test consisted of 6 drops (0.095 ml.) of a suspension of phage in broth. All inocula were taken from a single stock suspension of phage prepared by cultivation on solid media as described previously. This suspension was stored at 5°C in glass stoppered bottles and periodic plaque counts performed on it gave a mean count of 4.06 x 1010 particles per ml., the 99% confidence limits of the 10 mean being — 0.29 x 10 . The stock suspension was always shaken vigorously before withdrawing a sample from it and the sample allowed to reach room temperature before the drops were delivered. (b)Preparation of solutions of antibacterial agents.
Solutions of the antibacterial agent under test were prepared in 95.
the required concentration immediately before use as indicated on page 43 Ten ml. of the solution to be tested were pipetted into a glass-stoppered tube which was then partially immersed in a water bath
maintained at 25 (± 0.05)°C and left there for 30 minutes to allow the temperature of the solution toequilibrate with that of the bath.
(c)Inoculation of the antibacterial solution andsamplirig of the reaction mixture.
After allowing the temperature of the solution to stabilise, the phage inoculum was added and the mixture immediately shaken vigorously for 20 seconds. Without delay, approximately 1 ml. of the mixture was then drawn into a sterile dropping pipette and 6-drop samples of it delivered into a number of empty sterile test-tubes whose temperature had previously been equilibrated at 25°C by storage in the bath for 30 minutes and which were immediately returned to the bath. Normally 9 to 11 such.:samples were distributed and the whole operation, from inoculation of the antibacterial solution to the completion of the distribution of samples from the mixture, could be completed in about 2.5 minutes. The time of contact of the phage with the antibacterial was measured from the moment the first drop of phage inoculum entered the antibacterial solution. (d) Testing the reaction mixture for infective phage. At suitable intervals of time each reaction tube was withdrawn from the bath and 20 ml. of broth added to it from a flask fitted with an automatic tilt measure. After mixing the contents of the tube by shaking,
6 drops of an 18-25 hour broth culture of E. coli B were added Prom a dropping pipette, the tube was again shaken and placed in a water-bath 96. maintained at37°C.. At the end of, an experiment the tubes were transfered to an incubator at 37°C and their opacity compared with that of control cultures after 3 to 4 hours, 18 to 24. hours, 2 days, 3 days and 4 days.
The control cultures contained the same concentration of antibacterial agent and received the same host inoculum as the diluted reaction mixtures but contained no phage.
In the majority of the reaction mixture cultures containing active phage lysis was apparent at the 18 to 24. hour examination. The earlier examination was, however, found to be necessary to detect lysis in those cultures where large numbers of phage had survived exposure to the antibacterial agent. In these cultures, lysis of the sensitive host cells and growth of the resistant cells was rapid and after 18 hours incubation their opacity equalled that of the control cultures. The inspections at
2, 3 and IF. days were carried out since a number of cultures showed lysis only after two days incubation. In a number of experiments the cultures were examined at daily intervals for up to 7 days bat in no case was lysis obvious after longer than two days incubation.
The reliability of the periodic examination of the cultures in detecting active phage was confirmed on numerous occasions by placing loopfuls of reaction mixture cultures which had been incubated for 4. to
7 days on the surface of agar plates seeded with host by the method used in preparing plates for plaque counts. On incubating these plates dense bacterial growth was produced on the area of the plates on which the loopful of culture had been placed, but where the culture contained active phage, a clear halo of lysis surrounded this region of dense growth. 97.
The conclusions as to the presence or absence of active phage drawn from the plate tests in all oases corresponded exactly with the conclusions drawn from the periodic examination of the broth cultures. (e) The design of the experiments, The results of the tests for inactivation of free phage have been expressed as the Inactivation Time which is the time of contact of phage and antibacterial agent after which no active phage can be detected in the reaction mixture. Because of the considerable variation in Inactivation Time found for any one antibacterial at any one concentration 5 or 6 replicate determinations were made for each concentration of antibacterial tested. The mean of the individual Inactivation Times were then calculated and the result expressed as the Mean Inactivation Time (M.I.T.) Replication was facilitated by the fact that it was usually possible to carry out several replicate determinations simultaneously, the number of simultaneous replicates possible depending on the range of contact times and the length of time intervals used. In any one group of replicate tests the time intervals at which the samples of reaction mixture were diluted were selected so that, within the 9 to 11 samples taken from each reaction mixture, at least one sample contained active phage and at least the last two samples diluted showed no phage activity. Provided these conditions were satisfied the time intervals were made as short as possible in order to achieve as precise an estimate of the Mean Inactivation Time as possible. Each determination of a Mean Inact.vation Time therefore involved a number of preliminary experiments to obtain an approximate value. The required distribution of samples containing active phage was found to be produced in most cases by using reaction time intervals of approximately 1/8 to 1/10 of the expected M.I.T.
Atypical set of results, obtained from three separate determinations of the M.I.T. of 3 per cent phenol for the phage, are shown in detail in Table 14,
2. Reproducibility of the method.
The level of confidence with which the M.I.T. can be taken as an estimate of the inactivation of phage by an antibacterial agent is shown by the results of a series of replicate determinations of the M.I.T. for five different concentrations of phenol. These experiments are summarised in Table 15.
The individual estimates of the M.I.T. for each concentration of phenol are the results of separate experiments carried out at various times during a period of 3 months.
3. Estimation of the Mean Single Surviver Time. The results of the experiments referred to in the previous section were subjected to the analysis devised by Mather (1949) which permits the calculation of the Mean Single Surviver Time (M.S.S.T.). This time, which is the time at which there is, or the average, one surviving organism per sample volume, is calculated from the regression existing between the time of contact of the organism with an antibacterial agent and the log (-log proportion of negative samples) found at these times. The
M.S.S.T. is the time at which the proportion of negative samples is 0.03679 corresponding with log (-log) = 0. 99.
Table ]4. The determination of the Mean Inactivation Time (M.I.T.) of 3 for 250c. 2theno3ercer1 1 ihaeT6r•at + = active phage present; - = no phage activity.
Experiment Replicate Contact Time (Minutes) Inactivation Test 16 20'24 28 32 36 40 44 481 Time (minutes) I 1 + - 28 2 + - + + - - - - - 32 3 4. + - - + - - - 40 1.4 + + + - + - - - - 36 5 + + + + - - - - - 32
6 + 1 - 0. I.• ....• I... - •••• 24
M.I.T. 192/6 = 32.0 1 II 1 - - - + - - - - - 32 2 41.4 3 + + + + 4. - - - - 36 1 + + - + - - - - - 32 5 24 6 + + - + - - - - - 32 i
. M.I.T. 200/6 = 33. 3 r- III I + + + + + + - - 44 C \ I + + + 4- - - - 32 K 1 + + - - - . _ 24 - . .+ + + + - + -, : - 36 u 1 + 4- + + - ... .. 32 V, + + + + - - - 40
M.I.T. 208/6 = 34.7 1 Table 15. The Mean Inactivation Time (M.I.T.) of various concentrations of phenol for collligm_g1212)
Concentration Contact Number of Range of M.I.T. Mean 95 per cent of phenol. time replicate inactivation (minutes) M.I.T. confidence (per cent w/v) interval. tests in times. (minutes) limits of (minutes) each (minutes) mean M.I.T. determination. (minutes) 1 + 3.50 a 5 2i.- 4. 3.04 3.00 - 0.57 21 -4 3.20 2.-.i 4. .2.75 + 3.25 1 5 9 - 13 10.8 10.3 - 1.80 9 - 12 10.6 8 - 11 9.4 .....,...P. + 3.00 4 6 24 - 40 32 33.3 - 3.79 24 - 44 33 24 - 44 35 + 2.90 5 5 25 - 60 47 47.3 - 6.25 3o - 60 5o 40 - 50 45 2.66 20 5 180 - 240 204. 206 ± 25.4 180 - 240 208 101.
v- FIGURE I. Log-log regression for the inactivation of
coliphage T6r by pheno1,3.00 per cent,at 25°C.
(1B replicate determinations) 1.0 Airline fitted by eye
B=first calculated line
C=second calculated line
0.0
0
-1.0
-2 .0
I 1 I I I
16 20 24 29 32 36 Time of contact minutes) 102.
The data from the experiments are set out in Table 16 in a form suitable for the application of Matherts analysis. The values, y, which are the log (-logs) of the proportion of negative samples (p), are found by reference to Matherts tables, The analysis is illustrated using the results obtained from the inactivation of the phage by 3.00 per cent phenol. The values of y are shown plotted against contact times in Figure 1. A trial line was drawn by inspection and the M.S.S.T. read from the abscissa as 23,6
minutes. Expected values, Y, were read from this line for each contact time and converted to the corresponding expected proportions, P, by reference to Matherts tables. The weighting coefficients, w = P(log P)2 , and the working log-logs, y = y + (p-P) 11101••••••••1110•11•VIM•16.4111ew w 1-P p log P were then calculated for each time x. These, and subsequent calculations are summarised in Table 17. Table 16. The inactivation of coli hake T6r in aqueous solutions of phenol. x = contact times in minutes; p = proportion of negative samples; y = log (-log p)
Concentration Number of of phenol replicates (per cent w/v) • x 2.000 2.333 2.667 3.000 3.333 3.667 3.50 15 p 0.267 0.533 0.600 0.733 0.933 0.800 y 0.278 -0.463 -0.672 -1.169 -2.668 -1.500 x 6 7 8 9 10 11 12 3.25 16 p 0.187 0.313 0.500 0.688 0.562 0.875 0.938 y 0.517 0.150 -0.366 -0.984 -0:551 -2.024 -2.749 x 16 20 24. 28 32 36 40 3.00 18 p 0.055 0.278 0.389 0.444 0.778 0.889 0.889 1.065 0.247 -0.057 -0.208 -1.382 -2.140 -2.140 x 25 30 35 40 45 50 55 2.90 15 p 0.067 0.333 0.267 0.400 0.533 0.667 0.800 y 0.994 0.095 0.278 -0.087 -0.463 -0.904 -1.500 x 60 80 100 120 140 160 180 200 220 2.66 10 p 0.10 0.30 0.40 0.20 0.50 0,40 0.50 0,70 0.80 y 0.834 0.186 -0.087 0.476 -0.366 -0.087 -0.366 -1.031 -1.500 Table 1 . Calculation of lo lo re ression for reaction of .00 er cent •henol with coli•4asl2sLia 18 replicate experiments. x 16 20 24 28 32 36 40 x2 256 400 576 784 :1024. 1296 1600 P 0.055 0.278 0.389 0.444 0.778 0.889 0.889 Y 1.065 0.247 -0.057 -0.208 -1.382 -2.14.0 -2.14.0 Y 1.26 0.60 -0.06 -0.72 -1.38 -2.04 -2.70 P 0.029 0.162 0.390 0.615 0.778 0.878 0.935 w 0.3524 0.6403 0.5937 0.3835 0.2174 0.1265 0.0646 Yw 0.9822 0.2063 -0.0573 -0.154.1 --1.3800 -2.1331 -1.9675 Y2 0.9647 0.0426 0.0033 0.0237 1.9044 4.5501 3.8711 0.4023 -0.1073 -0.6169 -1.1265 -1.6361 -2.1457 Y1 0.9119
Sw = 2.3784 S(wx2) =. 1478.8912 S(wYw) = -0.3118 S(wx) = 57.5260 S(wxyw) = -18.6894 s2(wyw) = 0.0972 s(w4 ) = 1.8632 82(wx) = 3309.207 105,,
The following statistics were then computed:- - = 5. 260 = 24.1868, = :021,..3188 = 0.1311 2.3784 Jw 2.3784
8(w4) S2(wYw) = 1.8223 •••••1111111.1••••01Vil•
Sw s(wxyw) S(WY). S(wx) = ^11.1480
Sw
S(wx2) S2 (wx) = 87.5186 Sw b = = —0.1274 87.5f86
Total Sum of Squares = 1.8223 x 18 = 32.8011.4. S.sq. due to regression = 18(--11.1480)2 = 25.5603 87.5186 Remaining S. sq = 7.p411 An analysis of X.2 was drawn up as follows
Item Degrees of -, 2 P Freedom It.
Regression- 1 25.5603 less than 0.001 Remainder 5 7.2411 about 0.20 Total 6 32.8014
Thus the regression is highly significant although the Remainder k2 indicates some heterogeneity in the data. 106.
Next, the variances and standard errors of 'al (=Yvv) and fbl were calculated:- V = 1 = 1 = 0.02336 a FT;73 42.8112
1 , 1 = 0.000634.88 Vb = 771177571 87.5186 x 18
Therefore a = y = -0.1311 + 0.1528
b = -0.1274. + 0,0252 The regression equation is Y = -0.1311 - 0.1274 (x-24.1868) = 2.9503 - 0.1274x from which the M.S.S.T. was calculated as 23.1578 minutes. The standard error of the estimate as calculated from the expression
Sx ... 1. + R) 1 b S(nw) S(nw(x-x) 2
is + 1.217 minutes. From the regression equation a new set of values, Y1, were calculated for the corresponding values of x (see Table 17) and from these a second approximation to the loglog regression obtained. This second cycle of computations yielded the following values:- a = -0.1199 + 0.11111 b = -0.1343 10.07.0 Remainder X 2 = 2.6752 Regression equation: Y = -0.1199 - 0.134.3(x - 24.2233) M.S.S.T. = 23.3306 + 1.058. 107.
The first calculated regression line is thus better than that fitted by eye since the second approximation has yie3ded a diminished remainder, or heterogeneity, X 2 (from 7.203 to 2.6752) and smaller estimates of the standard error of M.S.S.T. (from 1.217 to 1.058) and slope (from 0.0252 to 0.0211). However, the values of tat, rbt and M.S.S.T. obtained from the first approximation fall easily within the standard erronsof ihe estimates from the second approximation. Further, the M.S.S.T. estimated visually from the line fitted by inspection lies well within the llmits of error derived from bath calculated values. Visual and calculated estimates of the M.S.S.T. for each of the five phenol solutions tested are recorded in Table 18 together with the standard errors of the calculated values and the probability levels to which the 2 reminder ;(- corresponds in each case. Table 18. Estimation of Mean Single Survivor Times (M.S.S.T.) for coliphage T6r exposed to various concentrations of phenol.
CALCULATED ESTIMATES Phenol Visual -- concentrations estimate per cent 171/v of M.S.S.T. First Approximation Second Approximation (minutes) M.S.S.T. Stananrd P M.S.S.T. Standard P error (remainder error (remainder (2) )(2)
3.50 2.03 2.117 0.176 0.5 - 0.7 2.120 0.172 0.5 - 0.7 3.25 7.45 7.255 0.392 0.3 - 0.5 7.353 0.364 0.3 - 0.5 3.00 23.60 23.158 1.217 0.2 23.331 1.058 0.7 - 0.8 2.90 36.75 36.536 2.094 0.7 - 0.8 36.7)46 2.063 0.7 - 0,8 2. 66 124.5 119.65 15.48 0.2 - 0.3 121.22 14.03 0.7 - 0.8 1 109. FIGURE 2. Relation between phenol concentration,
Mean Inactivation Time (M.l.T) and
Mean Single Survivor Time (M,S.S.T) f or soli phage T6r at 25°C.
X = points on cal culated 2.0 regression lines
1.5
) tes inu m ( ime t og I
I .0 M. S. S. T.
0.5
I i 044 0.46 0.48 0.50 0.52 0.54 log phe nol concentration per cent w/v) 11.0.
J.. Relation of M.I.T. to M.S.S.T.
The logarithms of both the mean values of M.I.T. and the best caJculated values of M.S.S.T. referred to in the two previous sections have been plotted against the logarithms of phenol concentration in Figure 2.
The relationship of log M.I.T. with log concentration of phenol and of log M.S.S.T. with log concentrations of phenol both appear to be linear and parallel to each other. These assumptions were tested by standard statistical analysis as described, for example, by Brownlee (1949). The relevant data is shown in Table 19. Table 19. Data for calculation of regression of log M.I.T. and log M. S. S.T. with los phenol concentration for ooliphage T6r.
Phenol concentration log M.S.S.T. log M.I.T. M.I.T. (minutes) M.S.S.T (minutes) =y1 = Y2 per cent log w/v per cent =x
1 3.5 0.54111 2.997 0.4767 2.120 a 0.3263 3.25 0.5119 10.27 1.0115 7.353 0.8665 3.00 0.4771 33.33 1.5228 23.33 1.3680 2.90 0.4624 47.33 1.6752 36.75 1.5653 2.66 0.4249 206.0 2.3139 121.2 2.0835
Sx = 2.4204 S(x- 7)2 = Sx2 (Sx)2/n = 0,00839737 S.yi = 7.0001
S(y, - 51)2 = (By1)2 /n S(x 70 (Y1 37.1)
= S(xW') (Sx)(Sy) = 1092944323
= -0.12713262 S(y2 --52)2 = 1.80802785
S(x 7c) (y2 - 3;2) = -0.12309532
112,
For the log M.I.Z. v log phenol concentration relationship, the calculated from the expression correlation coefficient (r1),
r = S(x ;)( Y1 — 3r1) s(c-3c' )2. gyi.—r1)2) is 0.9985 with 3 degrees of freedom. This exceeds the tabulated value for a level of significance of 0.001 and the correlation is therefore highly significant. The calculated equation for the regression of log phenol concentration upon log M.I.T. has the components
= y al = Sy1 = 1.40002
= S(Yl — Y1) (x ) . —15.13957 S(x 2)2
and 3.c = Sx = 0.008
The residual variance about the regression line(s1) calculated from N 2 2 2 , /n-2) is 0.0018905. 3, = r1). (Sur1 y ) Similarly, the log M.S.S.T. v. log phenol concentration shows a ) of 0.9988, which again shows a highly correlation coefficient (r2 significant correlation, and values of a2 = 1.24192 and b2 —14. 65879 The residnnl variance about the regression line(s2) is, in this case, 0.0014103. The regression coefficients, b1 and b2, were compared by first calculating the weighted average residual variance
s2 = (nisi + n2s2)/ /(ni + n2) = 0.00165025 113. where ni = n2 = 3, the degrees of freedom of each regression line.
The variance of the difference between b and b is then 1 2
S2 = 2f s2 = 0.38304 12 302 )
b = 0.48078/Q61887 = 0.7769 and t = b1 2 S12 The degrees of freedom for this t are n1 n2 = 6 so that it corresponds to a probability of 0.4 to 0.5. There is therefore no evidence that the regression coefficients, bl and b2 are different. It was therefore concluded that a linear relationship exists both between log M.I.T. and log phenol concentration and between log M.S.S.T. and log phenol concentration. Further, the regression lines representing these relationships are parallel, at least over the time range of 2 to 200 minutes. Finally, the mean ratio of M.I.T./M.S.S.T., obtained from the antilog of - - is 1.439. 35. - Y2
5. Two factors affecting the reproducibility of the M.I.T. (a) Effect of sample volume. Three determinations of the M.I.T. of 2.9 per cent phenol were carried out using the standard technique but distributing one-drop samples of the reaction mixture into the empty sterile test tubes in place of the 6-drop samples normally used. A contact time interval of 5 minutes wag,- used and 5 replicate tests performed for each M.I.T. The results of these tests are shown in Table 20. For comparison the results obtained by three determinations on the same strength of phenol but using the 334.
Table 20. Effect of reaction mixture sample volume on value of M.I.T. Inactivation of colipha e T6r by 2.9 per cent phenol i Sample Range of M.I.T. Mean M.I.T. 95 per cent Volume Inactivation (minutes) (minutes) confidence time limits of the (minutes) mean M.I.T.
One drop 30 - 50 37 32.7 + 10.1
20 - 40 29
25 - 45 32 --_-_ -__ + 16-drops 25 - 60 47 47.3 - 6.25 30 - 60 50 40 - 50 45 1:15. standardised technique throughout are included in the table although these have already been recorded in Table 15. Inspection of the results shows that the decrease in the sample volume produces apparent decrease in the M.I.T. together with an increase in variation between estimates of the M.I.T. It is, however, of interest to note that the ratio of the variance of the two treatments, 2 F = 2.605, indicates that there is no statistically significant difference 2 in the variances. Similarly comparison of the two means by the tti test gives a value of t = 1.5729 with 4 degrees of freedom. This corresponds to a probability level of 0.1 to 0.2 and indicates that there is no significant difference between the means. Further tests would show whether this apparent difference was real but the probability level of 0.1 to 0.2 for 14. degrees of freedom indicates that even if the difference, when tested with more degrees of -freedom, is real it is unlikely to be large enough to have any practical significance.
(b) Effect of variations in inoculum titer. Two determinations of the M.I.T. of 2.9 per cent phenol were made using as inocula for the reaction mixtures a 1 in 100 broth dilution and a 1 in 10 broth dilution of the phage stock respectively. Apart from the inocula variations the determinations were carried out by the standard technique and each consisted of 5 replicate determinations with a 5 minute contact time interval. The results are shown in Table 21. 12 6.
Table 21. Effect of inoculum titer on value of M.I.T. Inactivation cf coliphage T r by 2.9 per cent henol.
hage inoculum, Range of Inactivation M.I.T. 'lution of stock. Times(4iinutes) (Minutes) -2 10 5-25 14 -1 10 25 - 40 34 0 25. -60 47
Gross variation in the titer of the phage inoculum therefore has a definite effect on the value of the M.I.T. although it seems unlikely
that even 2 or 3-fold variations in the titer will have an appreciable
effect on the estimate of the M.I.T. obtained.
6. The Systematic examination of the antibacterial agents. (a) Modification of dilution procedure., The standardised technique results in the solution under test being diluted by approximately one in 200 before the host culture is added.
With the exception of nminacrine hydrochloride and formaldehyde, the concentrations of antibacterial agents tested were such that the standard dilution procedure reduced the concentration to or below the concentrations previously shown to be sub-inhibitory to the lytic action of even small numbers of phage (see Table 13). To ensure adequate dilution in the tests on aminacrine hydrochloride, the 6-drop samples of reaction mixture were distributed into 250 ml. glass-stoppered bottles and subsequently diluted with 200 ml. volumes of broth to give a dilution of .7 :17.
approximately one in 2,000. Similarly, in the testa on formaldehyde the reaction mixture samples were distributed into 100 ml., glass-stoppered bottles and subsequently diluted with 100 ml. volumes of broth, giving a dilution of approximately one in 1,000.
(b)Ranging Tests.
An approximate indication of the effectiveness of each antibacterial agent in inactivating free phage was first obtained. The previously described standardised technique was used with contact times of 10,
20 and 30 minutes, 1, 2. 4, 8 and 18 hours, 2, 3, and 5 days and the tests were conducted in triplicate. The samples to be tested for surviving phage after more than 8 hours contact with the antibacterial agents were placed in glass-stoppered containers to avoid concentration of the solution due to evaporation and were stored in an incubator at 25° C. The results of these ranging experiments are shown in Table 22 and have been quoted as "Inactivation Times", the term "Mean Inactivation Time" (M.I.T.) being used only for the results from tests with not less than 5 replicates and contact time intervals of approximatelyV8to 1/10 of the M.I.T.
The inactivation time quoted for chlorokresol was obtained from a test in which contact times of 3, 4, 5, 6 and 7 days were used. (c)Further examination of the more effective agents.
The antibacterial agents which showed an inactivation time of less than 10 minutes in the concentrations used in the ranging tests were subjected to further study. Each was used in a series of ranging tests using different concentrations of the antibacterial until the concentration 11.8,
Table 22. The virioidal activi of a ueous solutions of various antibacterial agents at 250C against coliphage T r.
Antibacterial Concentration Inactivation Time per cent w/v Molar liminacrine Hydrochloride 0.2 8.0 x 10 3 > 5 days -4 Cetrimide 0.01 2.7 x 10 <7.. 10 minutes Chloramine T. 0.1 3.5 x 10-3 <7 10 minutes -2 ChIorocresol 0.2 1.4 x 10 5 days Chloroxylenol 1 6.L. x 102 > 5 days (solution) (20) 2 Cresol (and 1 9.2 x 10 3 days soap solution) (2)
Crystal Violet 0.01 2.4 x 10 4; 10 minutes
Formaldehyde 1.0 3.3 x 10-1 < 10 minutes -1 Phenol 5.0 5,3 x 10 4:7 10 minutes Phenylmercuric -5 Nitrate 0.002 3.1 x 10 ) 5 days
Sodium Lauryl 0.1 2.5 x 10 3 > 5 days Sulphate 119. FIGURE 3 Relation between concentration of antibacterial agent and
Mea n 1 nacti vatio n Time (M, I. T.)
Cct rimi de Crystal Violet Phenol CI--
Chl orarnine T kr-- Formaldehyde
X = points on calculated lines
2 • 0
7 C
12. I - 5
0
I •0 0 I 5.8 —.4.2 4—.3 4.8 3-0 3.2 i4 1.0
log concentration (Molar) 1200
range giving conveniently measurable inactivation times was established.
Several concentrations in this range were then used in estimations of Mean Inactivation Times by the standardised technique. The results of these experiments are shown in Tables 23 and 24_ The results for phenol have previously been reported in detail but are included here for comparison.
For each antibacterial agent the logarithms of the Mean Inactivation
Times have been plotted against the logarithms of molar concentration in Figure 3 and the calculated data for the regressions shown in Table 25. The concentrations of each antibacterial agent which would be
expected to give Mean Inactivation Times of 10 minutes and 60 minutes were calculated from the regression of log. molar concentration upon log M.I.T. represented by the formula:- x = a' + b' Cy - Y)
where x = log, molar concentration; a' = b' = slope of x upon y and y = log M.I.T. These calculated concentrations are shown in Table 26, together with the 95 per cent confidence limits of the estimates; the limits being obtained from the residual standard deviation of x. The calculated concentrations were used to calculate the Dilution Coefficient ("n") from the expression:-
"n" = log t, - log t2
log C2 - log C1 where C1 = concentration giving a M.I.T. of 60 minutes (t1) and C 2 = concentration giving a M.I.T. of 10 minutes (t2)
The values of "n" are also recorded in Table 26. Table 23. The inactivation of coliphage T6r by Cetrimide and Chloramine T. in aqueous solution at 25°C.
Antibacterial Concentration Contact Range of Mean Inactivation Time Inactivation Time (Minutes) per cent w/v Molar Interval Time (Minutes) (Minutes)
Cetrimide 6,5 x 10 3 1.78 x lo- I 8- 14. 9.80 6.2 x 10 3 1.7o x 10-4 2 14 - 26 19.3 6.o x l03 1.65 x 10-4 3 21 - 33 26.o 5.7 x 10-3 1.56 x 10-4 5 40 - 7o 52.o 5.5 x 10-3 1.51 x 10-4 lo 7o - 120 94.0 -3 Chloramine T 6 x lo-2 2.13 x lo 1 9- 13 10.60 2 5 x lo 1.78 x 10 3 2 14- - 24. 19.3 4. x 110-2 1.42 x 10-3 3 24 - 33 27.5 3 x lo-2 1.06 x 10-3 6 48 - 72 57.0 -2 -4 2 x 10 7.10 x 10 lo 80 - 120 98.0 Table 2/4-. The inactivation of coliphage T6r by Crystal Violet, Forrn1dehyde and Phenol in aqueous solution at 25°C.
Antibacterial Concentration Contact Range of Mean Inactivation Time Inactivation Times per cent virbr Molar Interval Times (Minutes) (Minutes) (minutes)
Crystal Violet 4,3 x 10-3 1.05 x 10 4 1 7 - 10 8.20 4.0 x 10 3 9.80 x 10 5 2 14- 22 18.7 3.6 x 103 8.82 x 10-5 5 30 - 55 47.0 3.3 x 103 8.09 x 105 15 130 - 175 151
_ Formaldehyde 8 x 10 1 2.67 x 10 1 1 8- 12 9.6 7 x 10-1 2.33 x 101 1.5 10.5 - 15 13.25 6 x 10-1 2.00 x 10- 1 2 18 - 26 21.3 1 -1 5 x 10 1.67 x 10 5 35 - 55 46.0 4 x 10-1 1.33 x 10-1 10 80 - 110 92.0 i - Phenol 3.50 3.72 x 10 1 ia 24-- 4. 3.00 _1 3.25 3.45 x 10 1 8-13 10.3 -1 3.00 3.19 x 10 4. 24. - 44 33.3 -1 2.90 3.08 x 10 5 25 - 60 47.3' 2,66 2.83 x 101 20 180 - 240 206 Table 25. Data calculated from the regression of log. molar concentration and loeLLILAntibacterial Agents- x = log. molar concentration (coded to give positive values); y = log.
Antibacterial Correlation Degrees of -Slope of Slope of Residual coefficient freedom i y upon x upon Standard (r) x Yi Deviation (b) (b ) of x
Cetrimide 0.9972 3 0.214]. 1.4762 -13.2790 -0.0749 0.0025
Chloramine T. 0.9896 3 1.1215 1.4995 -1.9985 -0.4900 0.0314
Crystal Violet 09964 2 0.9664 1.5092 -10.470 -0,0915 0,0052
Formaldehyde 0.9966 3 0.2883 1.4119 -3.3517 -0.2963 0.0113
Phenol 0.9986 3 0.5105 1.4002 -15.2032 -0.0656 0,0028
At the 10 per cent 'level of significance, r = 0.8054.with 3 degrees of freedom and 0.900 with 2 degrees of freedom. Table 26. Molar concentration of 5 antibacterial agents giving M.I.2. of 10 minutes and 60 minutes Zcaloulated from the regression of log.mular concentration 1177117717717717171re dilution cooffi#ent of these antibacterials.
Antibacterial Molar concentration giving M.I.T. of:.,. Dilution 10 minutes 95 per cent 60 minutes 95 per oent Coefficient confidence confidence 11nu limits -. limits 4 -4 6 Cetrimide 1.78x10 ±3x10 1.55x10 13x10 12.99
-3 Chloramine T. 2.32x10 +6.0x10- ,4 9.66x10 4 +2.5x10 2.04 -4.8x10-4 -2.0x10-44 4 6 6 Crystal Violet 1.03x10 1.4x10 8.75x105 -3 x 10 10.78
-2 -1 2 Formaldehyde 2.57x10-1 4+2.2x10 1.51x10 +1.3x10-2 3.37 -2.0x10- 2 -1.2x10 ..... - 1 Phenol 3.44x16 1 +7 x 10-3 3.06x10 6 x 103 15.23. 125.
The regression coefficients for the regression of log. N.I.T. upon log, molar concentration for cetrimide, crystal violet and phenol were compared by the procedure previously described in Section F.4. The same procedure was used to compare the regression coefficients for chloramine T. and formaldehyde. The details of the tests are shown in Table 27. It was concluded that, while the values of t in all oases exceeded the tabulated values for P = 0.05, the regression coefficient for cetrimide is not significantly different from those for crystal violet and phenol. The regression coefficient for phenol does, however, differ significantly from that for crystal violet and those for chlornmine T and formaldehyde are also significantly different. Table 27. Comparison of regression coefficients for the regression of log. M.I.T. upon log, molar concentration.
Comparing regression Weighted Variance of Degrees of Corresponding coefficients of:— Average difference Freedom probability. Residual in regression t Variance coefficients
Cetrimide and Phenol. 0.00146 0.6222 3.092 6 0.02 to 0.05
Cetrimide and Crystal 0.001934 0.8504 2.860 . 5 0.02 to 0.05 Violet
Phenol and Crystal 0.002366 0.6009 7.283 5 about 0.001 Violet
Chloramine '1' and 0.002735 0.06729 20.11 6 less than 0.001 Formaldehyde 127,,
PART III, DISCUSSION. 128.
The Bacteriophase.
The investigations reported here have been limited to the reactions of a single strain of bacteriophage, the coliphage T6r, and its host,
E. coli B, to 11 antibacterial agents. This approach was adopted since it was felt that obtaining precise, quantitative estimates of the reaction of one phage to a relatively limited number of antibacterial agents would be of more value than the performance of a survey, such as those of a number of previous workers, giving a qualitative indication of the reaction of a number of phages to a large number of antibacterial agents
(Deutsch and Rohr, 1955; Hunter and Whitehead, 1940; Klein et al, 1945).
A coliphage was chosen because of the detailed information available on the structure and function of this type of phage. The decision to use
T6 was based on indications, to be found in the literature, that the
T-even phages are more resistant to inactivation by antibacterial agents than are many other phages.
Phase plaque
The investigations on the phage plaque counting method described were prompted by the lack of information on the reliability of the surface-drop method of plating in such counts, The method of plating developed was found to give statistically reliable results (see page 61 ).
The method of dilution of phage suspensions used, together with the standardised plating method, gave acceptable limits of error between replicate counts, the 95 per cent confidence limits of the mean of 5 counts using different dilution schemes being of the order of 10 per cent of the 129. mean.
In the present investigations on antibacterial agents, plaque counts were used only to standardise the phage inoculum. They therefore form an incidental, albeit essential, part of the present work. The investigation of the method has been presented here since, it is believed, it makes a contribution to plaque counting techniques in general and to the evaluation of the inactivation of phage by counting methods in particular. It is particularly suitable for the latter purpose since the dilution method using small but carefully controlled volumes of inocula provides a rapid means of diluting the phage. The method of plating by drops is economical in materials, rapid and reliable. Cultivation of phage.
The method for the cultivation of the phage in dilute liquid cultures of the host using phage inocula small enough to ensure single infection of the host gave final titers of the order of 7 x 109 phage particles per ml. and was simple to perform. Attempts to increase the phage yield, by modifying the cultural conditions to produce more vigorous growth of the host, failed, any additional phage produced presumably being lost by multiple infection of living cells and by adsorption to dead cells and bacterial debris. The yield is, apparently, characteristic of
T6r in fluid culture.
Higher titers of phage (about 7 x 1010 per ml.) were obtained by a method of cultivation on solid medium cultures of the host. The method was, however, less economical in time and materials and when large volumes of phage suspensions are required the liquid culture method is preferable. 130.
No comparisons were made in the present work of the action of antibacterial agents on phage produced by different methods of cultivation but the subject may be worthy of investigation.
Stowe of phage.
The investigations on the effect of storage on the infectivity of the phage (as measured by plaque counts) have shown the phage to be remarkably stable under a variety of conditions. The rate of inactivation increased slightly when the temperature of storage was increased from 5°C to 37°C but no difference in stability was detected between suspensions of the phage in broth and in 1 per cent dilutions of broth in water.
This inherent stability greatly facilitates the use of phage in experiments investigating the action of inactivating agents. It would be of interest to compare the variation in the resistance of phage to inactivation by antibacterial agents with the variation in its infectivity over prolonged periods of storage. The present results with phenol indicate no alteration in resistance over several months.
Inhibition of the action of the phase on its host.
The investigations carried out on the inhibition of the growth of the phage by antibacterial agents had the primary object of determining the maximum concentrations of the antibacterial agents at which small numbers of phage could be readily detected by their initiation of mass lysis in broth cultures of the host bacterium. The procedure used in these investigations will demonstrate only relatively gross effects on the rate of lysis of the culture and only pronounced selective inhibitions of 131. the growth of the phage without inhibition of the growth of the host will be detectable. The results obtained do, however, show some interesting features.
The effect of different antibacterial agents on the growth of the host cells alone differed in 3 respects: -
(a) The concentrations of different antibacterial agents required to
inhibit the growth of the host cells showed a very wide variation.
This is well-known and is presumably related to the different modes
of aotion of the agents.
(b) All of the antibacterial agents caused a reduction in the rate of
bacterial growth at concentrations approaching their final
bacteriostatic concentrations but the extent of this reduction varied
with different agents. The results obtained with chloramine T
and phenylmercuric nitrate were inconclusive but amongst the other
agents, and within the limits of precision of tha comparisons, no
obvious relationship existed between the extent of growth rate
reduction and the "dilution coefficient" usually attributed to the
antibacterial agent. (The "dilution coefficient" is a function of the
effect of concentration on the bactericidal activity of an
antibacterial agent, a high value far the dilution coefficient
indicating that the concentration of the agent has a pronounced
effect on its activity, a low value indicating that the activity
is less affected by changes in the concentration). For example,
both Cresol and Soap Solution and formaldehyde showed marked
reduction in the growth rate of the organisms in sub-bacteriostatic 3.32.
concentrations but the former has a relatively high dilution
coefficient, the latter a relatively low one. On the other hand,
the growth rate reduction effect was not so pronounced with either
phenol or nminacrine hydrochloride and again the former has a high
dilution coefficient, the latter a low one.
The method here of assessing the growth rate of the organisms
by the time, of incubation required to produce a visible turbidity
is relatively crude. While it gives an adequate indication of the
gross variations ingrowth rate occurring in the presence of different
antibacterial agents, the conclusions regarding the correlation of
the dilution coefficient of an antibacterial agent and the extent
of the growth rate reduction occurring in its presence must be
tentative.
(0) If the assumption is made that the width of the range between the concentration of antibacterial agent inhibiting growth in all
replicates and that permitting growth in all replicates is a measure
of the variation in resistance among the individual bacterial oells
of the inoculum, then this variation aiffered for differ4nt
antibacterial agents. The range of concentrations was widest with
cetrimide and sodium lauryl sulphate, but,in their case, was
probably the result of the clumping of bacterial cells that both of
these agents are known to cause.
In cultures of the phage, all the antibacterial agents caused an increase in lysis time at concentrations approaching the bacteriostatic concentrations. With the exception of aminacrine hydrochloride and 133. sodium lauryl sulphate, this increase in lysis time corresponded to that which might be expected from the reduced rate of growth shown by the host alone under the same conditions. When the concentration of antibacterial agent was low enough to permit a normal rate of growth of the host, the phage cultures showed a normal rate of lysis. There was, therefore, no evidence that these agents had a selective phage inhibitory action compared to their action on the host.
In the presence of sub—bacteriostatic concentrations of aminacrine hydrochloride a much more pronounced increase in lysis time was seen. The inhibition of mass lysis by the acridines is well established and the fact that it was detected here with aminacrine hydrochloride serves to prove that the method of examination was effective. The results obtained, however, provide no clear evidence that the lysis delay occurring does not correspond to the reduced rate of growth of the host cells. Mass lysis in phage cultures was found to be completely inhibited by sodium lauryl sulphate in a narrow range of concentrations which were just sub—bacteriostatic for the host. This inhibition is unlikely to occur by the same mechanism as that of aminacrine hydroohloride which is believed to act by combining with the phage DNA so preventing the maturation of the phage particle. Sodium lauryl sulphate has no effect on free DNA and phage inhibitory concentrations were later shown to cause no narked inactivation of free phage particles. It seems likely that, because of its high surface activity, it is preventirg the adsorption of phage by an effect on the host cells.
The selective inhibition of phage growth by crystal violet, reported by Graham and Nelson (195)f) for lactic streptococcus phases, was not found to occur with T6r.
The significance of these findings with regard to viricidal tests on free phage may be reiterated. It is essential that, when testing for phage surviving at the end of the inactivating process, the lethal agent must be neutralised chemically or diluted to a level which not only will permit the growth of the host cells but which has been shown to permit the infection of the host by small numbers of phage particles. It is not clear that this point has received the attention it merits in many of the previous investigations into the action of antibacterial agents on phage. 135.
Ipsotivation of free -Oar = The method of evaluation.
The principles of the method first described by Berry and Bean (1954)
for the estimation of the extinction time for bacteria exposed to an antibacterial agent have, with some success, been applied to the assessment of the action of antibacterial agents on bacteriophage. Some of the features and limitations of the procedure used are:-
The phage inoculum. The phage inoculum used consisted of a broth suspension of the phage. The 10 ml. of reaction mixture therefore contained, in addition to 4 x 109 phage particles, approximately 1 mg. of peptone and 0.5 mg, of sodium chloride. The particle weight of phage 13 T6 having been estimated to be of the order of 5 x 10 mg. (Adams 1959), the reaction mixture therefore contained about 2 x 10 mg. or phage material representing only about 0.2 per cent of the total nitrogenous material present. The content of non-phage material in the inoculum was constant but, when an antibacterial agent which combines with non-specific proteinous material is being tested, the proportion of the agent lost by combining with the non-phage material will depend on the concentration of the agent present. The phage titer of the inoculum was standardised by plaque counts
but the survival of the phage exposed to antibacterial agents was estimated by its ability to grow in fluid culture. Justification for this comparison can be found in the close agreement previously shown between
plaque counts and dilution end-point counts (hershey et al, 1943; Eleczkowski and Eleczkowski, 1951). 136.
Variations of less than 10—fold in the titer of the inoculum have
been shown to be unlikely to cause appreciable variations in the
Mean Inactivation Time of the phage in contact with an antibacterial
agent. This effect is no more than a reflection on the high titers
usea and the exponential rate of inactivation of the phage together
with the limits of error of the method of estimating the inactivation time.
All of the phage inonula in the present work were taken from the same
phage stock. No information was therefore obtained regarding the
effects of different methods of cultivation of the phage on its sensitivity to inactivating agents. This subject merits further study,
Alper (1950 having shown that the medium on which the phage stocks were cultivated affected the sensitivity of 2 coliphages to ionising radiations.
The reaction mixture sample. to minimise the errors in measuring the samples of reaction mixture the volume taken was made relatively large
(approximately 0.1 mi.) and measured using a dropping pipette; the accuracy of such measurements, determined using water, was such as to give 95 per cent confidence limits of approximately 0.2 per cent of the
mean volume delivered. The drop volume delivered by a dropping pipette varies between solutions of different composition. The variations in sample volume between solutions of different antibacterial agents at various concentrations are unlikely, however, to appreciably affect estimates of their inactivation times since alteration of the sample volume from 6 to one drop was shown to produce a barely significant alteration in the Mean
Inactivation Time estimate. Drop volume variations due to the composition of the solution will certainly be less than 6-fold.
The temperature of the reaction. All of the present studies on the inactivation of phage were carried out at a temperature of 25°C, chosen as
the lowest temperature at which the water-bath could be conveniently
maintained throughout the year. Further work on the effect of higher
temperatures on the inactivation of phage by antibacterial agents may yield further information on their mode of action.
Testing the reaction mixture for infective phage. The test applied to
each sample of the reaction mixtures at the end of the inactivation
period was intended to show simply whether or not infective phage had
survived exposure to the antibacterial agent. The number of host cells added to the diluted reaction mixture samples was approximately equal to
the number of phage originally present in the sample. While the cultures were examined at relatively infrequent intervals, cultures from samples giving increasing time of contact of the phage with an antibacterial agent showed, in general, an increase in lysis time corresponding to the increase in time shown by decreasing numbers of phage in cultures prepared from untreated phage. The tests therefore gave no indication that inactivated phage were adsorbed to and lysed the host cells or that multiplicity reactivation occurred. Avery few reaction mixture sample cultures showed a pronounced delay in lysis, clearing of the culture being apparent only after 2 days incubation. The phenomenon, although rare, was seen with all the antibacterial agents examined in detail and occurred only in cultures from samples quenched at times approaching the final inactivation time when the numbers of infective particles present were smn1; 138.
The rate of lysis was much slower than that shown previously in the presence of sub-bacteriostatic concentrations of the antibacterial agents (section E.3.) even for small phage inocula and no explanation of the effect can be offered at the present. A more detailed study cf the interaction between a phage and its host after exposure of the phage to antibacterial agents may be rewarding. Such a study should include an examination of the behaviour of both infective and non-infective phage particles.
In the present work, simple dilution of the reaction mixture samples was relied upon to over-come inhibition of the growth of the host or surviving phage by the antibacterial agents. The reaction mixture sample volume used was relatively large and with many of the antibacterial agents the concentration required to inactivate the phage was high relative to those showing a bacteriAV: action. With these agents, therefore, the procedure used either limited the maximum concentration of the agent which could be tested or necessitated the use of large volumes of broth. For example, in the tests on formaldehyde and aminaorine hydrochloride, volumes of broth of 100 ml. and 200 ml. respectively were required. The reaction mixture samples therefore had to be distributed into relatively large containers and the maintenance of these containers, in the large numb3rs required by the desired level of replication, at 25°C in a water bath was inconvenient. Further work on formaldehyde and some of the other agents discussed later, should include a study of the chemical neutralisation of their inhibitory effects on phage growth. 1;9.
The precision and reproducibility of the method.
To obtain a precise estimate of the inactivation time of the phage
by an antibacterial agent the contact time interval used in the tests
was short in relation to the final inactivation time. A wide variation
was, however, found between individual estimates of the inactivation times,
the range Letween maximum and minimum values obtained from 5 to 6 replicate estimations commonly being anything from 50 to 75 per cent of the mean
estimate. This variation is an indication of the distribution of resistance to inactivation between individual phage particles in the
phage inoeulum and similar ranges were found with all the antibacterial
agents for which precise estimates of inactivation times were rossible.
The variation found here is also generally similar to the variation found
by Berry and Bean (1954.) and Briggs (1955), using a technique very similar to the one used here, for estimates of the bactericidal action of phenol
on E. coli. This implies that the distribution of resistance among phage is similar to that among bacteria.
The need fox adequate replication in extinction time estimates of the inactivation of phage is amply illustrated by the variations found here.
The use of a 5 or 6 fold level of replication was found to give a mean estimate of the inactivation time which adequately reflected the distribution of resistant individuals and the reproducibility of which was therefore acceptable. The reproducibility is illustrated by the variation found between several estimates of the Mean Inactivation Time for several concentrations of phenol. Here the 95 per cent confidence limits were about 10 to 20 per cent of the Nban Inactivation Times. It should 3/1.0. be noted that the estimates on phenol were performed at intervals aver a period of some months So that the confidence limits incorporate the errors due to vacations in all aspects of the tests.
The Mean Single Survivor Time (.20:112:11. The successful application of the analysis of Mather (1949) to the results of those tests on phenol for which extensive replication had been used is conclusive evidence that the distribution of surviving phage between the samples towards the end of the inactivating process is the same as that found amongst bacteria, the distribution being Poissonian. Interpretation of the results of phage inactivation tests in terms of the Mean Single Survivor Time is therefore possible but requires the performance of a large number of replicate tests. The standard error of the estimations can be computed al+hough the cycle of computations is too complex for routino use. These errors have, however, beet calculated for the results with phenol and it has been found that the value of the M.S.S.T. estimated visually from the plot of contact time against log( —log p) falls within the calculated limits of error from both the first and second approximation. The reliability of this visual estimate is only acceptable when 10 or more replicate determinations have been performed using a contact time interval of 1/8 to 1/10 of the expected inactivation time (or 1/6 to 1/7 of the M.S.S.T.). 110
The relation of Mean Inactivation Time pg M.S.S.T. Interpretation of extinction time data by means of the M.S.S.T. undoubtedly gives the most accurate interpretation of these data bat it does require a high level of replication to have any significance. It has been shown that the relationship of the Mean Inactivation Time to the M.S.S.T. is constant over a wide range of extinction times. This is considered to indicate that at a high level of replication, the interpretation of the results in terms of the Mean Inactivation Times is as reliable as their interpretation in terms of the M.S.S.T. At a lower level of replication, estimates of the Mean Inactivation Time have been shown to have confidence limits of 10 to 20 per cent of the mean. These limits of accuracy were considered to be acceptable for the present work and the exam4nation of the effect of antibacterial agents on phage was carried out using 5 or 6 replicate determinations of the inactivation times, expressing the results as the Mean Inactivation Time. 1/12.
The systematic examination of the antibacterial agents.
The method by which the antibacterial agents were examined will
detect only those agents which are very efficient inactivators of phage
at the concentrations tested. Thus, in the ranging tests carried uut,
considerable numbers of the phage may have been inactivated by aminacrine
hydrochloride, Chloroxylenol Solution, phenylmercuric nitrate and sodium
lauryl sulphate but, because they failed to inactivate all the phage in
the inoculum within 5 days they were classed as inactive in the
concentrations tested.
Attention was concentrated, for the present, on those agents showing
a high degree of activity, comparisons of their relative activity being
made on the basis of the time required for various concentrations of the agents to completely inactivate the phage present or the concentration of the agents required to show a given Mean Inactivation Time. The comparisons are therefore based on the effectiveness of the agents in inactivating the most resistant members of the phage population and as such, while constituting a stringent test of the effectiveness of the agents, give a realistic estimate of their action on the population as a whole. The disadvantage of the method of examination is that it gives no information on the kinetics of the inactiTation of the less resistant individuals in the population such as is obtained by plaque counting methods. The possible significance of these kinetics in relation to the mechanism of inactivation of viruses generally by chemical means has recently been reviewed by Gard (1960).
A detailed examination of the action of chlorocresol, Chloroxylenol 143. Solution and sodium lauryl sulphate was not undertaken at the present since the testing of concentrations higher than those used in the ranging tests would require the use of large volumes of broth to dilute the samples sufficiently to overcome the inhibitory actions of the compounds. The experience, gained from the tests on aminacrine hydrochloride and formaldehyde, of the inconvenience of using these large volumes of diluent discouraged their use in further tests. The results obtained using phenylmercuric nitrate were inconclusive since only a very low concentration was tested. Further study was not undertaken as it was felt that, in the absence of pronounced activity at low concentrations, the adequate testing of the higher concentrations likely to be required would require the use of chemical neutralisation of the inhibitory activity.
Aminacrine hydrochloride was tested at a concentration giving an approximately 3/5 saturated solution (8 x 10-3M.). The survival of the phage after 5 days exposZto such a concentration indicates that, while this substance inhibits the growth of phage at low concentrations (about -5 5 x 10 M.) by, it is believed, combining with newly formed DNA, it can have little effect on the DNA of mature phage.
The relative effectiveness of chlorocrescl, Cresol and Soap Solution and phenol are of interest. It should be noted that if chlorocresol and cresol have the same high dilution coefficient as phenol (n=15) then the concentrations of these two substances required to give a Mean Inactivation
Time of 10 minutes are 2.2 x 10-2 M. and 1.4 x 10-1M. respectively, compared -1 to the required. concentration for phenol of 3.L. x 10 . Even if n=10 the -2 required concentrations for chlorocresol and cresol will be 2.7 x10 M. and
1.7 x 10-1M. Conversely, the concentration of phenol which can be expected,from the known data, to give a Mean Inactivation Time of 3 to 5 days is about 1 1: )1
2.3 x 10-1M. (compared to an inactivation time of 3 days from 9.2 x 10-2M. -2 cresol and 5 days from 1.4 x 10 M. chlorocreso1).
It therefore appears that the three phenolics can be arranged, in ascending order of activity against phage, as phenol, cresol and chlorocresol. This is the same qlmlitative order of their activities against bacteria and implies that their mechanism in inactivating phage is at least similar to that for bacteria. What this action is 1il ly to be is discussed later. A further study of the phenol derivatives with a view to determining their dilution coefficients will be of particular interest.
From the detailed examination carried out on 5 of the antibacterial agents they can be arranged, in ascending order of activity against phage, as phenol, formaldehyde, chlormine T, cetrimide and crystal violet, on the basis of the Molar concentration required to give a Mean
Inactivation Time of 10 minutes. The same order resulting from a comparison at a mean Inactivation Time of 60 minutes (see Table 26).
The most interesting feature, however, of the results of the experiments on these compounds is the variation found in the effect of concentration on their activity. This effect can be expressed as the dilution coefficient (n) which has, for all practical purposes, the same numerical value as the calculated slope of log Mean Inactivation Time against log concentration (see Tables 25 and 26).
The validity of using n as a basis for speculation on the mechanisms of the inactivation of bacteria by bactericides has been the subject of some argument (Rahn, 1945). Much of the early criticism of this use of 14 5. n was based on its variability, resulting largely from the use of imprecise methods of determination such as those used by Tilley (1939) on whose results Rahn's arguments were largely based. It now seems to be generally accepted that, provided the determination of n is sufficiently precise, significant variations in values of n between the action of different bactericides on one organism, between the action of one bactericide on different organisms or between the action of one bactericide on one organism at different temperatures, at least gives an indication that the mechanism of inactivation in each case is different. It has been concluded from the present work that the value of n has an equal significance in the inactivation of phage by antibacterial agents.
The values of n recorded here are considered to be sufficiently reliable since they were derived from the calculated regression of log concentration upon log Mean Inactivation Time, which in every case was found to show no significant deviation from linearity over the concentration ranges tested. Each calculated value for n is therefore ultimately derived from 4 or 5 estimates of the effect of concentration on the activity of the antibacterial agent, each of these estimates being based on 5 or 6 replicate estimates of the inactivation time of the agent at that concentration.
If this significance of n is accepted the 5 antibacterial agents examined in detail fall into 2 general groups; chlorronine T and formaldehyde with relatively low values of n (approximately 2 and 3 respectively) in one group, and crystal violet, cetrimide and phenol with values of n greater than 10 (approximately 11, 13 and 15 respectively) in the other. It should be noted that, within each group, the slopes of
the regression from which the values of n were derived have been found
to show significant departure from parallelity when tested statistically.
The possibility therefore exists that differences in the mechanism of
inactivation exist between the members of each group.
The question which now arises is, can any similarity be found between
the possible general mechanisms of inactivation of phage by the
substances in each group?
The action of formaldehyde on free phage has been fairly extensively
studied and the consensus of opinion seems to be that the phage is
inactivated by the combination of the formaldehyde with the protein coat
of the phage particle. In most cases no interaction between formaldehyde
and DNA has been reported and the effect on the protein coat is to
prevent ejection of the DNA but not adsorption of the phage to its host.
No attempts have been made to investigate the mechanism of action
on phage of compounds liberating chlorine. Indeed, little information
is available on the mechanism of action of these compounds on bacteria
although it seems clear that "chlorine compounds", including the
chloramines, attack bacteria through the undissociated hypochlorous
acid formed by the interaction of the liberated chlorine with water.
(Brazis et al, 1958; Hadfield, 1957; Narks et al, 1945.). Chlorine is known to have an intense reactivity with proteinous materials (Sykes, 1958)
and it seems very likely that the site of the attack on phage by
chioramine T is the protein coat.
Both formaldehyde and chlorine react strongly with organic matter so 11:7.
that the inactivation time estimates for these two agents will be
particularly affected by the presence of peptone in the phage inoculum.
The effect will be to make the Mean Inactivation Times, determined
experimentally, longer than the true inactivation time. This effect
will be most pronounced with low concentrations of the agents so that the
value of n obtained will be greater than the true value although the
size of the effect is a matter for speculation. Since the arguments
on the significance of the values of n found here are based on the
smallness of n for formaldehyde and chloramine the effect, in any case,
does not invalidate the argument.
The mode of action of crystal violet on intact phage has not been
investigated but it seems highly likely, from the known affinity of
crystal violet for DNA, that inactivation occurs by combinatitn of the
crystal violet with the phage DNA. That ability to combine with free
DNA does not necessarily confer the ability to produce marked inactivation
of intact phage is shown by the acridines, for which combination with free
DNA has been proved but which have little or no effect on free phage.
No attempt has been made to explain the action of the quaternary
ammonium compounds on phage but their effect on bacteria has been very
extensively investigated. The prevalent theory of the mode of action of
the quaternaries (including cetrimide) on bacteria is that their main
effect is to cause an increase in the permeability of the cell well which results in a lethal loss of intracellular constituents. (Salton, 1951;
Stedman et al, 1957). This effect has been suggested by Gilby and
Few (1957) to be caused by the interaction of the compounds with phospholipids in the protoplasmic membrane.
Phenol is known to have no effect on phage DNA and its actior on bacteria has been shown to be very similar to that of cetrimide (Gale and. Taylor, 1947; Maurice, 1952; Stedman et al, 1957). It is suggested that both cetrimide and phenol aot on phage in a similar manner, possibly by affecting in some way the association of the protein ooat of the phage with the DNA (without necessarily causing complete dissociation of the two)or by denaturing both the protein coat and the "internal protein" shown by Spizizen (1957) to play an essential part in initiating phage multiplication within the host cells. It is, therefore, suggested that, when the inactivation of a phage by an antibacterial agent has a low value of n, the agent is attacking the protein coat of the particle. When the value of n is high then the attack is on the internal structure of the phage, either by combination with the DNA, as in the case of crystal violet, or by affecting the association of the DNA with the protein content of the phage, as with cetrimide and phenol. These postulations will require further experimentation before they can be fully justified, but it is submitted, the determination of an accurate value for n may be a useful preliminary indioation of the site of aotion of an antibacterial agent on a phage. 149.
Phase reactions and other viruses.
As has already been pointed out, predicitions of the effioiency or mode of action of an antibacterial agent in inactivating animal viruses on the basis of its action on phage must be made with caution. There are sufficient dissimilarities in structure and function between phage and other viruses to make it clear that only when an agent is known to act on a specific feature of the phage and that feature is known to occur in other viruses will it be possible to predict if the action is likely to be similar on both kinds of virus.
If we accept the supposition that cetrimide and phenol act on phage by affecting the association of protein and DNA while chlormline T and formaldehyde attack the protein coat, then we might expect that cetrimide and phenol will be relatively less effective than chloramine T and formaldehyde against animal viruses where the strict orientation of
DNA and protein is less apparent than in phage. As far as can be judged from a number of relatively qualitative investigations into the action of antibacterial agents on animal viruses (reviewed in part by Dunham, 1957 and Sykes, 1958), animal viruses are, in fact, relatively resistant to inactivation by quaternary ammonium compounds and phenols but less resistant to inactivation by formaldehyde and chloramine T.No information is available on the action of crystal violet or similar compounds on viruses but one would expect them to be active at leaEt against those containing
DNA. These predictions must, for the present, apply only to relative efficiency in inactivating viruses since it appears that phage generslIy are more resistant than animal viruses to inactivation by at least 150„
some antibacterial agents, including a number of surface active agents.
One other parallel may be drawn between the inactivation of phage and animal viruses. The dilution coefficient for the inactivation of
purified p.oliovirus by formaldehyde appears to be about one (Shaffer,
1960) but little information appears to be available on the effect of
the concentration of ether antibacterial agents on their effectiveness in inactivating viruses. If the dilution coefficients for anims1
viruses are of the same order as some of those found here for phage
the effect on the chemical disinfection of materials contaminated with
viruses will be very significant. To illustrate the point, if a virus
suspension is completely inactivated by a given concentration of a
disinfectant in x minutes and n = 6, then halving the concentration will give an inactivation time of x hours, if n = 10 -11 the inactivation time
becomes x days and if n = 13 - 14 it becomes x weeks. The need for the determination of the dilution coefficient for the action of antibacterial agents on animal viruses is obvious. 151.
Suggestions for further work.
The investigations reported here could be usefully extended in a number of ways.
1. Several reports have been published on the departure from first order
reaction kinetics of the inactivation of phage by antibacterial
agents (Heicken and Spicier, 1956, 1959; Krueger and Baldwin,
1931+; Mills, 1951+). The inactivation is rarely followed for
more than 3 or 1+ log cycles and a study of the kinetics of the
process using both plaque counting methods and extinction time
estimates would enable the inactivation to be followed from a high
initial titer to a mean single survivor and to complete inactivation.
Differences in the shape of the time/survivor curves obtained from
different antibacterial agents may indicate differences in their
mode of action.
2. Alper (1951+) noted differences in sensitivity to ionising radiations
between phage stocks cultivated on different media. It would. be of
interest to determine if the composition of the media and other
conditions of cultivation affect the sensitivity of phage to
inactivation by antibacterial agents. If they do, care will be
required in the standardisation of phage stocks used in experiments
intended to elucidate the mechanism of inactivation.
3. A study is required of the effect of storage of phage stocks
under different conditions of tempe'ature and medium, on their
resistance to inactivation. The stability of phage, measured by
its infectivity, is well known; it would be of interest to
determine if its resistance to inactivation is equally constant. I52Q
The information is required for the adequate standardisation of phage
inocula in investigations of inactivation with antibacterial
agents. In addition, comparison of the dynamics of the slow
inactivation occurring under moderate conditions with that of the
more rapid inactivation occurring in the presence of antibacterial
agents may indicate different mechanisms in each case.
4. A wide field for investigation can be found in the effects on the
inactivation of phage by different antibacterial agents of alteration
of the pH of the reaction mixture or the temperature of the reaction.
The effects produced may indicate the chemical reactions involved.
Such an approach has been used, with some success, by Heicken and
Spicher (1959) in their studies on formaldehyde and the effects with
other antibacterial agents could be of interest.
5. An equally wide field for investigation is the modification of the
conditions provided for the recovery of phage after exposure to an
antibacterial agent. Such modifications could include the use of
chemical antagonists of the antibacterial agent, the dialysis of
the reaction mixture, variation in the multiplicity of infection of the
host cells provided for the growth of the phage at the end of the
inactivation process and variations in the cultural conditions
of these phage cultures. The effects of these modifications on the
recovery of the phage should yield valuable information on the
mechanism of the action of the antibacterial agents on the phage.
6. Attempts should be made to determine if phage which have lost their
infectivity by treatment with different antibacterial agents can
adsorb to and lyse the host cells as has been shown by Hershey and 153.
Chase (1952) and Bourgaux (1957) to occur with coliphages treated
with low concentrations of formaldehyde.
7. The investigations should be extended to other phages and eventually
herviruses when similarities in the reaction of different viruses to
particular antibacterial agents may indicate the particular mode
of action of these agents or, conversely, similarities in the
detailed structure of the viruses.
8. Information on the mechanisms of the inactivation of viruses by
chemicals may be obtained by a comparison of the kinetics of their
action with those of the attack of proteolytic enzymes on isolated
virus protein and of ribonuclease and desoxyribonuclease on nucleic
acids extracted from viruses. The reaction of free nucleic acidK
will certainly be different from its behaviour in intact virus particle
where it is protected by its associated protein but the comparison of
the kinetics may indicate the main site of attack of en agent
inactivating viruses. 15,
Adams, M.H. (1948) J.gen.PhYsiol., 31, 417. Adams, M.H. (1959) Bacteriophages, Interscience Publishers Ltd., London. Alper, T. (1954). J.gen.Microbiol., 11 313.
Andrewes, C.H. and Horstmann, D.M. (1949) J.gen.Microbiol., 1 290. Asheshov, I.N. et al (1952) Antibiot.Chemother., 2 366. Asheshov, I.N. et al (1954) ibid, 6 380 Astrachan, L. and Volkin, E. (1957) J.Amer.chem.Soc., 22, 130. Bailey, N.T.J. (1959) Statistical Methods in Biology, The English Universities Press Ltd., London. Berns, K.I. and Thomas, C.A. (1961) J.Mol.Biol, 1 277 and 289. Berry,H. and Bean, H.S. (1954) J.Pharm.Pharmacol., L61, 649. Bourgaux, P. (1957) Compt.Rend.Soc.Biol., 251, 1267. Bourke, A.R. et al (1952) J.Immunol., a, 75. Boyd, J.S.K and Bradley, P.L. (1951) Brit.J.Exp.Path., 397. Brazis, A.R. et al (1958) Appl.Microbiol., 6, 338. Briggs, A.(1955) Ph.D Thesis. University of London. Brownlee, K.A. (1949) Industrial Experimentation, H.M.S.O., London. Burnet, F.Land Lush, D. (1940). Aust.J.exp.Biol.med.Sci., 28, 141. Cade, A.R. (1937) J.Amer.pharm.Ass., Li, 1233. Chantrill, B.H. et al (1952) J.gen.Microbiol., 6 74. Cook, A.M. (1954) J.Pharm.Pharmacol., 6, 629. Cook, A.M. et al (1956) ibid, 1 721.
Cook, A.M. and Wills, B.A. (1954) ibid, 6, 638 Cook, A.M. and Wills, B.A. (1956a) ibid, 1 266. Cook, A.M. and Wills, B.A. (1956b) J.appl.Bact., 219, 155.
Cook, A.M. and Frills, B.A. (1959) J.appl.Bact., 22 180. Cook, A.M. and Yousef, R.T. (1953) J.Pharm.Pharmaool., 141, DelbrUck, M. and Luria, S.E. (1942) Aroh.Biochem., 2, 111. De Mars, R.I. (1955) Virology01, 83. De Mars, R.I. et al (1953) Ann.Inst.Pasteur, 113. Demerec, M. and Fano, U. (194.5) Genetics, Lt, 119. Deutsch, M. and Rohr, E. (1955) Arch.Byg., Berl., 212, 644.. Dickinson, L. (194.8) J.gen.Microbiol., 2, 154. Doermann, A.H. (1948) J.Paot., 257. Dunham., W.B. (1957) in Antiseptics, Disinfectants, Fungicides and Chemical and Physical Sterilization, G.F. Reddish, Ed., Kimpton, London. Elderton, P. (1902) Biometrika, 1,155. Ellis, E.L. and Delbrtick, M. (1939) J.gen.Physia., 22 365. Fildes, P. (1940) Brit.J.exptl.Path., 21 67. Fisher, R.A. (1948) Statistical Methods for Research Workers, Oliver and Boyd, Edinburgh. Fitzgerald, R.J and Dabitt, D. (1946) J. Immunol., 52, 121. Gale, E.F. and Taylor, E.S. (1947) J.gon.Microbiol., 1, 77.
Gard, S. (1960)Ann.N.Y Acad.Sci., a, 638. Gilby, A.R. and Few, A.V. (1957) Second Int.Congr.Surf.Activity, Lond., 11, 262. Gots, J.S. (1959) in Bacteriophages, M.H. Adams, Di., Intersoience Publications Ltd., London., Graham, D.M. and Nelson, F.E. (1954) J.gen.Physiol., a, 121. Gratia, A. (1936) Ann.Inst.Pasteur, 57, 652 (through Adams, M.H., 1959). Grossman, L. et al (1961) J.Mol.Biol., 1, 47. 356.
Hadfield, W.A. (1957) in Antiseptics, Disinfectants, Fungicides and Chemical and Physical Sterilization, G.F. Reddish, Ed., Kimpton, London. Hall,E.A. and Asheshov, I.N. (1953) J.gen.Physiol., 17, 217. Ha 11, E.A. et al (1951) Antibiot.Chenother., 20 369. Heivken, K. and Spicier, G. (1956) Zbl.Bakt., Abt.1,0rig., a], 97. Heicken, K. and Spicher, G. (1959) ibid, 112, 11 and 321. dtHerelle, F . (1917) Compt.rend.Acad.Sci. 1 373. (through Adams, M.H.
d'Herelle, F. (1920) Compt.rend.Soc.Biol., Pia, 93. (through Kleczkowskn, J.
Hershey, A.D. (1946) Genetics, 32, 620. Hershey, A.D. and Chase, M. (1952) J.gen.Physiol., Xi, 39. Hershey, A.D. et al (1943) J.Immunol., O, 267. Hofschneider, P.H. (1960) Nature, 186 330. Hotchin, J.E. (1951) J. gen. Microbio1.0 5.2, 609. Hunter, J.G.E. and Whitehead, H.R. (1940) J.Dairy Res., 11 62. Inkley, G.W.(1948) . Private communication in Sykes, G., (1958) Jones, D. (1945) J.Bact., 52, 341. Jones, D. and Schatz, A. (1946) ibid, a 327. Kalter, S.S. et al (1946) J.Bact., 237. Kay,D. (1959) Biochem.J., 21, 149. Kleczkowska, J. (1945) J.Bact., 52, 71. Eleczkwski, A. and AleczkaAski, J. (1951) J.gen.Microbiol., LL, 346. Klein, M, at al (1945) J. Immunol., 2, 389. Klein, M. et al (1948) ibid, 52, 135. Krueger, A.P. and TIaldwin, D.M. (1934) J.gen.Physiol., 21, 499. :157.
Kubinski, H. (1960) Bull.Acad.Polonaise des Soi., 8, 34.7. (through Bull. Hyg., Lond (1961), 16.2 85) Labaw, L.W. et al (1949) Science 110 275. Latarjet, R. and Morenne, P. (1954.) Compt.Rend.Soc.Biol., aa, 791, Luria,S.E. et al (1951) J.Dact., 61, 179. Luzzati, V. et al (1961) J.Mol.Biol., 1 634. McCulloch, E.C. (1945) Disinfection and Sterilisation, Kimpton, London. Marks, H.C. et al (1945) J,Bact., ha, 299. Mather, K. (1949) Biometrics, .50 127. Maurice, P. (1952) Proc.Soc.Appl.Bact., 25, 144. Mayers, V.L.and Spizizen, J. (1954) J.Biol.Chem., 210 877. Miles, A.A. and Misra, S.S.(1938) J.Hyg.,Camb., 732. Mills, R.F.N. (1954) Private communication in Sykes, G. (1958). Mills, R.F.N. (1955) J.gen.Microbiol., 12, 172. Mirsky, A.E. and Ris, H. (1951) J.gen.Physiol., 1, 475. Mutsaars, W. (1957a) Ann.Inst.Pasteur, 20 1.
Mutsaars, W. (1957b) ibid) 93, 754.
Mutsaars, W. (1959) ibid 2§.„ 93. Peacocke, A.R. and Skerrett, J.N.H. (1956) Trans.Faraday Soc., 5Z, 261. Puck, T.T. et al (1949) J.exp.Med., 211, 65. Putnam, F.W. et al (1949) J.biol.Chem., a, 303. Rahn, 0. (1945) Injury and Death of Bacteria by Chemical Agents, Biodynamica, Normandy, Missouri. Salton, Y.R.J. (1951) J.gen.Microbiol., 5, 391 and 405. Sauerbier, W. (1960). Nature, 188 327. Schaffer, F.L. (1960) Ann, N.Y. Acad. Sci., a 564. Schinzel, A. and Lingau, J. (1955) Arch.Hyg., Berl., 21.51, 265. Spizizen, J. (1957) Proc.Nat.Acad.Sci, Wash., 11, 694. 158.
Staehelin, M. (1958) Biochim.Diophys.Acta, ?„51., 410, Stearn, L.E. (1930) J.Bact., 2z 133. Stedmpm,R.L. et al (1957) J.Bact., .7.1 655. Steel, K.J. (1959) Ph.D. Thesis, University of London. Strelitz, F. et al (1955) Proc.nat.Acad.Sci., Wash., 0, 620. Strelitz, F. et al (1956) J.Bact., 21.„ 90. swanstrom, M. and Adams, M.H. (1951) Proc.Soc.exp.Biol., N.Y., 372. Sykes, G. (1958) Disinfection and Sterilisation, Span, London. Tamimi, H.A. (1959) J.infect.Dis., la, 27k.. Tilley, F.W. (1939) J.Bact., 8, 499. Williams—Smith, H. (1951) J.gen.Microbiol., Li, I.58. Williams—Smith, H. (1953) ibid, 8 116. Wills, B.A. (1955) Ph.D. Thesis, University of Lo.adon. Withell, E.R.(1938) Quart.J.Pharm., 11 736. Wooley, J.G. et al (1952) J.immunol., 21.0 523.