Novel high- and low-copy stable cosmids for use in Agrobacterium and Rhizobium

PLASMID

14,

171-175 (1985)

Novel High- and Low-Copy Stable Cosmids for Use in Agrobacterium and Rhizobium DANIEL R. GALLIE, STEVE NOVAK, AND CLARENCE I. KADO’ Davis Crown Gull Group, Department of Plan1 Pathology, University oj”Cal$ornia,

Davis, Calfornia

95616

Received May 23, 1985 Presentedare a set of cosmids based on the unit copy Agrobacterium plasmid, pTAR, and the high-copy-number mutant plasmid, pUCD500, of pTiC58. The addition of apar function derived from pTAR to the vectors allowed them to be stably maintained throughout the cell population in the absence of selective pressure. These vectors, designed for Agrobacterium and Rhizobium, also work in Escherichia coli. The vectors can be cotransfened to Rhizobiaceae from E. co/i with the helper plasmid, pRK2013. The pTiC58 origin containing vectors, pUCDlOO0 and pUCDlOO1 were found to be incompatible with a 250-kb plasmid harbored by R. meliloti RM102Zl. RM 1022 I(pUCD 1000) was still capable of nodulating roots in alfalfa. o 1985 Academic PRSS, lnc.

When designing stable cloning vehicles from novel systems, it is important to characterize the plasmid regions required for autonomous replication and for stable inheritance. Vector stability has been an important problem in maintaining genomic libraries and for analyzing bacterial-plant interactions. Existing cloning vectors for Agrobacterium and Rhizobitlm species (l-6) do not exist at high copy, lack stability, and therefore could present difficulties in maintaining DNA clones throughout the cell population. Recent information on the origin of replication of native Agrubacterium plasmids (7,8) is used here in the design and construction of a set of stable, highand low-copy vectors for use in Agrobacterium and Rhizobium. To construct a low-copy cloning vector which would be compatible with Ti plasmids and would have several selectable markers and unique cloning sites, we began with pUCD9c, which is a Hind111 deletion derivative of pUCDSB, that originated from pUCD5 (1) (Fig. 1). Introduced into pUCD9c was an EcoRI/PvuII fragment containing the tetracycline resistance gene of pBR322, to give rise to plasmid, pUCD497. The pTAR origin of replication was moved as a 2.7-kb EcoRI frag’ To whom reprint requestsshould be addressed.

ment from the pTAR derivative pUCD400 (8) into the EcoRI site of pUCD497, to result in the 9.1-kb cloning vector pUCD2000. The pTAR par locus, which is contained within a 1.3-kb fragment (unpublished results), had been introduced into the polylinker region of pUC4 (9) and is referred to as pUCD550. This 1.3-kb fragment containing par was moved into an EcoRI site of pUCD2000 to give rise to the 10.4-kb vector, pUCD2001, which is the stable counterpart to pUCD2000. The pTiC58 copy number mutant pUCD500 (7) was used to construct a high-copy-number vector. The Hind111 fragment of pUCD500 containing the pTiC58 ori and inc functions was introduced into the Hind111 site of pUCD9c, to result in the 9.5-kb vector, pUCD 1000 (Fig. 2). pUCDlOO0 contains a unique KpnI site. A unique Sac11 site was introduced into pUCD1000 by the removal of a 0.25-kb Sac11 fragment, to result in pUCD IOOO-1. pUCD 1000-2, containing a unique XbaI site, was constructed by linker conversion of the unique Sac11 site of pUCD 1OOO-1. Both pUCD1 OOO-1 and pUCD1000-2 still contain a unique KpnI site. pUCD 1000 is high copy but par- and therefore not stably inherited. The functioning of a partition mechanism is independent of the origin of replication. The pTAR par frag171

0147-619X/85 $3.00 Copyright 0 1985 by Acadcm~ Press. Inc. All rights of rrpmduction in any form rewrved.

172

SHORT COMMUNICATIONS

H

FIG. I. Construction of low copy cloning vectors pUCD2000 and pUCD2001. Construction is described in the text. Ap, ampicillin; Cm, chloramphenicol; Km, kanamycin; Nm, neomycin; Tc, tetracycline; cos. the X cos site; B, BumHI; RI, EcoRl; H, HindlIl; K, Kpnl; P, AwlI; Ps, Pstl; s, Sad; Sl, Sucl.

ment was therefore introduced as a BamHI fragment into a BarnHI site of pUCD1000 to result in pUCD 1001. This plasmid was markedly more stable than pUCD 1000, but did not exhibit a complete pTAR par’ phenotype. Introduction of the pTAR par fragment will completely stabilize pTAR par- or pSa par plasmids (unpublished results). Therefore, it is possible that the copy number mutation of the pTiC58 origin region has caused pUCD 1001 to behave in an unpredicted manner. When a second copy of the pTAR par locus was introduced in a tandem fashion with the first copy also within the BarnHI site, the resulting plasmid, pUCD 1002 was completely stabilized. The stabilities of the vectors in A. tumefaciens LBA430 1 and in R. meiiioti RM 1022 I were quantitatively determined. As antici-

FIG. 2. Construction of the high-copy cloning vectors pUCD 1000 and pUCD 100I. Construction described in the text. Abbreviations as described in Fig. I caption.

pated, pUCD 1000 and pUCD2000 are parand therefore were not stably inherited (Fig. 3). Both vectors are lost from the cell popu-

Generations

FIG. 3. Segregation of cloning vectors from A. tumefuciens LBA4301 under nonselective growth conditions. Data plotted as the percentageof cells retaining the vector against the number of generations of growth under nonselective (no antibiotics) conditions. Vector segregation data identical for RM 10221, except for pUCD1002 (see text). A, pUCD1000; 0, pUCD1001; A, pUCDlOO2; l , pUCD2000; 0, pUCD200 1.

SHORT COMMUNICATIONS

lation by the 50th generation of nonselective growth. On the other hand, pUCD1002 and pUCD200 1 were completely stabilized by the presence of the pTAR par locus. pUCD 1001 is intermediate in stability between pUCDl000 and pUCD1002. The stability of the vectors in R. meliloti RM 1022 1 is similar to that seen for LBA430 1, with the exception of pUCD1002. As RM 10221 is recombination proficient, the tandem arrangement of the par loci can resolve into a unit copy of the par locus and became identical to pUCDl001. Figure 4 shows qualitatively the relative copy number of the four vectors. Lane 1 rep-

FIG. 4. Plasmid isolation FromA. tumefaciens LBA4301 showing relative copy number differences between the vectors. Growth conditions and plasmid isolation were identical for eachstrain. Lane 1,pUCD500; 2, pUCD1000; 3, pUCDIOO1; 4, pUCD2000; 5, pUCD2001, 6, pUCD400. chr, chromosomal DNA. Open circle and linear forms of plasmid DNA are also present in lanes 1, 2, and 3. Plasmid DNA was isolated by the method of Kado and Liu (IO).

173

resentsthe pTiC58 high-copy-number mutant, pUCD500, which was the progenitor of pUCDlOO0 (Lane 2) and pUCDlOO1 (Lane 3). Lane 6 representsthe ori containing pTAR derivative, pUCD400, which was used in the construction of pUCD2000 (Lane 4) and pUCD200 1 (Lane 5). The copy number of each cloning vector was quantitatively determined (Table 1). It has been consistently observed that the copy number of plasmids containing this pTiC58 high-copy mutant ori is inversely proportional to the size. Therefore the copy number of pUCDlOO0 and pUCD 1001, at 40 and 37, respectively, is lower than the 80 copies/cell for the smaller sized pUCD500. pUCD2000 and pUCD200 1 contain a wildtype pTAR ori. No variation in copy number has been observed for vectors containing this ori. The small apparent increase in copy number of pUCD2001 over pUCD2000 is due to the stabilizing presenceof the pTAR par loci, causing the population of plasmid carrying cells to increase from 80% to 100%during selective growth conditions. In R. meliloti RM 1022 1, pUCD2000 and pUCD2001 remained as unit copy plasmids whereas pUCDl000 and pUCDl001 were present at 29 and 28 copies, respectively. Host influence on the high-copy mutant pUCD500 have been observedpreviously (7) which may account for the lower copy number in Rhizobium than in Agrobacterium. The pTiC58-origin-based vectors retain the Rh- 1 incompatibility function and are incompatible with plasmids belonging to this group. When pUCDl000 or its derivatives were introduced into RM 1022 1, a resident 250-kb plasmid was no longer detected by agarose electrophoretic analysis (data not shown), suggesting that the 250-kb plasmid may belong to inc Rh-1. Nevertheless, RM102Zl(pUCD 1000) still retained root nodulating ability on alfalfa roots (Medicago sativa cv. Moapa 69). This would suggestthat either the 250-kb plasmid is not involved in nodulation or no longer is present in an episomic form. No incompatibility was seen between pUCD2000 and the 250-kb plasmid.

174

SHORT COMMUNICATIONS TABLE I VECTOR CHARACTERISTICS

Unique cloning sites

Size

COPY

Vector

(kb)

number”

Stability (% loss&en)

pUCDlOO0 pUCDIOO1 pUCDl002 pUCD2000

9.5 10.8 12.1 9.1

40 37 37 0.9

2.3 0.9 0.0 4.3

KpnI, Sac11b, XbuI’

pUCD200 I

10.4

0.0

ClaI, KpnI, PsfI, Sad,

KpnI

KpnI BamHI, ClaI, KpnI, PstI,

SUCI,sun 1.0

sun

’ Copy number determined as described previously (7). bAdditional sites available in pUCDlOO0 derivatives pUCDlOOO-1 and pUCDlOOO-2.

Becausethe vectors, bearing either the origins of pTAR or pTiC58 also contain the ColEl type origin of replication, DNA manipulations may be carried out in Escherichia coli as well as Agrobacterium or Rhizobium. Plasmid DNA can then be introduced into Agrobacterium or Rhizobium by way of transformation. Alternately, the basis of mobilization, located near the ColEl type origin, makes the vectors capable of conjugation into members of the Rhizobiaceae with the aid of the helper plasmid RK2013 (3). pUCDl000, pUCD 1001, and pUCD 1002 contain the neomycin phosphotransferase II (NPTIQ2 gene from Tn5 and the kanamycin resistance gene from pSa. Both genes are alike in that they both encode kanamycin resistance.They differ in that the NPTII gene also encodes for neomycin but not gentamicin resistance and the pSa kanamycin resistance gene encodes for gentamicin but not neomycin resistance. Thereby, using neomycin or gentamicin, these two kanamycin resistance genescan function as different selectable markers. The vectors described have proven useful in the construction of Agrobacterium chromosomal cosmid banks with an average DNA insert size of 33 kb. In addition, the par- phenotype of pUCD2000 has been useful in isolating the par function from pTiC58 (unpublished re-

sults). The high-copy nature of pUCD1002 has been usedto study regulation of a pTiC58 gene involved in virulence. The vir gene, normally repressedin A. tumefaciens, was found to escape repression when carried by pUCD1002 (T. J. Close, personal communication). The incompatibility function (inc Rh-1) associated with the pTiC58 origin region has made pUCD 1000 and its derivatives useful in curing other inc Rh- 1 plasmids from Agrobacterium strains which were otherwise difficult to remove by standard curing techniques. Moreover, pUCD 1001 and pUCD200 1 should facilitate the stable maintenance of genomic libraries of Rhizobium and Agrobacterium. Materials and Methods. Bacterial strains, media, and plasmids. A. tumefaciensLBA430 1 (Ret-, rifr, UV”, vir-) was obtained from R. A. Schilperoort. R. meliloti RM 10221 was obtained from D. Helinski. pUCD9c, a derivative of pBR322 wasconstructed by D. Zaitlin. Media 79 was used to grow RM 1022 1 for transformation and plasmid isolation. Antibiotic concentrations for LBA430 1 were 5 pg/ml tetracycline, 30 &ml neomycin or kanamycin, and 75 pg/ml gentamycin; 5 Irglml tetracycline or 500 cLg/mlneomycin was used for RM 1022 1; 10 &ml tetracycline, 30 pg/ml neomycin or kanamycin, 3 pg/ml gentamycin, and 50 pg/ml ampicillin were used for HBlOl. Maintenance assay and copy number deter2 Abbreviation used: NPTII, neomycin phosphotransmination. Copy number and maintenance asferase II.

175

SHORT COMMUNICATIONS

sayswere described previously by Gallie et al. (7). RM 10221 recipients containing different plasmids were allowed to grow for 48 hr between assaysfor plasmid maintenance. Plasmid detection and isolation. The plasmid detection and isolation procedure of Kado and Liu (10) was modified for Rhizobium as follows. A 350-~1aliquot of RM 1022 1 culture was centrifuged and the cell pellet washed in 0.5 ml of 0.5% sarkosyl in 100 mMTris-HCl, 100 mM EDTA, pH 7.0. After pelleting, the cells were treated with 100 ~1 of a 5 mg/ml lysozyme solution in 10 mMTris-HCl, 1 mM EDTA, pH 7.5, for 30 min on ice. The solution was then treated as previously described. Transformation. The method used to transform Agrobacterium, described by Kao et al. (11) was modified for Rhizobium by substituting medium 79 as the growth media. Root nodulating assay. Surface-sterilized alfalfa seeds(Medicago sativa cv. Moapa 69) were sprouted and placed in Lenard jars containing a sterilized 50% vermiculite/50% perlite mixture. Rhizobium liquid culture was poured on the sprouted seedsand infectivity data collected after 6 weeks by observing the formation of root nodules.

ACKNOWLEDGMENTS We thank Dr. A. Dendenkr for the root nodulating assay protocol for Rhizobium, and Jeff Hall and Paul Hara for technical assistance. This work was supported by NIH Grant CA- I I526 from the National Cancer Institute.

REFERENCES 1. CLOSE, T. J., ZAITLIN, D., AND KADo, C. I., Plasmid 12, Ill-118(1984). 2. DITTA, G., SCHMIDHAUSER, T., YOKOBSON, E., Lu, P., LIANG, X.-W., FINLAY, D. R., GUINEY, D., AND HELINSKI, D. R., Plusmid 13, 149-153 (1985). 3. DITTA, G., STANHELD, S., CORBIN, D., AND HELINSKI, D. R., Proc. Natl. Acad. Sci. USA 17,7347-

7351 (1980). 4. FRIEDMAN, A. M., LONG, S. R., BROWN, S. E., BUIKEMA, W. J., AND AUSUBEL, F. M., Gene l&289-

296 (1982). 5. KNAUF, V. C., AND NESTER, E. W., Plasmid 8,45-

54 (1982). 6. TAIT, R. C., CLOGE,T. J., LUNDQUI~T, R. C., HAGIYA,

7. 8.

9. 10.

M., RODRIGUEZ, R. L., AND KADO, C. I., Bio/ Technology 1,269-275 (1983). GALLIE, D. R., HAIGIYA, M., AND KADO, C. I., J. Bacteriol. 161, 1034-1041 (1985). GALLIE, D. R., ZAITLIN, D., PERRY, K. L., AND KAW, C. I., J. Bacteriol. 157,739-745 (1984). VIEIRA, J., AND MESSING, J. Gene 19,259-268 (1982). KADO, C. I., AND LIU, S.-T., J. Bacterial. 145, 1365-

1373 (1981). 11. KAO, J. C., PERRY, IS. L., AND KADO, C. I. Mol. Gen. Genet. l&425-432 (1982).