ML20070S103

Doubly Labeled Water (3-HH-18-0) Method:Guide to Use, Ucla Publication 12-1417

ML20070S103
Person / Time
Site:	Purdue University
Issue date:	06/30/1983
From:	Nagy K CALIFORNIA, UNIV. OF, LOS ANGELES, CA
To:
Shared Package
ML20070S086	List: ML20070S083 ML20070S088 ML20070S103
References
NUDOCS 9104020139
Download: ML20070S103 (43)

Text

_, , .. . _ _ , . _ _ _ . . - - . - - - - - - -- - -- - ---- - - - ~ ~ ^ - - - ' ~ ~ - " ~

~ $ / U L L.t 6 zqO THE DOUBLY LABELED WATER 180(31!H

) METHOD:

A GUIDE TO ITS USE Kenneth A. Nagy Laboratory of Biomedical and Environmental Sciences University of California .

900 Veteran Avenue Los Angeles, California 90024 June 1983 i

UCLA Publication 16 12-1417 4F page; l F9104020139 91031e l F;DR ADOCK 05000182 POR

_ _ _ _ _ . . . - -. --.- -- - - - - - ---- '~~"' ' ~~ ~~ ~ ~ ~ ~~" ~ ~ ~ ~ ~ ~~

. c 2

TABLE OF CONTENTS Page I n t r o d u c ti o n . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 Types of labeled water . . ........................ 3 Isotope suppliers .... ....................... 6 E s t i m at i ng co s t s . . . . . . . . . . . . . . . . . . . . . . . . . . . . . fl P re p ari ng i nj ec ti on sol u ti on . . . . . . . . . . . . . . . . . . . . . . 10 f i e l d p ro c e d u r e s . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12 Supplies and equipnent ......................12 I nj ec t i ng a n im al s . . . . . . . . . . . . . . . . . . . . . . . . . 13 Ob ta i ni n g s am pl es . . . . . . . . . . . . . . . . . . . . . . . . . 14 Recapture interval ........................17, Laboratory procedures .........................19 i Injection syringe calibration and standard preparation. . . . . . . 19 Distilling samples . . . . . . . . . . . . . . . . . . . . . . . . 21 E qu i pm en t a n d s up pl i es . . . . . . . . . . . . . . .' . . . . . 21 P rocedu re . . . . . . . . . . . . . . . . . . . . . . . . . . 23 P i pet t i ng s ampl es . . . . . . . . . . . . . . . . . . . . . . . . . 26 Equi pnent ..........................26 Procedure . . . . . . . . . . . . . . . . . . . . . . . . . . 27 T r i t i um a n a l y s i s . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30 0 xy g e n - 18 a n a l y s i s . . . . . . . . . . . . . . . . . . . . . . . . . . . 31 C a l c ul a ti o ns . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32 Con ver s i on f ac to rs . . . . . . . . . . . . . . . . . . . . . . . . . . . 34 Errors.................................36 Permits ................................39 Ac k n owl e d gem e n t s . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40 Literature cited . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42 A pp e nd i c es ( c al cul a ti on f o rm s ) . . . . . . . . . . . . . . . . . . . . . . 44

-3 ,

INTRODUCTION The doubly labeled water method for measuring C03 production and water flux rates is being used with increasing frequency in studies of energy and material balance in animals living in their natural habitats. Nathan lifson and his colleagues originated and developed this method beginning more than 10 years ago, and they sumarized their efforts in 1966 (Lifson and McClintock 1966). However, few investigators used this method during the 1960s, apparently because of the high costs and technical dif ficulties involved.

More recently, isotope costs have declined as much as 901 (for 100 ),

technical aspects of sample analysis have been streamlined and improved (Wood et al.1975), and potential errors have been evaluated (Nagy 1980, Nagy and Costa 1980). ,

Our research group has been using a form of this method routinely for over eight years. We have received numerous requests for infonnation and assistance regarding this method. This paper is intended to help answer the more comonly-asked questions about our materials and prodedures, and to encourage investigators to add this research tool to their repertoire of ca pa bili ti cs .

TYPES OF LABELED WATER Doubly labeled water is prepared for injection by mixing sppropriate amounts of water containing a hydrogen isotope together with water containing an oxygen isotope. Once this mixture is diluted in an animal's body water, the concentrations of both isotopes are generally so low that few individual water molecules actually contain both isotopes (despite the implication given by the name of this method),

.. 4 Two isotopes of hydmgen, deuterium and tritium, can be used. Deuterium is a stable isotope (not radioactive, so it doesn't decay away on the shelf).

Advantages of deuterium are 1) its use does not require radioisotope pennits (see PERMITS section below). 2) it is relatively inexpensive (about $0.35/g of 2

99% H2 O when purchased in bult), and 3) it can be used in conjunction with tritiated water (3HHO) in studies, where water transfers between labeled organisms (e.g., mother end suckling young mamals) are of interest. The major disadvantage of deuterium is the relative dif ficulty and expense of measuring it accurately. At concentrations between 0.1 to 2 atom %, deuterium can be analyzed accurately by conventional mass spectrometry (Mullen 1973) and by inf rared absorbance (Zweens et al.1980). At lower concentrations (0.017 to C.05 atom %, equivalent to 0.002 to 0.04 atom i excess above the nr.tural

~

backgrounc of about 0.015 atom %) reliabic measurements can be done with an isotope ratio mass spectrometer (Schoell er et al .1980). Both mass spectro.

metric mettods require conversion of H in H 2 O (distilled from the biological sample) to H2 gas in a good vacuum line. The infrared absorbance method also requires distilled water. We are experimenting with a Bechnan Acculab 2 1R spectrophotometer equipped with a small volume (25 microliter) sample chamber, and have found that accurate measurements become difficult at concentrations below about 0.1 atom %. Thus, animals would require relatively large doses of deuterated water with this method, and this may cause problems due to 1) disruption of water balance status by the large volume of water injected, and 2) possible toxic effects of heavy water (Samis et al.1974 Sutton et al .1974).

Tritium is radioactive (radiological half life = 12.3 years). It is a sof t beta-emitter, having a maximum radiation distance of less than 1 m, so it is relatively safe to use compared with other radioisotopes. A major

l oS- ,.

advantage of tritium is the ease and accuracy with which it can be measured using a liquid scintillation counter. It is also relatively inexpensive, and very small injection volumes can be used. A disadvantage is the necessity to obtain permits for its use, both in the field and in the laboratory (see PERMITS t ection below).

Tw isotopes of osygen,18 0 andI ,70 , are available in the form of enriched water. Both are stable isotopes, and both are relatively expensive.

Oxygen 17 may be useful in doubly labeled water studies, but this has not been investigated. Oxygen 18 can be used at two levels of enrichment in animal studies, depending on the range and accuracy nf the measurement technique to be used. The hi,aer (and thus more expencive) enrichments (0.25 to 2.0 atom 1, equivalent to 0.05 to 1.8 atom i excess above naturai bactground of about 0.20 atom 1) should be used in conjunctinn with analyses done with

• conventional mass ;pectrometers (Boyer et al.1961) or with the. proton I

activation method (Wood et al.1975). Much lower enrichments (0.001 to 0.02 atom i excess) can be used wher samples are analyzed by isotope ratio mass spectrometry (IRMS). This method has been validated in humans (Schoeller and With this technique, costs for I8 0 cre much lower but van Santen 1981).

IB O costs for analysis of are high (ca. $40/ sample vs. about $15/ sample for proton activation ana'ysis). There are everal real and potential drawbacks to the !RMS technique: 1) high cost of analysis, 2) relatively large sample 18 voluc.1es required (several c1), 3) dif ferences in natural 0 levels (background) between animals and within a single animal through time may be large relative to the levels of enrichment used. We are examining these questions in an attempt to improve the acceptability and utility of the low-level enrichment method. At this moment, the high-level method is best for small animals, where costs for isotopes are low but many samples are

I obtained in a single study. For large animals, high isotope enrictnents are usually too expensive, and the low level method may be the only choice.

Moreover, costs for sample analysis by isotope ratio mass spectrcnetry are

! comparatively 1cw because fwer animals are usually injected in large animal studies, so there are fewer samples that require analysis.

The remainder of this paper describes the use of water labeled with tritium and with oxygen-18, the latter used at the higher enrictnent.

Analysis of tritium is done by liquid scintillation counting and oxygen-18 is analyzed by proton activation of 18 0 to I8 F with subsequent counting of gam a- emi t t i ng I8 F in a gama counter.

This method is reliable and has been validated in several species of animals (sumarized by Nagy 1980).

ISOTOPE SUPPLIERS Some sources of water labeled with IOO , 2H and3 H are listed in Table 1. This list was generated from catalogues that we had on hand at the moment, and it is not at all complete. Some catalogues ar'e several years old, so the prices listed may not be current. The price of I8 0 labeled water has been declining (1 g of 97 atom % H 0 cost more than $500 ten years 2

ago), but prices of tritiated water have been increasing. Some companies sell l 18 H

2 0 at lower enrichments, such as 10 and 20 atom 1, and the price per g isotope may be lower. However, we do not use these enrichments because of possible disruptive effects of the large volumes of water that must be given to animals to enrich their body water to sufficient levels. Some companies I

include shipping and handling costs (which are substantial for small orders) into the list price. They may off er lower prices for large orders.

,- . - - - - -. w- --.c - w. . . . -1,- - , , , -- r ,.,- , --yv 9 m-s- y e

l, 7 .

Tabl e 1. Some sources of H2 180 , 2 H O and 3HHO.

2 Prices i

parentheses were calculated f rom data given by supplier.given Prices may not inbe current, l

i

Suppit er Enri ctrnent Cost

OXYGEN-18 Per 9 isotope Per g water l (DOE or other federal grant)

Stable Isotope Sales, Monsanto Research

• Corp. P.O. Box 32 95+1 $52.50 ($44.63)

Miamisburg, Ohio 45342 (comercial)

(513) 865-3501, Attn: Al Ruwe 95+1- 569.50 (559.08) i (for 2 to 9 grams)

YEDA Stable isotopes, P.O. Box 97% (574.20) 565.00*

95, Rehovot, Israel, phone _(for__10___to99gramsL 054-70617, Attn. Michael Epstein 97% (563.0) $5F0 F (for 100 499 grams) 97% IT52.51) 54P

• (Plus costs of shipping, insurance, handling, 5% import duty)

PROCHEM,19 0x Bow Lane, Sumit NJ, 07901, phone, 99% ($81.37) $72.50*

4 (201) 273-0440, Attn: John Kilby -

• (Lower prices for large orders)

Isotec, Inc .,1029 Senate Dr. ,

Centerville, OH, 45459, phone 90-99% $123.50 ($100.04-110.04) )

(513) 435-4669, Attn: Vince Avena ,

MSO lsotopes, Merck Chemical Division, (For1_g)

P.O. Box 2951, Terminal Annex, 95-99% ( $175.44-168.35) $150.00 CA. 90051 Phone (For 10 g) *

(800) 423 4977 ( $1F2.05-145. 90) $130.00 DEUTERIUM MSO Isotopes., Merck Chemical Division P.O. Box 295i, Terminal Annex, CA 90051 99.8% $43/100g Phone, (800) 423-4977 4 Stohler Isetope Chemical, 49 Jones Pd, d

MA 02154, 99.8% $40/100g Waltham[617)

Dhone. 891-1827 KOR isote. pes, 56 Rogers St . ,

99.75% $38/100g Camt Phone (617 dge,)MA 02142 661-8220

-8 Table 1 (Continued)

Suppli er E nri chment Cost DEUTE RIUM Bio-Rad laboratories 32nd & Grif fith Ave.

Richmond, CA 94804 99.8% $40/100g Phone (415) 234 4130 ,

ICN Chemical & Radioisotope Division 2727 Campus Drive, Irvine, CA 92715 99,8% $40/100g Phone (714) 833-2500 TRITlUM Ncw England Nuclear, 549 Albany St. I mci /g H 2O $26/g H 2O Boston, MA 02118 25 mci /g H 2O $26/g H 0 Phone (617) 482-9595 100 mci /g H 2O $39/g H 0 1000 mci /g H 2O $105/9 20 Research Prods. Intl. Corp. ,

410 N. Business Center Dr. 10-1000 mci /g H 2O $150/g H 2O Mount Prospect, IL 60056 Phone (312) 635 7330

• ICN Chem, & Radioisotope Division 2727 Campus Drive, 100 mci /g H 2O $35/g H 2O Irvine, CA 92715 Phone (714) 833-2500 ,

knersham Corp., 2636 S. Clearbrook Arlington Heights, IL 60005 5 mci /g H 2O $35/g H 2O Phene (800) 323-9750 1

=. - , - - , , _ _ __ . . . _ . . . - , . . . _ - . - - _ . -

4 9 ..

ESTIMATING COSTS

\

The following is an example of calculating the major costs for doubly labeled water in a typical study. Asstne that mean body mass of the species is 100 g. Assume that we wish to compare field metabolic rates of adult males with those of adult females. Minimum acceptable sample size is 8 for each group (for statistical reasons) so 16 recaptures are needed. If recapture

> success is about 40%, then 20 males and 20 females should be injected. 40 animals x 100 g/ animal = 4 kg of animal to be labeled. For H 2 0, we use 18 0 per kg, which yields an initial a dose of 3 ml of 90-99 atom 1 H2 180 level in body water of about 0.6 atom 1 (0.4 atom % excess). Thus, 4 kg-8 x 3 ml/kg = 12 ml 90-99 atom % H2 0 are needed. A ml of 95 atom 5 IO 0 18 0 isotope (due to the higher density of H contains roughly I g of 2

H I8 0 compared to H O). 12 g I8 0 as 90-99 atom % H 2

I80 , at '

2 2 3

about $50/g, will cost $600. We inject about 1 mci H/kg of animal, so 4 mci 3H are needed, at a cost of about $30 (Table 1). We expect to obtain 16 recapture samples, so the 16 initial samples for these animals also require isotope analysis (initial samples from animals that are not recaptured are

{ discarded). Total number of samples will be about 36 (including background samples and standards). Tritium analysis by liquid scintillation counting costs about $0.75/ sample for vials and scintillation chemicals, so 36 x $0.75 18

= $27. The wheel used for proton activation analysis of 0 (see below) will hold up to 37 samples (in addition to the necessary comparators and standards). Thus, all samples in this study can be analyzed on a single wheel, at a cost of $550. Miscellaneous supplies (blood containers, syringes, tape markers, torches, capillary tubes, etc.) should cost about $80. To sumarize:

i

4 .

]

. 10 18 H

2 0 $ 600 >

I 3 HHO 30 Tritium analysis 27 18 0 analysis 550 Supplies 80

. $1287 18 For animals smaller than 100 g, isotope costs are lower, and the 0

analysis becomes the major expense, in larger an%alsi the isotope costs become the major expense. For example, a single 50 kg dace would require 50 kg x 3 g H2180/kg x $50/g H2 0 = $7,500, and a study involving 18 eight deer (assuming high recapture probability) would cost $60,000 for 0 4

alone.

PREPARING INJECTION $01.UTION First, we calculate total body mass of all animals to be injected (see above), multiply this by 3 ml/kg, and transfer that volume of 90-99 atom 5 18 H

2 0 to a dry, sterile serum-stoppered glass vial. Then, we calculate total tritium needed on the basis of tritium counting ' statistics. In order to attain 1% counting error (2 standard deviations) _in a liquid scintillation counter, about 50,000 counts must be accumulated. To achieve this in a reasonable amount of counting time, the scintillation vial containing the initial sample from the animal should contain at least 12,000 counts per minute (CPM). We usually count -10 microliters of-sample per vial, so'that-10 pl of body water should contain 12,000 CPM. Thus 1 kg of animal, containing about 700 ml of body water (for reptiles and mamals- for birds, we use a

typical value of 630 ml TBW/kg), should have 12,000 CPM /10 ul x 1000 pl/ml x 700 ml = 8.4 x 108 CPM af ter isotope injection. One millicurie (mci) of tritium is equivalent to 8.8 x 10 8CPM with a typical counting efficiency of 401, Accordingly, the tritium dose should be about 1 mC1/kg body mass. The appropriate volume of tritiated water stock is then calculated and added to the H2 8 0 in the serum stoppered vial. A tape label is placed on the vial and aarked appmpriately. Nacl may be added to the injection solution, if desired, in the amount of 0.9 g NaC1/100 ml H2 0. We generally do not add NaC1, because 1) injection volumes are small relative to body mass (0.3% of mass), 2) it is dif ficult to add Nacl without losing expensive isotopic water or causing excessive dilution of the injection solution, and 3) adding Nacl appears to f acilitate microbial grcwth in the injection solution, a consideration if the solution will not be used right away. .

Tritiated water is available in a variety of enrichments (Table 1). We use 25 mci /m1, because the volumes we nonnally transfer are large enough to be measured easily with a 1 ml disposable tuberculin syringe, but they are still small enough that the H2 0 is not diluted much by addition of the H2 0.

Af ter shaking the vial to mix the injection solution thoroughly, we withdraw a small amount to check the tritium level. 5 pl of injection solution are pipetted into 1.0 ml distilled water and mixed. This is approxin.ately the same dilution as will occur in the animals (3.0 ml/kg injected into 700 ml TBW/kg). 10 91 of this are transferred to 10 mi scintillation fluid in a counting vial (see details below) and counted to I

insure that there are about 12,000 CPM in the vial. We usually retain ca. 30 pl of the injection solution, flame-sealed in a capillary tube, for later use i in preparing standards.

l 4

When it may be difficult to recapture injected animals before their i isotope levels become too low (af ter about 3 half lives -see below). either because CO2 and water fluxes are very high (e.g., hunningbirds) or because animals are only intennittently capturable (e.g., sea birds having long foraging trips), we have increased the isotope doses given to each animal.

However, this quickly becomes expensive. The isotope loss curve is exponential, so doubling the dose provides only one more half-life, and the dose must be quadrupled to obtain two additional half lives. This approach also raises the risk of inducing toxic effects caused by heavy water (Samis et al .1974, Sutton et al .1974).

For very small animals (0.5 to 1 g), removing 35 pl of body fluid for isotope analyses may be harmful, thereby defeating the goal of obtaining ,

measurements on healthy, undisturbed animals, in this situation, we have been able to do isotope measurements with samples as small as 10 pl. The results were more variable, but still reasonable. We used one microliter for tritium analysis, and increased the amount of tritium injected as.nec6ssary.

Oxygen-18 analyses were done using 4 91 Microcap pipettes. Each pipette contained 3 ul for a total of 9 p1 in the three pipettes we prepared for each s ampl e.

FIELO PROCEDURE >

Supplies and equipment We use a tackle box containing: The vial of injection solution, a syringe and needle for making injections, a container of heparinized microhematocrit tubes for blood samples, Critocaps for temporarily sealing blood tubes (if necessary), a small butane torch (e.g., Microflame " CUB "

l

l 12 Microflame Co., 3724 Oregon Avenue S., Minneapolis, M155426, phone (612) 935 3777) for flame scaling blood tubn. tape and waterproof marker for labeling blood tubes, paper pad and pencil for recording data, pesola spring scale or an appropriate balance for weighing animals, paint or tags or dye for r,arking injected animals, plastic or glass vials for storing and protecting sealed blood samples, tissues for wiping up blood, a plastic bag to contain any radiocctive wastes, small scissors for cutting tape markers and clearing hair from veins, a needle f or pricking veins, and a butane cigarette lighter for lighting the torch. For larger animals, we use heparinirtd 4 ml Yacutainers for taking blood samples from veins. A minimum maximum thennometer is placed in the study area in the shade during the measunnent period. Surveyor's tape may be useful for marking capture locations or coordinates in the study area. Animals are captured by hand, with nooses, in baited u os, or in mist nets as appropriate, injecting animals We prefer to use the same syringe and the same volump of doubly labeled water for all injected animals, if possible. When there are large differences in body mass between animals, two, perhaps three dif f erent injection volumes (all about 3 ml/kg) will be used. This procedure greatly simplifies determinations of volumes injected. These are measured later in the  !

laboratory by calibrating the injection syringe (s).

We use different routes of injection, depending on the situation.

Intetvenous injections equilibrate ir. body water the f astest, so captured animals can be released sooner. However, a sufficiently large vein must be j accessible to do this accurately. A? so, the injection syringe of ten becomes l

l contaminated with blew, in tt 's ev(nt, we use a new syringe on each animal, j because rinsing c9 drying i contaminated syringe takes too long, and rinsing i

1 1

I

m

-14 but not drying a syringe causes contamination of the injection solution by rinse water, intramuscular and intraperitoneal injections equilibrate at about the same rate, but slower than I.V. injections, and subcutaneous injections are the slowest to equilibrate. We use intraperitoneal injections in most situations. These are simple and accurate. In birds, the presence of abdominal air sacs complicates intraperitoneal injections, so we inject deep into a pectoral muscle in these animals.

We of ten use the dilution space of injected isotopes (usually IOO ) as an estimate of total body water volume (required in the CO2 and H2 O flux cal cul a ti ons). It is important that injections are done carefully and reproduc ibly. We carefully align the syringe plunger to the desired volume when filling it, and use the same pressure on the plunger at the end of .

injection (important when using disposable syringes with rubber plungers).

Then we pull the skin to one side while slowly withdrawing the needle to cut off the needle's channel in the animal and reduce leakage, if any problem or error occurs during injection, we record it, and that TBW,value may then be discarded on that basis. The same person that does the injections should also do the syringe calibrations. We have found considerable variation among ourselves in this regard.

O_btaining samples Af ter injection, we hold the animats until the isotopes are thoroughly mixed in the body water betore taking a sample. No food or water is available during this time. For small animals (ca. 300 g or less) given 1.V.,1.p. or 1.M. injections, one hour is usually suf ficient. More than I h is needed in reptiles if they are cold, if we have doubts about the correct equilibration period, we measure it in the laboratory beforehand by taking serial.15 min 3

samples in an animal given HHO only. Large animals (above 1 kg) require 1

1

l 2-4 h for equilibration, and some ruminants may require 10 h (see Nagy and Costa 1980). During the equilibration period, the animals can be weighed, sexed, measured and marked for later identification.

Fluid samples can be blood, saliva (Schoeller et al.1982), tears,1pph, urine or feces. Urine and feces are not recomended as initial samples because of a probable lag in complete' equilibration of injected isotopes with water in these substances. Urine or feces have proved adettuate for recapture samples, except for animals having short food passage times (i.e., minutes).

f)

We usually take blood samples, as these are easy to obtain and process. In reptiles and small mamals, we use the suborbital sinus behind an eye. A non-fire-polished (sharp), heparinized henatocrit tube is carefully inserted l between the eyelid and eyeball, and gently pushed, with a twisting motion, ,

against the orbit. Upon slight withdrawal of the tube, it will usually begin filling with blood, in larger mamals, a leg or ear vein is pricked with a needle, and the dmp or two of blood is taken up in heparinized capillary tubes, in birds, the brachial vein in the wing or various veins in the leg and foot can be used. Bleeding is stopped by a combination of direct pressure, upstream pressure to occlude the vein, and wiping away excess blood and blowing on the wound to hasten clotting. Gelfoam (a comercially-available, topical clotting agent) has been useful in some situations. Be sure to take scce blood f rom an uninjected animal for 18 measurement of background 0 level .

The minimum volume of fluid needed for routine isotope analyses is 35 pl . A microhematocrit tube holds about 70 pl when completely full. We usually take about 45 pl in each of two tubes (one for analyses, one for ba kup). This leaves room at each end of the tubes to flame-seal both ends (with the small torch) without burning the blood. The two tubes are checked

16-to make sure the seals are complete and they are taped together, marked for l identifiration, and placed in a larger container for protection. The time and d6;r are recorded. Samples are refrigerated until they are distilled in the i h ratory. We prr fer not to freeze samples, because some tubes have sh6ttered and were lost. Samples should not be allowed to rot, because such samples do not yield clean water upon' distillation and contaminated water plays havoc with the I8 0 activation analysis.

Flame seals are important. We have had many problems with Critocaps, clay seals, wax seals and parafilm. These often fail, especially during changes in barometric pressure (such as in airplaines). Failed seals open the possibilities of 1) loss of sample through evaporation, 2) enrichment of sample via isotopic f ractionation during evaporation, and 3) dilution of ,

sample by instillation of water vapor from other samples in the same container l

or ambi ent water vapor.

When the species being studied is so small (ca. 2 g or less) that taking l

an initial fluid sample would be sufficiently hamful to offect nonnal behavior a f ter release, we have deleted this step and used estimates rather than measured values for initital isotope levels. In this event, study animals are released immediately af ter injection, weighing and marking.

Initial isotope levels are estimated from data on additional animals captured at the time of injection. These animals are processed as above, except that initial fluid samples are taken for isotope analyses, and the animals are then l killed and dried at 60-70 00 to determine water content. These animals should be selected to span a wide range of body sizes. Body water volume (ml) is plotted ag: inst live body mass (g), and the least squares regression line is calculated. Then, this line is used to predict the body water volume of the released animals from their live body masses. The equation (ml body water

-17 voltne) X Iisotope specific activity) = constant is used to calculate constants for citium and oxyaen-18 fr m the isotope results for each of the animals that were sampled just af ter injection. The means of these constants are then used in tte same equation to predict isotope specific activities for animals that were released, f rom their predicted body water contents.

Recapture interval The most reliable results are obtained when animals are recaptured 18 between one and two half-lives of the 0 isotope, if sampled earlier than one half-life, insuf ficient isotope turnover has occurred, and small errors in isotope measurements cause large errors in the results, if sampled af ter 18 three half-lives, the 0 level is too close to background to yield reliable 18 results. We distrust any 0 values that are below about 0.23 atom % (0.03 .

atom % excess) with our analysis methods. Tritium is accurate for more than 5 half-lives, so calculated water fluxes are reliable for a longer period than are calculated CO2 production rates.

Biological half-life of I8 0 in animals varies with body mass and '

differs between major taxa. Half-lives can be estimated from the equations in Tabl e 2. These equations were derived from results of doubly labeled water studies done in the field. Each equation has limitations (see footnotes in Table 2), so they should be used only as a guide. In general, animals living in moist habitats should have shorter half-lifes than predicted, and vice versa for desert ani...als. Similar adjustments stould be made for succulence of diet, extent of drinking f ree water and stage of breeding cycle. Most I8 0 isotope in labeled animals is lost in the form of water,

(>80%) of the so water flux (not metabolic rate) is the major determinant of 180 hal f-li f e.

Upon recapture, we take another blood sample and record time, date and body mass. The animal can then be released for a second measurement period

18-

. Equations for estimating biological half-life (T1 in days) of 180 Table]0 as H2 in terrestrial vertebrates in the field, from body /2 mass (ing).

REPiltESI 0

/2 = 9.55 9 106 BIROS2 0

TU2 = 0.152 9 37 MARSuplAL MAMMALS 3 EUTHERIAN MAMMALS 4

. 0.147 '

t 1

Reptiles living in arid and subarid habitats, during activity season (spring and sumer) only. Tmpical reptiles have shorter half-lives, and longer half-lives occur in Xantusiid lizards and in arid reptiles during inactive seasons.

2 Based on relatively few data, but representing a wide variety of habitats and di ets . Desert species have longer half-lives than predicted from this equation.

3 All smaller marsupials in this data set were Dasyurids, which have high water flux es . Thus this equation may understimate half-life for small, non-dasyarid marsupials.

4 Desert species have longer half-lives than predicted from this equation, i

, , , - - -- , - - - - - , - , , ,----y _ , - . -

l i

(the recapture sample serving as the final sample of the first period and the

! initial sample of the second period), if the animal is thought to contain j suf ficient isotope.

4 LABORATORY PROCEDJRES k

injection syringe calibration and standard preparation Measurements of injection volumes and standards can be done either in units of volume (m) or 91) or in units of mass (g, mg or pg). Care must be taken to use the same units throughout all calculations. We use volume rather than mass, because we have found this to be aster and more accurate overall.

Our procedure follows. .

The injection syringe (s) is flushed with distilled water, filled to the appropriate mark with distilled water, the outside is- dried, and the syringe is weighed to 0.1 mg. Then, the syringe is emptied using the same plunger pressure as in the field and weighed again. Injection' volume = difference in mass. The same person that did the injections in the field should calibrate the syringe. The calibration should be repeated at least three times, watching for consistency in results. -We use the mean value in subsequent cal cul a tions. We assume 1.0 mi distilled water weighs 1.0 g. Isotopically labeled water has a- higher density than distilled water, and should not be -

used to calibrate the syringe when using the. procedure based on volume units.

Our isotope counters are not capable of measuring the high levels of3g 18 and 0 in typical injection solutions, so the injection solution must be diluted before measurement. In theory, the best way to do this is to use the-injection syringe to inject the same volume of injection solution used in the 1

field into a- beaker containing a measured volume of distilled water l

i 20-representative of the total body water volume of a typical animal in the fi el d. (It this procedure is used. the injection syringe need not be calibrated). This procedure works well when injection volumes are small (so using a full injection amount for standard preparation is not prohibitively expensive), and when only one syringe and one injection volume was used in the fleid (multiple syringes or injection' volumes require multiple dilutions).

However, in practice, we have found that using Microcaps to dilute injection solutions is the most reliable procedure.

Using a previously calibrated Pipetman P1000 autopipet, we transfer 1.00 ml distilled water to a clean (less vial. Then, 5 microliters of injection solution are added using a 5 91 Microcap (Gold t.abel), and the vial is stoppered and shaken to mix it completely. This yields plenty of standard ,

water for duplicate measurements of tritium content (if desired) and triplicate measurments for oxygen-18 (required if body water volumes will be estimated fra oxygen-18 dilution space), and conserves expensive injection salution. ,

Standard dilutions of the injection solution can be done on the basis of mass rather than volume. This requires that the amount of injection solution added to the distilled water be weighed accurately, and this is difficult when using small volumes and Microcaps because.of the small masses involved and the continual loss of injection solution by evaporatiori during the process.

However, this disadvantage is offset somewhat by the advantage of potentially more accurate measurements of volumes injected in the field. These can be done by accurately weighing the injection syringe before and af ter injection.

Syringe calibration is unnecessary with this- procedure. Combinations of the voltne method and the mass method can be used, provided that the density or specific gravity of each injection solution is carefully measured in order to p ermi t i, %ns between units of mass and volume.

1

- . - - - . . . . ~ _ . . . . - . . . , . . - - . , . . - . _ . , - . . _ . - . - . . . - . _ . . _ _ _ _ , , _ . , _ _ _ . , . _ .

-21 Distilling samples Equp,ent and supplies

1. Slide warmer with cover, to hold temperature at about 500 0 (hotter will cause volatilization of organic compounds and a dirty distillation). Cut slots in side of cover so that the tips of Pasteur pipettes laying on top of the wanner can extend out through the slots (see Fig.1).

2. Small vacuum bcttle containing liquid nitrogen - optional (see below).

3. pasteur (transfer) pipettes, 9 inch size. These should be stored in a sealed container along with Orierite or silica gel to insure that they are absolutely dry before use. The pipettes made by various companies have dif ferent properties- some cre.k easily while others have a thin section that .

is too narrow to contain much distillate. Kimble provides a good pipette.

4 Good vacuum pump with hose equipped with a stopcock at the distal end and a rubber stopper with a hole that will provide a tight fit around the thin end of a Pasteur pipette. The stopper and glass tube provided with each vial of Microcaps works well here (Fig.1).

5. Welder's torch with two tubes - one for bottled or natural ;as, the other for bottled oxygen. This is needed because bunsen burners don't produce a flame that is hot enough or narrow enough for precise and rapid melting of pipettet,

6. Beaker for tot glass fragments.

7 Asbestos pad for hot pipettes.

8. Marking pen, sharpening stone or file or glass cutter, scissors, needle-nosed pliers.

22 1

Pasteur pipette, Slide warmer evacuated, f?ame set at 5000 sealed with Beaker of distilled simp 1'e inside water with tissue Slide warner Sample Paper wick .

cover } tube)

{ .m- .~ _ . I .

. i i

i

- - - m m- - - 2haC /

(,,7 ,

[

( -

N . ,

N t

Stopper and tube supplied witl. microcaps T *

~

Dispossble tubercu111 7 syringe barrel Tubing - < /

./ t. Vacuum line i [

T ,_ _gQ' ~ f

' Q.

~ n, C -

r to good vacuum pump g 6\d ' . We 's ,} ,],.

Disposable 3 way stopcock l

Fig. 1 Slide warrer distillation setup, and vacuum line plumbing.

.o -.

Procedure

1. Set samples out to warm to room temperature to avoid problems of ambient water condensation on the outside of the tubes. Be sure dry Pasteur pipettes are ready. Turn on torch (oxygen on last, off first) to yield gentle, but hot flame.

2. Remove a dry Pasteur pipette' from the container and mark the sample identification on it.

3. Remove tape marker from hematocrit tube containing sample. Save marker for later use.

4 Scratch tube near one end and break open. A small sbarpening stone works well for scratching. Open end with large air pocket first, to minimize chances of sample blowing out. We work over a sheet of Parafilm, so that if some sample is dmpped, it can be retrieved into a . lean hematocrit tube kept handy just in case. Don't saw on the tube -only a small scratch is necessary. Scratch and break off other end. Scratch in middle and break this section in half. The long middle section is broken in half because a long tube full of sample can spurt during evacuation and contaminate Pasteur pipette. It is imrortant to keep the sample horizontal throughout so the fluid doesn't run out of the tubes. If a sample is clotted, scrape the ends of the two middle sections together before separating them to ' cut the clot off. If sample stinks, distillate will probably be contaminated and it should be distilled a second time.

5. Place all four tube pieces carefully'through the large end of the Pasteur pipette, past the ~ constriction, into the middle segment. The two end pieces are included because they often contain fluid-as well. Try not to.get any sample on the inside of the pipette, as this may burn during flame sealing; and contaminate the distillate with smoke.

l

F

6. Holding the Pasteur pipette horizontal, irove it over the flame so that the flame heats the large end midway between the opening and the constriction. Roll the pipette so that the tube is heated evenly. When the glass softens and the end begins to droop, grab it with a pair of needle-nosed pliers and slowly pull it out of the flame. The rest of the pipette will thereby be sealed. Keep the rest of 'the pipette in the flame briefly, rotating, until a small button of glass forms on the end. This is a critical step. Do not expose the large open end of the pipette to the flame. The flame is producing " metabolic" (oxidation) water at a high rate, and if this water gets inside the pipette, the sample will be contaminated by unlabeled water. Fonnation of the glass button af ter sealing is important because it helps reduce stress in the glass. Stressed glass often cracks during ,

temperature changes, and cracks cause loss of vacuum in this procedure. it is important to keep the pipette horizontal- during this procedure.

7. Gently lay the pipette on the asbestos pad to cool.

8. Turn on the vacuum pump with the stopcock closed,so that the hose is evacuated. When the pipe'..e is cool, inspect the flame seal for cracks. if a crack is present, scratch and break open the pipette, remove the sample and repeat the above steps, with a new Pasteur pipette.

9. At this point, two things can be done depending on sample volume. - For samples larger than about 10 pl, use step 9A; smaller samples - use step 98.

9A. Tilt the pipette and tap it so that the tubes containing the sample slide to the large end. Insert the small end through the stopper on the vacuum hose. Open the stopcock slowly, and then within 2 to 4 seconds, flame seal the small part of the pipette near the stopper, while still evacuating, Place pipette on asbestos pad to cool. Close stopcock dDd removi h7! gl? d stub f rom stopper with pliers. You should end up i ~

l 25 .

with a sealed, evacuated pipette with the sample in the.large end. If l the small end implodes upon flame sealing, the flame is too hot. The sample may bubble and spurt during evacuation. Check to see if any sample splashed into narrow end. If so, distillate will be contaminated and must be redistilled. The rapid sealing after opening the vacuum is necessary to prevent loss of muc'h sample volume. Sample water evaporates rapidly in a vacuum, end vapor will be sucked out by the vacuum pump.

The pipettes are evacuated in order to speed up the distillation process.

98. For snail samples, volume is critical, and the sample is frozen before evacntion to help immobilize it. Tilt the pipette so the sample tubes f all to the large end. Shake out as much fluid from the tubes as pos si bl e. Full tubes often shatter upon freezing, and flying debris can, contaminate the narrow end of the pipette. Insert the narrow end into the stopper in the vacuum line to isolate the pipette air from ambient air. Immerr ' the large end in liquid nitrogen until the LN2 stops bubbling. Open the stopcock to evacuate the pipette and flame seal the small end near the stoper. Place . pipette on asbestos pad to come to room -

temperature. Check each pipette to be sure narrow end hasn't become-contamir.ated by splashing sample,

10. Lay sealed, evacuated pipettes on top of cool:(room temperature) slide warmer so that narrow portions extend about I cm out tnrough slot in -

~

side of cover (Fig.1). When all pipettes are cn slide wanner, turn it on so that tube heating is gradual (reduces temperature-induced cracking). Sampl e water will distill from hot (large) end to- cool (narrow end). Distillation should be complete in 6 hr (small sample volume) to 18 hr -(70 pl sample). For larger samples, the narrow end of pipettes should be periodically pulled further out of the slots to minimize bubble formation in the narrow end.

7FT-- ---' --

s _,-,_,,%,nw ,. . .s w y, y.._, ,, .4._m.., , ,.,,.,.y ,-,7,,y,,-, ,_,,,,,ww.m,._,.3-

i

. Pipettes that distill slowly probably are cracked. Allow 3-4 days for _ complete distillation of _ these. We have usually been unable to demonstrate significant chan5es in isotope specific activity in samples due to pipette cracking upon comparison with uncracked replicates. However, differences do occur on occasion. Optional : distillation may be speeded up somewhat by draping tissue paper strips over protru{i,ng narrow ends, and wetting th m by capillarity from a pan of water under them. This will cool the narrow ends a bit more.

11. When distillation is complete, as indicated by the sample residue in the large end appearing dry, flame seal the narrow portion of the pipette behind the column of distilled water. You should now have a sealed tube of distilled sample water. Place the appropriate marker tape on this tube and retain for isotope analyses.

Pipetting samples Equipnent and supplies

1. Torch, sharpening stone, marking pen as above.

l 2.

l L.iquid scintillation vials and fluid.-- We use low-potassium glass l vials with nite, foil-lined caps. The scintillation fluid we use is ~

inexpensive and eff ective.

l Pour.13.61 g PPO (2,5-diphenyloxazole) intr' a new, 3.78 liter bottle of scintillation-grade toluene and shake. Mix four parts PPO-toluene solution with one part Triton X-100. Transfer .10 ml of this to scintillation vial, using Brinkman Dispessette or Repipette.

Two mouth pipettors. We make ours by cutting off and discarding _

narrc+ end of Pasteur pipette, attaching narrow rubber tube to tapered end of Posteur pipette, and Gelman SAFE mouthpiece to end of rubber tube. (SAFC mouthpiece contains a micmpore membrane barrier that prevents liquid isotope ingestion. Houth pipetting affords much more control and precision than does

)

hand pipetting small volumes). Insert rubber stopper supplied with each vial i

- }

of Microcaps into large end of Pasteur pipette. l

4. Containers (plastic tubes-or vials, small envelopes, etc.) for )

10 0 analysis tubes, one for each sample and standard.

holding

5. Microcaps, GOLD LABEL. One, of appropriate volume (usually 5 or 10 pl), for tritium analyses. One or two vials of 100, 20 p1 GOLD LABEL Microcaps for oxygen-18 analyses. A full wheel of microcaps for activation analysis contains 129 Microcaps (two vials). We have found that 20 p1 Microcaps f rom different production runs at the Drummond factory can yield variable results in our analyses. Either include 6 more 0.201 atom !

comparators for second vial of Microcaps in the wheel', or insure that all 129 Microcaps on a wheel are from same production run. We order directly from .

J. A. Walker, Sales Manager. Drummond Scientific Co., 500 Parkway, Bromall, Pennsylvania,19008, phone (215) 353-0200, specifying "Please make sure that all vials sent are f rom the same production lot in order to insure coastancy of diameter between vials of 100." .

6. Sheet of Parafilm for work station, small beaker of distilled water for rinsing HTO Microcap, tissues for rinsing.

' Procedure

1. Set up two scintillation vials with 10 mi scintillation fluid for each sample and standard to be measured- for tritium content. Mark sample ID on caps.

2. Mark sample 10 on containers used to hold Microcaps for oxygen-18 analyses.

3. On a sheet of white paper, line up 20 p1 GOLD LABEL microcaps from one or two vials of microcaps. (Eighteen Microcaps per wheel are needed ~ for J

4 comparators, standards and backgrounds. Three microcaps are needed per s ampl e. A full wheel of 129 Microcaps accomodates 37 samples ((37 samples x 3 Micocaps/suple) + 18 comparators, etc) = 129, requiring two vials of 100 microcaps. One vial of microcaps is-adequate for 26 samples plus comparators, leaving 4 Microcaps extra.) Place straioht edge across Microcaps and mark all of them with a line exactly 22 m from one rd. Put microcaps back in vial.

4. Open the scintillation vial and corresponding oxygen-18 Microcap container for the first sample. Place the Microcap to be used for tritium into one mouth pipettor, and put a 20 p1 Microcap for oxygen-18 into the other mouth pipettor. Rinse the tritium Microcap by sucking it full of distilled water and blowing it out on a tissue. Repeat at least 6 times. Then, suck air through Microcap until completely dry, ,,

5. Select the first distilled sample and shake the water down to one end. (Remember that these tubes are evacuated, if they are opened while breaks still exist in the water column, much sample could splash out.) Score and break open the opposite end. Score and open the water end as close to the l tip as possible. Fill the tritium pipette with distilled sample, making sure i

l the meniscus at each end is flat, and no water adheres to the outside of the pipette. Blow out the sample into scintillation fluid. Rinse pipette 3-4 times in the scintillation fluid in the scintillation vial, and cap the vial.

Then fill pipette with distilled water and blow this out on a tissue. Repeat this wash at least- 4 more times. Suck air through pipette until completely dry and clean. Now the pipette is ready for reuse. ( Olexane-based

- scintillation cocktails are very hard to wash out of microcap pipettes. We try to use the same tritium pipette for all samples in a run, in order to eliminate possible errors due to slight variance in volume between pipettes.

The most frequent cause of error in tritium analyses is pipetting error. It's

-_ , . . _ . - _ . _, -, . - _ . _ _ _ . _ . _ . ~ . . _ _ - _ . . _ . . . _ . , . _ . _ . . _ - . _ , _

-- -- .- .. -~- - -. - . _ _ . . . _ _ _ - . . . - _ _ _ _ _

-29 best to practice before working with actual samples.) Seal the scintillation vial, swirl to mix it and wipe off fingerprints from the glass. Now it is ready for counting. Repeat, to obtain two vials per sample.

6. Pick up the Microcap in the other pipettor, fill it as closely as possible to the 22 m mark with sample water. The exact level is not critical, but it should not exceed 22'12 m. . Suck the water column away from the end of the Microcap a bit and flame seal that end. [Microtorches are available ("The Little Torch", Tescom Corp., Instrument. Division, 2600 Niagara Lane North, Minneapolis, Minnesota 55441) that produce a small, hot flame.

This is useful for rapid flame sealing of microcaps without causing the sample to boil.) Remove the Microcap from the pipettor and flame seal the other 18

end, place the sealed Microcap in the marked container. Make two more 0 i tubes as above (3 tubes per sample), and put these in the same container.

i 7. Flame seal the tube containing the remainder of the sample, ano save 18

) it in case reanalysis is necessary. (There are two -steps in the 0 analysis procedure where an entire wheel of samples can be lost if the equipment malfunctions. This has not hap,)ened yet, but it could. Thus it is wise to have enough extra sample to repeat the 0.-18 analysis if necessary.)

! 8. Prepare 6 I8 0 Microcaps for comparators. Comparators contain water i

! of a known I8 0 content. Since each run comes off.the cyclotron at slightly different activities, comparators must be included in each run. Also prepare 3 more Microcaps with comparator water to be run as unknowns, to serve as a check on the entire procedure. Prepare 3 Microcaps each of 1) water from an uninjected animal (background), 2) a 1 to 200. dilution of injection solution in distilled water, and 3) the distilled water used in making the 1 to 200 f

dil uti on. The latter values are used in the body water volume calculations, f

l

4 TRITIUM ANALYSIS The scintillation vials containing 5 or 10 91 of distilled sample water are counted in a liquid scintillation counter to 1% error (ca. 50,000 counts accumulated) or to 20 minutes. We use a Beckman LS 230 counter 2 at the following settings: gain 330, channel A (full tritium window) preset error 1,0%, preset minutes 20, low sample reject OFF,. single cycle, external standard ratio ON, print mode CPM. Samples that do not attain 11 counting error during the first cycle are recounted with preset minutes at 100, External standard values are used only to indicate unusual quenching, not to correct sample counts for quenching, because of problems with quench corrections (Springell,1969). If a sample is suspicious, it is discarded a,nd another scintillation vial is prepared and counted.

Because the doubly labeled water calculations depend .on the relative (not absolute) isotope concentrations in initial and final samples, factors that affect absolute values (such as quenching, counting efficiency, etc.) are not important as long as their influence is identical for both_ initial and final samples f rom an animal . If this condltion is met, it is not necessary to convert CPM values to disintegrations per minute (DPM) or to milli-or microcuries. The toluene-based scintillation fluid we use remains clear with addition-of up to about 50 pl of water. -More than this causes cloudiness which changes quenching, so we avoid these problems by using small sample volumes. It is' also not necessary to' correct CPM values for radioactive decay that occur < 11 between the times of injection-and sample analyses, provided that inital samples, final samples and injection solution dilutions are all counted together; because decay will. have been identical, on e proportional basis , i n all - sampl es .

e i

i OXYGEN ANAL.YSIS -

When the oxygen-18 tubes are processed, they are given unique color code marks for identification, centrifuged to eliminate bubbles, placed on'an aluminum wheel, bombarded in the cyclotron, removed from the wheel, inverted and recentrifuged, opened and the irradiated end of the Microcap discarded, and the other end now containing the kater is -counted in a gama counter. The data are fed into a computer for semilog least squares regression analysis and extrapolation to time zero. These intercepts are then compared to the I8 intercepts of tubes containing water of known 0 content (the 18 "comparators") which were processed in the same run, and 0 contents of the I unkowns are calculated (in triplicate for each original sample). When samples .

18 o f known 0 content are processed along with dif ferent runs _, the results ,

sometimes ccee out slightly different for reasons as yet undetennined. Thus, it is desirable to process all samples from a given study in one cyclotron run if possible. If two _or more runs are necessary, it is advisable to include at least three tubes of comparator water, designated "known 180 " so they won't be treated as comparators, in each run so that the results from separate runs l can be normalized against each other. It is highly ihadvisable-to put the IS O run and the final sample = from initial sample from a given animal in one 18 0 run,

! that animal in another 18 C02 production equations require 0 in terms of atom- % excess, which l is atom % with background subtracted. We have not found-an-adequate constant 18 0 in uninjected animals.-and in any event, the occasional l for background j variability of I8 0 results from.our analysis suggests that use of a constant  !

a l is not advisable, especially _if final samples are close to -background., l 0 Therefore, at least 3 tubes of water distilled from samples taken from an 18 uninjected animal should be included in v' run to -provide a value for 0

~. . . . - - . . - . - . - . - - - . - . . . _ - . - . - . . - - . - - . _ ~ _ - _ . - - - _ -

s bac kground. Six tubes of background water from two different animals would be better, t

C ALCULAT!]NS The data for blood sampling times, body masses, body water volumes,-

tritium and oxygen-18 are recorded on' the data sheets shown in Appendix 1 and

2. Body water volume at the time of injection (TBW1 ) is calculated as oxygen-18 dilution space according to the equation shown in Appendix 1. If body mass did not change greatly during the measurement period, we usually calculate TBW2 as = BW2 x (TBWi /BW3 ). Alternatively. TBW2 can also be measured as isotope dilution space by reinjecting recaptured animals, after blood sampling, with a small ' dose of H 3 2

0, or with HHO (but see

~

3 errors associated with HHO dilution in Nagy and Costa 1980). If the animals are small and abundant (e.g., beetles, small lizards), 'T8W2 can be measured by killing and drying to constant mass at 6500. If this is. done, we then calculate TBW3 as = BW 3 x (TBW2 /BW 2

) because drying is more accurate than oxygen-18 dilution space. TIME (days) in Appendix 2 is expressed with portions of days given in decimals, not in hours or minutes.

Rates of CO2 production, water influx, water efflux and body mass change are calculated using a single calculator program and the data between the double-lines in Appendix 2. This program includes the following equations:

m1CO

3) 2 = 51.86 (TBW 2 -TBW 3 ) -in ((0 3 -O g )(T2 -T B )/(02 -O g )(T3 -TB N gh (BW3 + BW 2 )

in (TBW 2

/.TBW) i TIME

, where 03 , 02 and Og are oxygen-18 in atom % for initial, final and background samples,3 T , T2 and-TB are tritium values in CpH for initial, final and background samples, TBW is total body water in m1, BW is body mass in 9.

e TIME is days (in decimals) between blood samples and in is natural logarittn.

l "I H 2O out , 2000(TBW p - TBW ) in [(T 3 -T g)(TBW3 )/(T2-T)(TBW)I g l

2) 3 2 ,

kg day (BWg + BW2 ) in (TBW2 /TBW3 ) TIME "I H 2O i n , ml H2O out ,2000(TBWp - TBW3 )

3) kg day kg day (BWg + BW 2 ) TIME

4)  % BW Change , 200 (BWp - BW3 )

day (BW 3 + BW2 ) TIME Equations 1) and 2) describe situations where body water volume changes linearly through the measurement period. Equations are also available (Lifson and McClintock 1966, Nagy and Costa 1980, Nagy 1980) for calculating C02 and H2 O fluxes from isotope data when body water pools change exponentially or ,

! do not change at all during the measurement period. We use the linear-change equations for almost all situations. Rapidly-growing animals usually have exponential increases in TBW, so the exponential equations would be best in-these situations. For adult animals, TBW 2 is rarely exactly equal to

TBW 3 , so the above equations apply in most of our nudies. When TBW2*

f TBW 3

, we still use the -linear-change equations (to save the trouble of l reprogranning the calculator); but we change TBW3 or TBW2 by a small i

This has no influence on calculated results but it

~

amount (say 0.00001-ml).

allows- the calculator to proceed normally, if TBW2 exactly equals TBW3 in l the linear change equations, some values become zero and the equations will l not compute.

Arranging for results to appear in units that are mass-specific rates--

) (e.g., ml CO2/ g h and ml H 2O / kg day) was done for the convenience of physiologists, who are familiar with such units. Doubly labelled water actually measures CO2 production and water flux per whole animal per

!r

l I .

measurement period. Dividing these values by mean body and mass and by time incorporates the errors in these measured parameters to the overall error in the flux values. Also, dividing by body mass is usually done to account for i

differences in whole animal values that are Jue to differences in body size between individual subjects. Using straight body mass implies that the -

appropriate mass correction is 91 0 .There is much evidence that the ,

exponent is actually lower than 1.0 (Heusner 1982), so mass specific values may be in error for this reason as well . The appropriate units to use should depend on the questions being asked in a particular study. Because of this, r

care must be taken when doing statistical tests with doubly labelled water resul ts .

CONVERSION FACTORS The factors we use for converting metabolic rates from units of CO 2

production to the equivalent units of heat production (in joules), oxygen consumption and metabolic (oxidative) water production are shown in Table 3.

l The factors for carbohydrate, f at and protein were calculated from values '

l given by Schmidt-Nielsen (1952) and Schmidt-Nielsen (1975) assuming 4.184 kJ/kcal. The factors for the-desert plant and the mealwonn diets are based on amounts of carbohydrate, fat and protein assimilated and metabolized from these diets as determined in laboratory feeding trials (Nagy,1983). The factors for the seed and anchovy diets were estimated from gross chemical composition of these foods by assuming that the gross composition reflects +k same proportions of carbohydrate, fat and protein as in the dry matter that is acturtly metabolized from the food. The factors on the right side of Table 3 are per gram of metabolized (not gross dietary) dry matter.

~ - - , , , - _ .- ,... . .._. ___.-.._ _ . _ _._ _... _ _ _ _ _._._ _ _ .-

_ _ _ . _ _ = ._ _ _ _ _ _ __ . __ _ _ _

! Table 3. Conversion factors relevant to metabolic rate measurements made using doubly labeled water.  !

l . j i

(

microliters microllters  !

I metabolic metabolic j' J ml 0 2 "20 formed (kJ) m1CO 2

ml 0 2

H2 O formed  !

ml CO ml CO ml CO g dry g dry g dry g dry  ;

2 2 2 matter matter matter matter  :

1 ,

t

. CHEMICAL COUMP00NDS l Carbohydrate 20.8 1.00 0.662 17.5 840 840 556 i l i Fat 27.7 1.41 0.'754 39.3 1420 2000 1070 5 Protein, urea as  ;

! end product 23.1 1.23 0.509 18.0 778 960 396 l

! Protein, urate as i .end product 24.8 1.35. 0.546 17.8 718 -

970 392 .;

i i

! DIETS 3

Plants, mixed,  ;

l eaten by desert

herbivorous lizard 21.7 1.08- 0.637 l Mealworms (Tenebrio i larvae), eaten by -

l insectivorous lizard 25.7 1.33 0.660 i ~

l Seeds (millet) eaten '

by granivorous bird 21.9 1.08- 0.658

- t I

l Fish (anchovy) eaten .

l by penguin 25.8- 1.37 . 0.617  :

i

! O f 4 19 1

1 i i

+

f

" *~

t t . - - - .- - - - . _ _ _ _ _ _ _ _ _ _ _ _ _ _ ___ _ ______ __

l l- The literature-contains conversion-factors that differ from those of Schmidt-Nielsen. King and Farner (1961) suggested that the following factors are appropriate for birds: carbohydrate, 21.14 J/ml CO  ; fat 27.75 J/mi 2

002 ; pretein (urate), 27.25 J/ml C02 . Their factor for protein differs from that of Schmidt-Nielsen (Table 3) by 12%. We lack the expertise needed to decide which factor is correct.

ERRORS There are many assumptions and potential errors in the doubly labeled water met hod. These have been discussed by Lifson-and McClintock (1966) and analyzed by Nagy and . Costa (1980) and Nagy (1980). We are concerned here with the errors that are most likely to cause problems during the field work and laboratory portion of a study.

In the field, common errors include mistakes during injection, contamination of blood samples by exogenous water, uptake of exogenous CO 2

by labelled animals- and isotopic fractionation in some small tropical animals. -Mistakes during injection cause errors in TBW values estimated from isotope dilution space. These can be minimized by careful filling of injection syringes and noting (then discarding the resulting TBW values) when leaks occur during injections.

Contamination of blood samples in the field by exogenous water should not e ff ec t C02 production values, but will affect H2 O flux values. In most circumstances, exogenous water that gets into blood samples will not be enriched with tritium or o ngen-18 Common sources are rain water and the water produced by the flame to seal the blood tubes. (To minimze the latter.

l in ert the tube into the flame with a slight angle at the sealing end above horizontal, or tilt the flame away from the tube opening, using the base of l

l w w e-- w- , .%-= .- - e e % . .--r~~ .,.3m  ?,r,,11 -~we,*a , ~w.--u-r-wwr+ + e- w

• m-'+ 7'

t

• e the flame for sealing.) Because both the triti9m and the oxygen-18 in the contaminated sample will be diluted to the same extent, and because CO2 production is calculated from the difference between the washout rates of the

~

two isotopes, calculated CO2 production should not be changed when one blood sample is contaminated by unlabelled water. However, water flux is calculated from tritium data alone, and this value will be incorrect if one sample contains a lower tritium concentration than it should, if a labelled animal in the field inhales air containing unlabelled CO 2 , this CO2 should equilibrate with body CO2 and hence increase isotopicaI'y measured CO2 produc tion. A likely situation where this may occur is with animals that live in groups in enclosed nests .or burrows. This problem can be reduced by labelling.all animals in the group, so inhaled CO2 -

18 contains 0 enriched' to the same level in all animals. .

A problem with doubly labelled water measurements has occurred in some small tropical lizards we have studied. CO2 production values were unbelievably high, apparently because differential fractional ' evaporation of H

2 0 and 3HHO from the highly permeable skin of these animals.

. Attempts to correct these data by usin3 published fractionation factors for the various types of labelled water and CO2 (see 1.ifson and McClintock 1966) have been unsuccessful . Accordingly, we are reluctant to recomend doubly labelled water measurements in small animals with permeable skins living in moist habitats.

In the laboratory, samples can be accidentally lost at several stages of i the procedure, especially those involving vacuums, fluid transfers, and flame seals that are leaky. We strongly suggest practicing the entire procedure beforehand, using unlabelled blood samples.

What are the comon errors in sample processing in the laboratory that cause errors in calculated CO2 and water fluxes? Errors in tritium data are almost always due to pipetting errors. If there are doubts about any tritium value, we repipette and count the sample agsin, it is a good idea to remeasure both initial and final samples from a given animal if one of these samples looks suspicious. Small erro*rs in tritium values cause relatively small errors in water flux estimates, but C02 production estimates are much more sensitive to error in tritium values. However, CO2 production estimates are also very sensitive to errors in oxygen-18 values, so strange-looking metabolic rates can result from errors in any one or more of the four isotope measurements for a given animal, or from inaccurate isotope background values. Tritium concentrations are usually high relative to tritium background, so errors in the latter are usually inconsequential.

l However, oxygen-18 background (ca 0.202 to 0.204 atom %) is a substantial f raction of oxygen-18 in blood samples (0.25 to 0.7 atom %), so errors here can have a large effect because the background _value_ is subtracted from every other oxygen-18 value in the study. Investigators may wish to include six Microcaps of background to insure accurate measurement Of this important value.

Similarly, the oxygen-18 levels of the diluted injection solution and the i distilled water used in that dilution are important, because these values are 1

i used to calculate body water volume in all animals. Errors here can also have a substantial effect on calculated flux values via errors in TBW estimates.

However, T8W values derived from oxygen-18 dilution-can be checked qualitatively by comparing them with tritium dilution spaces calculated for the same animals (tritium spaces are- usually 2-6% of body mass higher than oxygen-18 spaces), or with spot-check measurements of TBW using drying to constant mass. It may be advisable to increase the number of oxygen-18 measurements for these solutions as well .

4 An important source of errors in the oxygen-18 analysis is contamination of the water in distilled samples. Water containing traces of organic matter will- bubble and cavitate when exposed to the cyclotron beam. Thus, less H

2 0 is being irradiated than intended, yielding falsely low values.

Seriously contaminated samples may explode during. bombardment or later when being processed for counting. Becaush the samples are being centrifuged continuously during bombardment, bubbles may not remain in the tubes when the samples are removed from the cyclotron, so it is difficult to detect these pro bl ems . The main cause of contamination is poor distillation in the Pasteur -

pipettes. This is usually due either to blood spattering into the narrow end of the Pasteur pipette during or shortly af ter evacuation, or to decaying of samples before distillation. If samples are odiferous, that means that ,

volatile chemicals are escaping from them, and these will distill over along with the water. It's best to keep samples refrigerated to minimize rotting.

If there is any suspicion that a distilled sample may be contaminated, we redistill it at least one more time, perhaps two if it is.still smelly. We are careful to note any samples that contain bubbles or are pressurized following proton bombardment, so these values may be discarded.

Unfortunately, when a sample is contaminated, all three Microcaps of that sample are usually pressurized.

PERMITS l

0xygen-18 and deuterium are stable isotopes (not radioactive), and no isotope permit is required for their use in field or laboratory. Tritium is radioactive, and radioisotope permits are required before using it in most circumstances. (Studies on very small animals may involve amounts of tritium that are below the license-requiring level of the state involved.) The total

Li -

i l

~

amount of tritturr in all the blood samples from a typical study is usually i

below licensable levels, because only a small fraction of the isotope that was injected is contained in the blood samples. Many colleges and universities are licensed to use radioisotopes on campus, and obtaining permission to use tritium in the laboratory may be stratcht forward. Pemission to use tritium in the field can be much more complicated, because state and local governments as well as landowners may have to be contacted for pemits or letters. We have found our campus Health and Safety Officer to be of great help with the permit process. The main drawback with using tritiated water is the time and effort that may be involved in obtaining permits for its use in the field.

However, in many instances, obtaining pemits has been quick and easy.

Other kinds of permits for field stt. dies of wild animais may also be ,

requi red. These include projects on endangered, rare, or threatened species, i

game species, migratory birds (federal permit required), some economically important species, and species involved with various human or domestic animal diseases, as well as studies done in areas that are national or state parks or reserves and other restricted properties. If blood samples will be transported across state or national boundaries, pennits may be required from-the appropriate public health or agriculture departments, because of possible transmission of diseases, l

ACKNOWLEDGMENTS This work was funded by Contract AT (n4-1) GEN-12 between the U.S. Atomic Energy Commission (which became the U.S. Energy Research and Development Administration) and the University of California, and by Contract DE-AM03-76-SF00012 between the U.S. Department of Energy and the University of California. I am grateful to my students and colleagues who have participated l

n 41 in the development of this technology during the last 10 years: Craig Adler, Justin Congdon, Paul Cooper, Eric Edney, David Eisenberg, Paul Franco, Ronald Gettinger, Bill Karasov, Bill King, Nathan Lifson, Bill Hautz, Philip Medica, Park Nobel Sally (Rockhold) Pixley, Howard Reiss, Robert Scott, Robin Severeid, John Stallone, Lynn Taylor, Mike Vensky, Stan Wakakuwa, Wes Weathers, Joe Williams and Bob Wood.

• 1 thank Bill Karasov and Joe Williams for reviewing the manuscript.

)

4 j

l l

l

r l

o- ,-

42-LITERATURE CITED

/Boyer, P.D., D.J. Graves, C.H. Suelter, and M.E. Denpsey. 1961. Simple procedure for conversion of oxygen of ortho-phosphate or water to carbon dioxide for oxygen-18 determination. Anal . Chem. 33:1906-1909.

Heusner, A.A. 1982. Energy metabolism and body size. I. Is the 0.75 mass exponent of Kleiber's equation a statistical artifact? Resp. Physiol .

48:1-12.

King, J.R. and 0.S. Farner. 1961. Energy metabolism, thermoregulation and body temperature. In A.J. Marshall, editor. Biology and Comparative Physiology of Birds, Vol. 2. Academic Prets, New York.

Lifson, N., and R. McClintock.1966. Theory of use of the turnover rates of body water for measuring energy and material balance. J. of Theoret.

Biol. 12:46-74.

/ ull en, R.K. 1973. The D2 0 method of measuring the energy metabolism of free-living animals. Pages 32-43 in J. A. Gessaman,- editor.

Ecological energetics of homeotherms. Utah State Univ. Press, Logan.

-Utah.

Nagy, K.A. 1980. CO2 production in animals: Analysis of potential errors in the doubly labeled water method. Am. J. Physiol . 238: R466-R473.

Nagy , K. A.1983. Ecological energetics. Chapter in R.B. Huey, E.R. Pianka and T. Schoener, editors. Lizard ecology: studies of a model- organism.

Harvard Univ. Press. (in press).

Nagy, K. A., and 0.P. Costa. 1980. Water flux in animals: analysis of potential errors in the tritiated water method. Am. J. Physio 1.

238 :R454-R465, s

e

. v o

,l Samis H.V., M.B. Baird, and H.R. Massie, 1974. Deuterium oxide effect on temperature-dependent survival in populations of Drosophila mel a no ga s ter. Science 183:427-428.

Schmidt-Nielsen K.1975. Animal physiology: adaptation and environment.

Cambridge University Press, London.

Schmidt-Nielsen,-X., and B. Schmidt-Nielsen.1952. Water metabolism of desert mamals. Physiol . Rev. 32:135-166.

l Schoell er, D. A., E. van Santen, 0.W. Peterson, W. Dietz, J. Jaspan and P.D.

Kl ei n. 1980. Total body water measurement in humans with 10 0 and 2H i

labeled water. Am. J. Clin. Nutr. 33:2686-2693.

Schoell er, 0. A., and E. van Santen.1982. Measurement of energy expenditure in humans by doubly labeled water method. J. Appl . Physiol . 53:955-959.,

Schoell er, D. A., W. Dietz E. van Santen and P.O. Klein. 1982. Validation of l saliva sampling for total body water determination by H 8 0 2

l dilution. Am. J. Clin Nutr. 35:591-594.

Springe11, P.H. 1969 Inadequacy of an external standard in correcting for quenching in liquid scintillation counting of tritiated-water in biological material. -Int. J. App. Radiat. Isotopes. 20:743-746.

Sutton, S.P., R. Dam, and M.K. Nielsen. 1974 Biological effects of D 0 2

adminstration to Coturnix , japonica. Life Sci. 15:2097-2108.

\ ood, W R. A., K. A. Nagy, N.S. MacDonald, S.T. Wakakuwa, R.J. Beckman and H.

Kaaz, 1975. Determination of oxygen-18 in water contained in biological sampies by charged partic1e activation. Anal. Chem, 47:646-650.

Zweens, J., H. -Frankena, A. Reicher, and W.G. Zylstra. 1980. Infrared-spectrophotometric determination-of2 0 0 in biological fluids.

P flugers. Arch. 385:71-77

.. _-