Release and Uptake of Gene Transfer Agent by

Transcription

1 JOURNAL OF BACTERIOLOGY, Aug. 1975, p Vol. 123, No. 2 Copyright 1975 Amerian Soiety for Mirobiology Printed in U.S.A. Release and Uptake of Gene Transfer Agent by Rhodopseudomonas apsulata MARC SOLIOZ, HUEI-CHE YEN, AND BARRY MARRS* Department of Biohemistry, St. Louis University Shool of Mediine, St. Louis, Missouri Reeived for publiation 4 Marh 1975 Many strains of Rhodopseudomonas apsulata are apable of exhanging geneti information via a reently disovered gene transfer proess involving the release and subsequent uptake from the medium of partiles ontaining geneti information (gene transfer agents, GTAs). No viral ativities are observed to be assoiated with this system. An assay has been developed to quantitate gene transfer in R. apsulata. Conditions are desribed for whih the number of ells aquiring a new geneti trait is diretly proportional to the number of GTAs and independent of the number of reipient ells. These onditions were used for the assay of the uptake and release of GTAs by ells. The maximum fration of reipients that aquire a given geneti marker is -4 x Free GTA appears in a growing ulture in one or two abrupt waves near the time of transition from exponential to stationary phase. During these waves, the titer of GTA for a given marker may reah 2 x 106ml. A omparison of the frequeny of singleand double-marker transfers suggests that most of the ells in early-stationaryphase ultures are ative reipients. The ultraviolet inativation spetrum of GTA resembles that of the small ribonulei aid phages. The inativation ross setion a, for GTA was alulated to be 1.7 x 10-1 m 2photon at 265 nm. We have reently desribed the disovery of the first geneti exhange system for a nonsulfur purple photosyntheti baterium (3). Certain strains of Rhodopseudomonas apsulata are apable of releasing partiles that ontain samples of geneti information representing all parts of the genome. These partiles an subsequently provide geneti information to other bateria of the same speies, thus effeting geneti transfers. The proess resembles transdution; however, the partiles are muh smaller than any known transduing bateriophage. We propose to all these partiles gene transfer agents (GTAs), and ells that express new geneti markers reeived via GTAs are termed transferants. It is our goal to be able to use this geneti system to study energy metabolism and the regulation of photosyntheti membrane formation via a biohemial genetis approah. This report desribes the development of a quantitative assay for R. apsulata gene transfer, the use of the assay to study some of the harateristis of the system, and further evidene about the physiohemial nature of the vetor mediating the gene transfer. MATERIALS AND METHODS Baterial strains. R. apsulata B10 and B6 have been desribed previously (3). The following strains 651 were derived from B10: BB101 (arries marker rif-10 whih onfers resistane to rifampin [RifR]); BB102 (arries marker str-1 whih onfers resistane to streptomyin [StrRI); BB1012 (arries rif-10 and str- 1); YB1020 (arries str-2 whih onfers streptomyin resistane and trpa20 whih auses tryptophan auxotrophy [Trp ]). Media. Peptone-yeast extrat medium (PYE; 0.3% peptone [Difo] and 0.3% yeast extrat [Difo] in deionized water) was used for growth of reipient and donor ultures. For strain YB1020 this was supplemented with 10 Ag of L-tryptophan per ml. RCV, a minimal salts medium (5), supplemented with 0.3 ug of L-tryptophan per ml was used as a seletive medium to detet tryptophan-independent ells. Gene transfer assay. Reipient ells were harvested from early-stationary-phase photosyntheti ultures by entrifugation and then suspended in buffer ontaining 1 mm MgCl,, 1 mm CaCl2, 1 mm NaCl, and 10 mm tris(hydroxymethyl)aminomethanehydrohloride (ph 7.8) (G buffer) supplemented with 500 Mg of bovine serum albumin (Sigma Chemial Co., fration V) per ml. Reipients were routinely harvested shortly before use, but oasionally were stored on ie for several hours without serious loss of ativity. GTA was prepared routinely by filtering a donor ulture through a 0.45-Mm membrane filter, sine removing ells by low-speed entrifugation results in a 50% loss in gene transfer ativity ompared to membrane filtration of the same ulture. The resulting ell-free filtrate may be used diretly or purified further to remove inhibitory substanes. GTA may be

2 652 SOLIOZ, YEN, AND MARRS partially purified and onentrated by diafiltration against G buffer plus 500 jg of bovine serum albumin in an amion thin-hannel filtration apparatus fitted with an XM300 membrane. Diafiltered onentrates may be stored at -70 C provided bovine serum albumin is present. Reipient ells and donor filtrate were mixed in G buffer and inubated aerobially at 35 C for about 60 min. During this inubation, the proess of gene transfer is sensitive to inhibition by a wide variety of substanes, inluding peptone and yeast extrat, so the inubation mixture is onstituted to minimize arry-over of residual medium from donor and reipient ultures. After this inubation, reipients were diluted, plated, and inubated at 35 C under nonseletive onditions to allow phenotypi expression of newly aquired geneti traits. Different geneti markers require different times for phenotypi expression. Rifampin resistane gene transfers were titered by plating reipients in a molten, soft-agar layer on PYE plates and inubating for 4 h at 35 C before overlayering with molten agar ontaining suffiient rifampin to bring the final onentration to 67 ugml. Prototrophi transferants were seleted from tryptophan auxotrophs by plating diretly in RCV agar supplemented with 0.3,g of L-tryptophan per ml. This level of tryptophan allows phenotypi expression but limits the growth of the parental auxotrophs to miroolonies. UV light irradiation. The stoks of bateriophage T2 and GTA used throughout the experiments had titers of 5 x 10 1 plaque-forming units per ml and 2 x 106 rifampin resistane gene transfer units per ml, respetively. The GTA stok was prepared by onentration and diafiltration as desribed above. All dilutions were done with G buffer. T2 and GTA stoks were diluted and mixed to give a titer of about 106 plaque-forming units per ml for T2 and 2 x 104 resistane gene transfer units per ml for GTA. Portions (1 ml eah) of this mixture were irradiated in a quartz uvette (1 by 1 by 4 m) with magneti stirring. As a soure for monohromati light, we used a spetrophotofluorometer (Amino-Bowman) equipped with a high-pressure xenon lamp. A dose of 100 ergsmm2 at 260 nm was delivered in approximately 200 s. The irradiated samples were assayed for the remaining GTA gene transfer ativity as desribed above. T2 plaque-forming ability was measured by plating with E. oli S. All handling of ultraviolet (UV)-irradiated samples was done in red light to avoid possible photoreativation. RESULTS Assay onditions. Two riteria were used in developing the assay. First, onditions should be found for whih the number of transferants is diretly proportional to the amount of input GTA and invariant with the number of potential reipient ells present. Seond, onditions should be optimized so that a fixed number of GTAs gives rise to a maximum number of transferants. The assay onditions desribed in Materials and Methods meet the first of these J. BACTERIOL. goals as demonstrated by the results shown in Fig. 1 and 2. When a onstant number of reipient bateria are exposed to inreasing numbers of GTAs (Fig. 1), the resulting number of transferants inreases proportionately until about 4 x 10-5 of the ells have expressed a given gene transfer. Further inreases in GTA per reipient result in a nonlinear inrease in transferants until the frequeny of transferants reahes a maximum at about 3 x 10-i to 4 x This saturation ours at GTAreipient ratios on the order of 100 times greater than the maximum ratio giving a linear response. The values on the linear portion of the doseresponse urve may be used to alulate a "titer" of the number of rif-10 GTA per ml of solution; however, sine the overall effiieny of gene transfer is not yet known, this titer represents only assayable rif-10 transfer units. Furthermore, sine the titer is measured for only one geneti determinant, the total number of GTA, is greater than this titer by at least a fator equal to the moleular weight of the genome divided by the average GTA equivalent moleular weight. We estimate this fator to be large (103 to 104) beause a 70S partile (3) with a buoyant density of gm3 in CsCl (un- 30 _ _ Input r,f-lo GTA per pl olulated from inear region FIG. 1. Number of rifampin resistane gene transfer events among a onstant number of reipients treated with various amounts of GTA. Rifampin-sensitive reipient ells (B6) were grown, harvested, and suspended at a final density of 9 x 107 olony-forming units (CFU) per ml in G buffer plus 500 Ag of bovine serum albumin per ml ontaining various dilutions of a onentrated, diafiltered preparation of GTA from strain BB101. The mixture of ells and GTA was inubated for 2 h at 35 C and then diluted and plated for determination of rifampin resistane. The bakground frequeny of spontaneous rifampin resistane in this experiment was less than 4.0 CFUml. The input titer of GTA is alulated assuming one RifR CFU arises from one rif-10 GTA in the region where the dose-response urve is linear, and is an "assayable" titer only. The inset is an expansion of the initial portion of the urve.

3 VOL. 123, 1975 published data) would be expeted to arry a nulei aid moleule of relatively small size (-10 Ito 106 daltons). Above a ritial ratio of reipients per GTA, the assay is insensitive to variations in the number of reipients (Fig. 2). At higher reipient ell onentrations, 02 beomes limiting and a gradual derease in transferants is observed (data not shown) unless steps are taken to improve aeration. We have also observed that 1 mm KCN strongly inhibits the gene transfer proess (data not shown), and these two results taken together suggest that this proess requires energy metabolism during the initial stages. Our suess in meeting the seond assay riterion, that of maximal response, annot be judged until more of the moleular details of the system are known. We have examined the effets of ioni omposition and onentration (and ph) on both the stability of GTAs and the assay system itself, and the onditions desribed above seem optimal. We have been impressed by the wide variety of normally innouous agents that inhibit the system; e.g., if either 50 mm NaCl or 0.3% peptone is added to the omplete assay system, a marked inhibition of transferant prodution is observed (see below). GTA uptake. Figure 3 shows the results of an experiment in whih the rate of appearane of potential transferants is ompared to the disappearane of free, assayable GTA in our standard assay mixture. Potential transferants are measured as the number of ells that are suffiiently ommitted to gene transfer so that dilution and plating in G buffer soft agar does not interrupt the proess. A small orretion has been applied to eliminate those gene transfers atually initiated in the soft agar after plating. Sine the kinetis of disappearane of free, assayable GTA omplements the appearane of potential transferants, we tenatively identify the disappearane of GTAs with their uptake by reipients. The ontinued rise in transferant titer after all GTA has been taken up (about 30 min) is paralleled by an inrease in total ell number and may represent the repliation of newly aquired geneti markers. If diluted samples from an assay mix are plated diretly into PYE-soft agar, we observe the previously mentioned inhibitory effet of PYE. The separation of these two transferant appearane urves suggests that the PYE-sensitive phase is over about 10 min after uptake of GTA. Ative reipient population. Analysis of the frequeny of double gene transfers (Table 1) is interpreted to indiate that nearly all of the ells in the population are ative reipients in E GENE TRANSFER IN R. CAPSULATA u _ I -jo ^ * o * o 0 * Reipient Cell Conenlrotion CFU ni J - -~-0 t^ sa r^ ^A ^r FIG. 2. Number of rifampin resistane gene transfer events indued by a onstant amount of GTA in various numbers of reipient bateria. The gene transfer assays were performed as desribed in the legend to Fig. 1, exept that eah assay ontained the same amount of GTA and various amounts of a washed reipient ell suspension. The same amount of the same preparation of GTA (stored at 70 C) was assayed on two different days using different ultures of reipients: assay 1, 0; assay 2, 0. CFU, Colony-forming units. our assay. If we assume that the gene transfers for rifampin resistane and tryptophan independene are independent events, i.e., no geneti linkage between these markers and no exlusion of a seond gene transfer by a first, then we an ompare the observed frequeny of double events (9.6 x 10-9) to the expeted frequeny alulated as the produt of the frequenies of the single events (7.2 x 10-9) We onsider this agreement to be exellent and to indiate that total ell number is a good estimate of the number of ative reipients in the assay. GTA prodution kinetis. To determine when during the growth yle GTA was released, samples of ultures were filtered and assayed for GTA at various times after inoulation of photosyntheti ultures of strain BB101 or BB1012. The surprising result is shown in Fig. 4, whih summarizes the data from four independent trials. Free GTA onentration inreases abruptly in filtrates shortly before the transition from log to stationary phase. However, before stationary phase is reahed, GTA titer dereases as abruptly as it appeared, only to reappear at even higher titers as turbidity beomes onstant. Oasionally the seond wave of free GTA does not materialize, but the peak of GTA titer during the transition from exponential growth to stationary phase has been observed in all trials to date. The sharp drop in GTA titer at about 16 h might be due to a short, intense period of GTA uptake by the produing ulture. To examine this possibility, we removed ell samples from ultures during this

4 654 SOLIOZ, YEN AND MORRIS 4 '-3 0 -~ Time (min) FIG. 3. Kinetis of GTA uptake and appearane of potential transferants. Three flasks of G buffer plus 500 ug of bovine serum albumin per ml were inubated at 35 C with gyratory mixing. At zero time, flask I reeived reipient ells (strain B6) and GTA (diafiltered; from strain BB101), flask II reeived GTA alone, and flask III reeived ells alone. The zero-time ell onentration in I and III was 4.7 x 108 olony-forming units per ml. The imput GTA titer in I and II was 5.7 x 103 rif-10 GTA per ml. At various times, samples were removed and assayed for unadsorbed GTA and ells ommitted to beome rifampinresistant transferants. Unadsorbed GTA were measured by filtering the ell-gta mix on 0.45-gum ellulose aetate membrane filters and assaying the filtrate for GTA (0). Potential transferants at various times were measured by diluting a sample from flask I 10-fold in G buffer plus 500,ug of bovine serum albumin per ml and then 12.5-fold into G buffer plus 0.6% agar (molten) in an empty petri plate. These plates were inubated at 35 C (after the agar gelled) until 2 h had elapsed from the initial mixing of ells and GTA, at whih time all plates were overlayered with 10 ml of 0.3% PYE soft agar and inubation ontinued for 4 h at 35 C. Then PYE-soft agar ontaining rifampin was poured over the plates as usual, and the plates were inubated until rifampinresistant olonies formed. Dilution and immobilization in agar were intended to prevent further assoiation between ells and GTA. To test the effetiveness of this proedure, small volumes were removed from flasks II and III, similarly diluted, mixed on empty petri plates with the addition of 5.0 ml of G buffer-soft agar, and then treated exatly as above. Transferants arising from plate adsorption were most numerous at period and tested the ells for their ability to bind GTA. Cells were spun down, suspended at a onstant turbidity, and exposed to a known saturating titer of genetially marked GTA. After suffiient time for omplete adsorption, mixtures were filtered and free GTAs were titered. This experiment (data not shown) indiated that ells from ultures in transition stage had the same adsorption ativity as early-stationary-phase ells. It would thus seem that flutuations in the rate of GTA uptake do not aount for the observed flutuation in free GTA titer. Furthermore, strain H9, whih produes but does not adsorb GTA under the onditions desribed in Fig. 3, shows flutua- TABLE 1. Frequeny of gene transfer J. BACTERIOL. New phenotypesa CFUml' Frequeny Single events, GTA indued Rif'-RifR 1.1 x 10' 1.0 X 10 4 Trp- -Trp+ 8.0 x x 10 5 Single events, spontaneousd RifS -RifR Trp- -Trp+ 2.5 x 10' 8.7 x x x 10-' Double events, GTA indued, observed Rifs' Trp --Rif'8 Trp x 10e 9.6 x 10 9 Double events, GTA indued, alulated' Rif'sTrp RifR Țrp x 10 Double events, spontaneous, observedd Rif' Trp -.Rif, Trp 08 <7 x 10' ayb1020 ells, 1.1 x 109ml (optial ount), were treated with filtrate from BB101 ontaining 2 x 106 rif-10 GTA per ml, and transferants were seleted. b CFU, Colony-forming units. Frequeny of partiular new phenotype obtained by dividing onentration of new-type ells by total ell onentration. d No GTA added. e 106 CFU observed. I Expeted frequeny alulated as produt of single-event frequenies; CFU per milliliter alulated as: expeted frequeny times total ells per milliliter. 9 No CFU observed per 1.4 x 109 reipients. early times (650mI at 1 min) and dereased gradually as inubation ontinued. The differene between the number of RifR olony-forming units on these two types of plates was plotted as transferants per milliliter (0). Samples from flask I were also removed, diluted, plated diretly in PYE-soft agar (instead of G buffer-soft agar), and then treated as desribed for the previous plates. The numbers of RifR olony-forming units appearing on these plates are also plotted (A).

5 VOL. 123, 1975 E 105 I - a *v Time (hrs) FIG. 4. Time ourse of GTA release. Samples photosynthetially growing ultures of strains BBI 0of or BB1012 were diluted 1:3 in G buffer plus 500 Ag of bovine serum albumin per ml, filtered through um ellulose aetate membranes, and frozen - 70 C. The following day, samples were thawed atnd assayed for either rif-10 (0, 0, A) or str-1 (V) G7"A. BBIOJ ultures (0, 0, A) were grown in srew- ulture tubes, two tubes being harvested for eah ti? point. Strain BB1012 (V) was grown in a plas syringe, and samples were removed periodially by depressing the plunger. Culture growth was followed by turbidity measurements, whih are plotted relatiive to the stationary-phase turbidity. Time sales wi then normalized so that 50%o maximal turbid; 'ity ourred at a fixed time. The resulting growth urt)es were essentially superimposable and are depited by GENE TRANSFER IN R. CAPSULATA 655 produing ulture but is not ative in the in vitro mixture. Inativation of GTA by UV light. The rate of inativation of gene transfer ativity of GTA by UV light was determined. Sine suh measurements depend strongly on the intensity of the light soure used, the extent of internal absorption and light sattering of the sample, and the sample geometry, we alibrated our - system with bateriophage T2, whih has a w known UV sensitivity. Idential irradiation on- ditions were assured by using a mixture of T2 9 and GTA throughout the experiments. Figure 5 depits inativation urves at 260 nm. Appropriate ontrol experiments showed that T2 and GTA did not interfer with eah other in either the assays or the irradiations. The experimental urves an be represented by the equation n = no exp - ad, where no is the initial titer and n is the titer after a dose (D) is delivered; the inativation ross setion, a, is a onstant whih is proportional to the sensitivity of a partiular biologial material (e.g., virus) at a given wavelength and orresponds to the slope in Fig. 5. The dose was alulated from the a value for T2 as determined by Rauth (6). The inativation of GTA was found to be exponential over at least three logs. at Ation spetrum. In Fig. 6, the inativation ross setions for GTA and T2 are plotted as a funtion of wavelength. The inativation speap trum of GTA shows a broad maximum at 265 me nm and is very similar to that of T2 and other ti viruses, suggesting that nulei aid is the target of the UV inativation. The value of a for ere 100 tions in free GTA titer similar to those shown in Fig. 4. The possibility remains that onditi)ns \ of GTA uptake in the ulture are not preservied E lo _ in our uptake assay and, therefore, uptake mlay _ still aount for the unusual hanges in titer A seond possible explanation for the drop in GTA titer at about 16 h is that an inhibitor of our assay might be produed and later ( deple stroyed. This has been ruled out by a sim] mixing experiment whih showed that exolge nous, genetially distint GTA, when mixed with filtrates of samples from a growing ultuire, Dose (ergninm2) retain a onstant titer, whereas the endogen( DUS FIG. 5. UV inativation of bateriophage T2 and geneti marker flutuates as shown in Fig. 4. GTA at 260 nm. Symbols: 0, perent survival of T2 This leaves open the possibility of a labpile plaque-forming units; 0, perent survival of rifampin inativating fator whih destroys GTA in tthe resistane gene transfer ativity.

6 656 SOLIOZ, YEN AND MORRIS -C 2 l, E z 2 0 > 101 v C) U 2U Wavelength (nm! FIG. 6. Ation spetrum of inativation of the plaque-forming ability of bateriophage T2 (0) and rifampin resistane gene transfer ativity (0). T2 at 265 nm is about 42 x m'photon; that for GTA is about 1.7 x m2photon. The ability of GTA to transfer genes is thus approximately 25 times more resistant to inativation by UV light of that wavelength than the plaque-forming ability of T2. Of the phages for whih has been determined (6), the small ribonulei aid phages MS2, R-17, fr, and 7-S most losely resemble GTA, eah exhibiting an inativation ross setion of about 1.2 x m2photon, whereas, among the deoxyribonulei aid phages, IX174 and fd, with a values of approximately 10 x m2photon, are most similar to GTA in their UV sensitivities. DISCUSSION The quantitative gene transfer assays desribed here have permitted the study of prodution, uptake, and UV inativation of GTA and have helped define onditions for maximal effiieny of geneti exhange with this system. The high frequeny of gene transfer per reipient (-4 x 10-4 for a single marker with exess GTA) ompares favorably with most generalized transdution systems and is high enough to be used easily for strain onstrution and geneti analysis. The observation that gene transfer reipients are saturated when only a small fration of the ells has aquired a partiular new marker may reflet an inherent low effiieny for inorporation of inoming geneti material or it may be due to a limited number of T2 290 J. BACTERIOL. adsorption sites for GTA on reipient ell surfaes. The analysis of double gene transfer events makes it seem unlikely that a small ative fration of the reipients aounts for all the geneti exhange. The unusually abrupt appearane and disappearane of GTA in growing ultures are intriguing. It does not resemble the phage prodution seen in either lysogeni ultures (2) or infeted ultures, whih release small phage ontinuously (4). If it is assumed that eah partile is relatively effiient in gene transfer and thus the titer by bioassay of a partiular geneti marker is a good estimate of the number of genome equivalents released, it is possible that the lysis of only 105 to 106 ells per ml may aount for all the observed GTA prodution, and thus some sort of lyti phenomenon annot be ruled out as the mehanism for release of GTA to the medium, sine lysis of this small fration would not have been deteted. Alternatively, living ells may shed GTA into the medium without lysing. Release by either mehanism learly involves a signifiant degree of synhrony whih remains unexplained. The rate of inativation of GTA ativity by UV light is omparable to that observed for transduing ativity (1), but transduing ativity usually shows an ativation by low doses of UV light whih is not observed with GTA. The UV inativation spetrum of GTA is similar to that for baterial viruses. Rauth (6) notied an empirial orrelation between the shapes of UV inativation spetra and strandedness of nulei aids arried by viruses, namely, the ratio for a values was 2 for double-stranded and 1 for single-stranded nulei aids with very few exeptions. The GTA spetrum resembles the single-stranded lass. Work on the diret haraterization of the nulei aid of GTA is in progress. The gene transfer system of R. apsulata would seem to be either generalized transdution by a defetive virus of a size lass hitherto not known to transdue or a baterial sex mehanism unrelated to viral ativities. The definitive resolution of these two possibilities will probably be diffiult, but in the absene of evidene for viral involvement we favor onsidering this system as a baterial mehanism that exists beause of the seletive advantage whih the apability for geneti exhange might onfer. We have disovered reently that GTA an be used to demonstrate geneti linkage between the genes for arotenoid biosynthesis and those for bateriohlorophyll biosynthesis, and mapping of this important region of the R. apsulata genome is in progress.

7 VOL. 123, 1975 GENE TRANSFER IN R. CAPSULATA 657 ACKNOWLEDGMENTS We thank Sandra Bilyeu for her exellent tehnial assistane, N. Melehen, J. Wall, P. Weaver, and H. Gest for many helpful and stimulating disussions. This work was supported by Publi Health Servie grant GM from the National Institute of General Medial Sienes and by grant GB from the National Siene Foundation. LITERATURE CITED 1. Benzinger, R., and P. E. Hartman Effets of ultraviolet light on transduing phage P22. Virology 18: Jaob, F., and E. L. Wollman Lysogeny, p In F. M. Burnet and W. M. Stanley (ed.), The viruses, vol. 2. Aademi Press In., New York. 3. Marrs, B. L Geneti reombination in Rhodopseudomonas apsulata. Pro. Natl. Aad. Si. U.S.A. 71: Marvin, D. A., and B. Hohn Filamentous baterial viruses. Bateriol. Rev. 33: Ormerod, J. G., K. S. Ormerod, and H. Gest Light-dependent utilization of organi ompounds and photoprodution of moleular hydrogen by photosyntheti bateria; relationships with nitrogen metabolism. Arh. Biohem. Biophys. 94: Rauth, A. M The physial state of viral nulei aid and the sensitivity of viruses to ultraviolet light. Biophys. J. 5: